holy mackerel, almost 1k comments in less than 24 hours! Believe it or not I've read each and every single one of them. NOTE: There is nothing wrong with the welder. I'm not trying to diagnose / fix anything. This was purely exploratory and trying to make sense of something I saw with the old welder. Anyway, excellent suggestions all around, some contradictions :) but, armed with new information, expect a follow-up video. PS if you suggested turning it on and off again: I love you.

@williamwakely1398

6 жыл бұрын

Have you tried turning it off and on again? ;P Seriously though, good luck, and I can't wait for the next video!

@manickn6819

6 жыл бұрын

They say curiosity kills .... I am not convinced. I am curious about the answer too. PS. When you get the most likely candidates for the correct answer please make another video.

@TexasTimelapse

6 жыл бұрын

Manick N I'm curious too. Hopefully we find a definite answer.

@manickn6819

6 жыл бұрын

TexasTimelapse it would really be productive if one of the viewers is involved in a design company or if a reputable firm like Lincoln Electric can take this challenge as an opportunity to do a "friendly" technical discussion. Maybe Tony can consider reaching out to one of these firms. Its practically free advertising for them so I can't see them turning him down.

@schwarzarne

6 жыл бұрын

You could diagnose wether what you see comes from your setup or your welder by flipping the current clamp, make the grounding wire go through it the other way. If it is the clamp you will see no change on the scope, if it is the welder the output will be flipped, with more positive current than negative current.

Hi Tony, some notes for you, from an old electrical engineer: 1) So the AC clamp meter is an inductive pickup. The output of this will not have a direct reference to the DC offset showing up on the scope. The clamp acts the same as a capacitor, blocking DC. 2) The symmetry you are talking about is called _Duty Cycle_ [%] where 50% is spending 1/2 the time on. 0% Duty cycle is off all the time. 3) The DC offset you are seeing on the scope is likely caused by the scope being set to *DC* input (coupling) on that channel. The offset would be affected by the duty cycle, kind of like an average of the voltage over time. Setting the scope to *AC* coupling should stop this offset from occurring. Your going to need a different pickup to read DC offsets coming from the welder. 4) The high frequency start is bad juju for the scope. High energy RF is getting coupled into the scope, either from the probe or just having the scope in close proximity to the arc. The energy is overwhelming the input protection on the channel, if coupled next to the scope. RF (if from the lead) could be reduced by choking the leads (looping them through a torroidal choke or similar system of creating a big inductor in the lead). For proximity coupling, move the scope further away or try and shield it using some metal sheeting. 5) For sure, trying to keep measuring that as-is will likely destroy the front end on that scope. It's not happy right now. Trying to couple high energy or high power current flow to an instrument that can pick up noise from your fluorescent lighting (very sensitive pickups) can be challenging. Thanks for all the vids, loving them. - Eddy

@king4aday4aday

6 жыл бұрын

You're right, I think from the welder's perspective it's symmetrical because it doesn't have the same ground reference, and the DC offset influences this.

@PixelSchnitzel

6 жыл бұрын

Deathlok67 (Eddy) is correct on all points. Weld balance in AC TIG is a matter of pulse width and the relative duration of positive vs. negative portions of the cycle. www.millerwelds.com/resources/article-library/understanding-ac-tig-balance-control

@JohnDoe-rl9pp

6 жыл бұрын

This was my first instinct too, but the clamp is supposed to be picking up CURRENT, not the voltage. Zero voltage reported to the scope should correspond to zero current (technically, zero magnetic field) through the clamp. There's a lot of potential sources of error here, though. If I were trying to track this down, I'd want to put a true voltage probe onto the welder electrodes. Theoretically that should be fine - they only run at a couple volts, and the welder shouldn't provide much current to the scope. Get anything wrong though, and you'll be out shopping for new test equipment. (EDIT: Be REALLY SURE where all your ground references are before you try something like this! And as Eddy says, you'd want some good signal filtering before letting it near your scope!) If the welder is keeping constant +/- voltage, and adjusting to maintain only the average current, the electrodes might be responsible. I'm just guessing, but the tungsten electrode is probably better at emitting electrons than the steel bench.

@zaprodk

6 жыл бұрын

The clamp meter is based off of an hall sensor, and will read DC just fine. If it were AC, you would not see these nice square waves on the scope, but something more alike a sinusoidal wave.

@Audio_Simon

6 жыл бұрын

zapro_dk There are inductive AC clamps too, but I thunk you are correct. So the scope needs to share the welder ground. Should be no harm in that back at the wall. Just donut ground the scope art the wall and at the welder. Ground loops are naughty!

Flip the clamp meter over. If the waveform flips over on the scope then the scope and clamp meter are fine and the machine is outputting an asymmetric wave. If the waveform stays the same on the scope then it's just a measurement problem.

@origamihawk

6 жыл бұрын

This sounds so stupid it might just work

@jsincoherency

6 жыл бұрын

Why does it sound stupid?

@iwtommo

6 жыл бұрын

Literally this. Just posted this before finding this more upvoted comment. Those clamps are usually 0-60a and not -60 to 60a :P Im surprised it does any -a at all :0 EDIT: If you need some brainbending electronics skill, you'd best be asking mikeselectricstuff.

@robmckennie4203

6 жыл бұрын

I expect what the commenter means by "stupid" is just simple. This explanation could also explain why the waveform is nice and square on one side, but weirdly sloped on the other

@Split10uk

6 жыл бұрын

I was thinking alone those lines. Try and repeat the error in the first instance.

turn the oscilloscope upsidedown , if the positive side increases and the negative decrases,then that tells you its a gravity problem , so just put the oscilloscope on its side and they should equlize :P

@paultrgnp

6 жыл бұрын

mario kovac Touché!!!

Why does it feel like you keep mentioning that it was a different issue that caused you to replace your old welder, is done just in case the wife decides to watch this one? :)

I'm going to get spoiled having 2 videos in a week

The problem is the blue weld cart

@johnpossum556

6 жыл бұрын

I'd usually say it needs some John Deere green on it 'cept with the red welder it'd look too much like a Christmas decoration all year long and we can't have that!

@ArmchairDeity

4 жыл бұрын

The hell you say.... blue is my favorite color! 🤨

You should be able to verify if it's the clamp by simply clamping it 180 degrees to the direction it is now. Current clamps measure in one direction, flip the direction around and the waveform will either be mirrored, or it will be the same. If it's mirrored, then the clamp is probably working correctly.

@geobas1

6 жыл бұрын

EFormance Engineering still won't work. The clamp outputs a voltage as voltage is the only thing scopes can actually measure. That voltage it is inducing isn't referenced to any zero so absolute measurements are impossible best this set up can do is measure peak to peak and show the waveform.

@eformance

6 жыл бұрын

I think you misunderstand what I said. A current clamp's output is + or - based on the direction of the current flow. If you reverse the direction of the clamp, the output signal is also reversed. This has nothing to do with reference to ground or anything. His clamp is battery powered and DC coupled to the Oscope, so the ground reference is provided by the battery in the clamp and the "zero" calibration.

@ppdan

6 жыл бұрын

That's a very good advice!

@pjtruslow

6 жыл бұрын

I had a similar idea. sounds like a pretty good idea for me. additionally should check that the scope is set to DC coupling, although if it were set to AC, then the square waves would be a lot more rounded, and more simmetrical.

@eformance

6 жыл бұрын

With AC coupling he wouldn't have any DC offset in the waveform.

Rotate clamp 180, if waveform flips then its the welder, if waveform stays negatively biased its the scope

@fellipec

6 жыл бұрын

For some reason this make some sense.

@LEDtest

6 жыл бұрын

My first thought, exactly. Also, you could try the higher current range on the clamp (switch to top position). And the scope is probably going crazy, because of EMF generated by the high frequency startup. Try putting the scope at a greater distance from the welder, and if needed, shielding it.

@ArkanoidZero

6 жыл бұрын

This is one of those things that doesn't make sense if you think about it only a little bit, but is actually correct, the reason is that style of probe is actually measuring the inductance of the wire, so a positive voltage traveling through it creates a curved electric field via the right hand rule, so flipping the probe around will cause the measurements to swap polarity.

@LEDtest

6 жыл бұрын

As near as I can figure the probe is for AC as well as DC current. It should be using a hall effect sensor to measure the magnetic field, which is also the reason for the battery in the clamp. But there may be a problem with residual induction in the clamp somewhere, which would show when flipping the clamp.

I work in high voltage electronics (large, angry pixie wrangling) for a living. Not welders, but things like make a lot of arcs none the less, but hopefully not on purpose... It could be a number of things you are seeing. You’re using a cheap clamp on current probe - where I would use a high dollar toroid current transformer that has been calibrated for the current level I would be measuring. They can also be frequency sensitive i.e. “meant to be used at x to y frequency” and anything outside of that you will see distortion... And those are for AC waveforms. For DC you need a Hall effect type of device. Also the shape of your waveform is complex. Meaning not a sign wave... which can mess with measurements. And what are you comparing to? What did the manufacturer of the welder choose to base their measurements on? RMS? No RMS? Are you sure the welding waveform is supposed to be perfectly symmetrical? Depending on the circuit arrangement, it may be that they get a nice clean square wave in one direction, but some ringing or other type of garbage in the other direction. Actually, I rarely see such clean square waves. Usually overshoot on a rising edge is a fact of life if not full blown ringing. I see that wave form and see nothing wrong because I don’t have anything to compare to or schematics to look at and determine it’s just the way it’s designed. I tend to try to think like my doctor when it comes to “problems” like this. Is it really a problem or am I just perceiving it as one because I don’t have 30 years of medical experience to tell the difference between a cancerous growth or an ingrown hair... I would let a real welder (like Jody?) who has more experience welding put the thing to work and see what they think... If they have a problem, then it’s a real problem, otherwise you may be overthinking things. Oh, and the RFI from the HF start could break your scope. Not so much from actually frying something, but it could erase or add bits in flash that could brick it.

I call for a collaboration with ElectroBoom. He can fix it and hurt himself in the process.

@nicktrousers

6 жыл бұрын

Gdsmith504 good idea

@tmdcbass

6 жыл бұрын

I now have a picture in my head of ElectroBoom licking a TIG torch at full blast. Thanks!

@Parents_of_Twins

3 жыл бұрын

That dude is freaking hilarious. Stumbled onto his channel on accident and watched like 40 videos in a couple days. My kids were referring to him as "The Crazy Guy".

Hey Tony, I don't know how the current should look like(not a welding machine designer) but I know a lot about power electronics (and making measurements on them - just writing up my masters on power electronics). 1) The "I nearly fried my scope" its not your fault its the scope. The owon scopes are crap. I bought one for my first scope and in a couple of weeks I had problems with the knobs, the fan, the trigger, it shows wrong stuff on the screen sometimes, the record lenght is not right ... for more you can look at a review EEVBlog made on them. So the wierdnes you get form the hf start its the probe picking up strong elector magnetic interference (from the high frequency, high voltage sparks) and passing it to the scope and the scope treking out coz its made badly. I would recommend rigol over owon anyday (for cheep new scopes). 2) You have your clamp set to 1mv/10mA. So it gives you 1mv out for 10mA that you measure. It also has a 1mv/100mA range. This ranges are so you can get more sensitivity and the lower end. But - all of this clamps have a battery in them. Now if the internal supply is 5V and you mesure 90A, so 90000 mA so 9000 mV so 9 V would be the output of your clamp, but it can't give you that becouse it has only 5V, so it gives you that. This is why you have your 1mv/100mA range. In the manual it says for 1mV/10mA: range form 10mA to 2A and for 1mV/100mA 20A to 65A. So 1 step is to change the range. 3) You don't need to run your scope on battery power. You are not connected to the welder any way, because you are measuring current with an insulating method. That you don't have your BNC - your scope - ground referenced may be causing problems in point 1). In labs we always measure with all the instruments grounded and insulate them with measuring methods - especially for current measurement. It gets a little bit more advanced if you would want to measure voltages, then battery scopes (better common mode rejection) and insulation probes come on to play. But if you only have your current probe plugged in you can have your scope connected to the mains. (I always do). 4) All hall effect sensors have a balance button, make sure to press it, rotate it, (its different on different clamps) if you see an offset when you are not measuring any current. 5) The current rating on the probes would be RMS presumably. Or DC. Or peak(amplitude). But defiantly not peak-peak. 6) The shape of the waveform does not sit right with me. For one side to have some shape to it, and the other to be dead flat - i would defiantly look at the measurement equipment first and then at some incorporated woodo black magic. Thoe I don't knock the htp folks, but to have suck a nice squre wave and high amps its a bit iffy. I would expect to see some overshot or a more "regulated" looking curve. Google pid regulator for the curves I'm talking about. Hope this helps. Leme know if you have any other questions. I learned so much of machining form your videos that I'm happy to help. Best regard, Marko

@ljsaldana4739

6 жыл бұрын

Marko, with the slam dunk.

RF start interference : Sounds like classic EMI. I would look to try... 1) connect scope ground (the outer shell of the BNC input connector) straight to the welding table using a short as possible strap (think 1" long). Any noise picked up from the arc on the outer shield of the probe cable will return to the welder ground through this rather than via the sensitive internal electronics of the scope. High frequency noise wants to get back to its point of origin in as short a loop as possible and will find every path it can! 2) Minimise current loop area of weld torch output to ground clamp. Route the two cables next to each other or loosely twist them together. AC current in a loop = magnetic field = EM field = interfere. 3) increase distance between scope and welder arc to minimise effect of large EM radiated field from arc. 4) Add a metal "counterpoise" or shield directly between arc and scope that is resting on and connected to weld table for its length, not just at one small point. Again to try and reduce effects on scope from arc. 5) Depending on the frequency of the high frequency start you can put ferrite cores around the probe cable to reduce conducted noise on the shield but selecting the right ones has several variables. Last resort. I've been a fan of your videos for what seems like ages so I hope these comments are helpful. Ping me if you need more info.

To rule out the clamp, put it on opposite on the wire so it runs back-to-front instead of front-to-back , this will reverse the waveform on the scope, and if the a-symmetry keeps being the same, the probe is not to blame.

@superbone8724

6 жыл бұрын

Was about to say the same thing. He should try this.

@rr974

6 жыл бұрын

That was my idea as well. If he flips the clamp the waveform on the screen should flip upside down if it really is an asymmetric current. If it stays exactly the same the current is symmetric but something in the measurement equipment is not.

this was hysterical and informative. I never would've hooked up an oscilloscope to welder or a tig, I'd be worried about blowing it up. Glad to be able to see this on KZread. I couldn't afford to do this experiment, especially since I haven't access to a scope, muchless able to afford one. my unit is just a Sears Home-N-shop 30-200 Arc welder.

I must say the level of information the viewers possess is the most impressive thing I've seen in a long time.

I think severianvoid found the missing puzzle piece here, see ""Exploiting Advances in Arc Welding Technology" By Woodhead Publishing ". As noted in the on page 12 in the image the DC is clipped by arc rectification (rectalfrycation is the only way I can remember this). This is an interesting topic one can google. I'm not an electrician and I'm a home hobby welder who has one welding class at the local community college, but this does make sense to me. Tony, you're on a roll with the videos. Keep up the great work.

The hantek clamp has a max frequency of 40khz.. it error percentage is 8% +- 0.3A measuring AC.. compared to 4% +- 0.3A measuring DC.. you also need to "zero" the current clamp ... duty fully ripped from the manual. >>. it is needed to turn the zero adjustment knobs on the clamp until the multimeter reads zero. That means pushing down the zero adjustment knobs on the clamp until the multimeter reads zero. Then let go hold of the zero adjustment knobs.When using a scope,set DC coupled mode to measure DC current. If you are just setting the "zero" point on your oscilloscope you most likely have a dc off set from your clamp compared to your scope

@erg0centric

6 жыл бұрын

tl,dr: RTFM

@maesto

6 жыл бұрын

Based on the scope settings on the screen I would say he hasn't set the offset in the scope.

My pet theory is that This Particular Old Tony is trollin' y'all for the spectacularly poorly written comments, which he's going to read out loud to his family while they all wet their pants, roaring with laughter. Good job so far, fellas!

@imagineaworld

5 жыл бұрын

Hahaha... hey, wait a minute!!

Turn the clamp meter around. If it still clips on top, it's the meter/scope. If it clips on the bottom, it's the welder. My money is on the meter.

@roboticus3647

6 жыл бұрын

Excellent idea! (Owner of 7 scopes (at the moment...))

You are killing it with the videos this year.... Love watching your videos. Without crackpot ideas, there would be no progress.

Have you turn'd it off and on again? :P

@thebateman7949

6 жыл бұрын

Haha! I said the same thing. Hafta go delete my comment now. :(

@davydmir6565

6 жыл бұрын

hahaha lol!

@Twisty1024

3 жыл бұрын

Tony loves you

Your videos are inspiring and I am always excited to see a new one pop up

Before all, do a self calibration on the scope with the built in probe calibration output. After, For try, just hook up the clamp to a 200W lightbulb, then You will know if the clamp is out from calibration. You have to see a total symmetrical sine wave. Do 2 turn around the clamp, then You need to see 2x the current. If not symmetrical, You need to open the probe and find the balancing calibration pot and adjust.

@PSUQDPICHQIEIWC

6 жыл бұрын

If it's a hall effect sensor with a steel yoke (or some other material with significant retentivity), it's entirely possible to magnetize the yoke by measuring strong DC pulses. This tends to offset all other measurements, and doing an electrical calibration to remove the offset might not fix the problem if the remanent field is strong enough. I know I've had to manually demagnetize my clamp meter more than once because I couldn't cal it back to zero. That said, it's trivially demagnetized by measuring strong AC signals, soooo I don't think it's necessarily an issue.

@pjtruslow

6 жыл бұрын

The probe compensation output on the scope is to adjust the compensation capacitors on your passive probes, not to calibrate the scope itself

@BMRStudio

6 жыл бұрын

pjtruslow just to check levels, timing and general A/D quality

@BMRStudio

6 жыл бұрын

HammyJamPants that’s why he need to hook to ac line with 200w bulb...

@pjtruslow

6 жыл бұрын

the probe compensation output is a square wave from zero to 5V, you can't even expect it's voltage to be precise. the only characteristic about that output that is important is that it is a very sharp square wave, and as such you can't really use it to diagnose the oscilloscope.

I'd put my money on the current clamp. That is, I wouldn't put my money on a CC-65... The positive half-period seems too flat, as if the clamp is clipping the signal. When you turn off the welder at 6:49 the "wavy" parts of the current have similar amplitude, but the two edge spikes from the negative portion are "flat spots" in the upper part.

Hay Tony, Reading through some of the comments there is only 1 I can see that states the obvious. Start with a known signal, put the clamp on the Live wire of a Lamp or even the welder, check the waveform, should be a perfect sine wave. You don't need a ground line unless you have noise or are looking for a comparison between that ground and the voltages being observed. As with all electronics, start at the beginning with something you know, then progress from there.

Hey old Tony, Do your tests again, but now with the Amp-clamp rotated 180 degrees. (So that the ground wire enters de clamp from the other side). Now you measure the current in the oposite direction and the picture on the scoop should also flip 180 degrees, so now more positive than negative. If this is the case your Amp-clamp is working fine...and is the issue something else. 👍😁

@n2gpin

6 жыл бұрын

The fact that the positive portion is so evenly clipped leads me to believe it has hit saturation as opposed to the negative portion, which shows some overshoot and correction. I don't know whether that implies the welder over the test setup, but since two welders show virtually the same signal on one test setup, I would guess the test setup is where the issue lies. Or the welders could be set up like that by design. Sorry, no answers here...just an observation.

I was a firecontrol technician in the navy for 22yrs, When we had to troubleshoot the radar transmitter, we needed to make sure that the scope/meter was grounded at the same potential as the equipment we were wanting to test. Also HF start will affect your scope because it doesn't appear to be of a shielded variety.

Pretty likely that the clamp is the culprit, and is saturating at a peak-current. The asymmetry of the internal settings makes it so that the +threshold is different than the -threshold. Please check the following (in sequence): 1) A simple test: *Reverse* the clamp direction (put it on backwards on the same grounding-wire) - Does the waveform on the scope invert? If it does, it means the source is the welder - if not, it is your clamp. 2) If it _is_ your clamp - are the batteries in the clamp okay? Bad batteries affect the DC-balance. 3) I notice that there's a range-switch on the clamp - is it on its least-sensitive setting? A clue to all this is a freeze-frame at 6:49 ish that shows a pretty waveform towards the end of a welding-cycle - the upper and lower clipping-levels are clearly visible, whereas the rest of the current is a nice sinusoidal and _symmetric_ 'ringing'. Cheers!

As a practising pixie wrangler, it looks like your scope may have some setup errors. I would recommend that you check your scope's settings. Typically you will have three choices for coupling AC which will remove any DC bias in the signal, DC which will show you the DC levels and Ground which will give you a flat line without any other signal. When I test a scope I like to set it to AC and then adjust the vertical position to make sure that the 0 signal is on the center horizontal line. Next I will use either the calibration output of the scope or connect to AC main power and verify that in AC coupling I have a 60 up 60 down sine wave. Once I know that the scope is properly displaying some easy to verify test signals I can look with some confidence at other things. As always with this sort of thing you may find it useful to read the manuals for both the scope and your torch - and keep in mind that when going from theory to practice your mileage may vary.

@zman97211

5 жыл бұрын

This is absolutely true. But simply grounding the scope input and centering that on the display should be enough for this. The clamp is AC coupled to the cable it's measuring already. DC is likely introduced because the clamp is saturated.

Everybody thank Tony for sharing with the group.

@TheFeller1554

6 жыл бұрын

Thank you, Mr. This Old Tony

I'd say that it's actually the welder giving you the asymmetric wave. For AC welding, the positive half is used to clean the oxidation off of the metal and it's actually heated/melted in the negative half and it's not 50/50. Because that's not really an industrial grade welder (not an insult), when they write the controller software, they probably use the same general wave form for AC and DC which is why you're seeing the positive half kinda chopped off. When they say "balanced", it'd be "balanced" for a non-ferrous metal and your adjustments are from that base wave form.

@ounben

6 жыл бұрын

I've scrolled the whole comment section to see, if somebody wrote the right answer. And you got it absolutely right. Nobody wants a symetric WIG welder ... greetings from switzerland

Put the clamp meter on the ground wire "the other" way. If the machine side exits the "front" of the meter, flip the meter so that the machine side is on the "back" or the meter. If that's the actual current waveform, the new waveform will be exactly the negative of what you saw before.

@kevinhoober1863

6 жыл бұрын

My thoughts....exactly

I agree with BMR Studio trying it on a lightbulb. A water heater would be even better, they are around 1.8-2kW. Heaters and lightbulbs are fine since they are resistive loads, and have no reactive component. The current probe claims having a Hall sensor, so it can measure DC offset too. Try it with a DC circuit, like car battery and a heater. If both work fine the probe should be reliably measuring both AC and DC. This probe is capable up to 65A, but only in 1mV/100mA setting. If your switch stays at 1mV/10mA it can do only 10A! Depending on the core it might measure still OK since they usually have a big headroom, but if the manufacturer pushed the datasheet declaration to its limit the core might saturate and display lower values. There are sometimes protective Zenner diodes inside and they clip higher voltages. Try and switch your probe to 1mV/100mA and your scope to 100mV/div resolution. As for scope resetting my first thought was the same as yours - grounding. If you are running the scope already from battery, it is certainly not the grounding. Try and keep the scope and the probe wire away both from the thick metal table, the welder stick and the ground clamp. All they create a loop where huge currents are flowing and coupling to your scope if near enough. If you have a spare milling machine around you might use it as RF shield between the scope and the welding equipment :)

Never ran into this before but if I question my equipment (I'm an electrical contractor) I use it on known source, for instance, if my amp probe isn't showing RMS, I'll simply connect to a wall wart or a circuit that I can add in a known amperage item to see if the clamp reads as correct.

Hi Tony. Can you turn the current probe around to measure the inverted current to check the the current probe is clipping on the ‘negative’ (below zero) side? If this wave form on the ‘scope looks ok on the ‘positive’ (above zero) then I would suspect the current probe is saturating on the negative measurement. Hope hope that makes sense to you.. John (retired electronics and electrical engineer from UK)

It kind of makes sense that they would want to add a negative bias to the waveform. If you had a positive bias (i.e., negatively charged tip), the positive ions from the rod/metal would gradually deposit onto the tip.

It is my understanding that AC/DC Current Clamps are non-migratory, be it African or European.

I have no idea what your problem might be, or if there even is a problem. I do however know that I like your videos big time, and it's fun being led in on your daily troubles like this, despite not fully understanding everything.

This video every few days is amazing.... I could get used to this.

A few things that may influence the result: - Check if the output of the clamp is the same even if You put it on backwards. It SHOULD by symmetrical, but it is not always the case. Flip it. - Check the current while not welding on the table, but on an isnulated piece of metal. Just a wild guess that it might change something..... - Try buying a LEM transducer rated for higher amps for current measurement. From my experience, the current measuring clamps from oscilloscopes are often no good.

Hi Tony, to test if the Clamp is clamping the waveform - pun not intended - try putting it the other way round on the cable, then, does the Waveform flip on the scope? Not using wall power with the HF-start is probably smart. For even better results try putting the scope on some styrofoam, and coiling up the leftover clamp-cable near the scope on the metal table.

Hi Tony- I have spent a good deal of time researching welders to fix my own SRC and IGBT systems- some of what I learned is straight forward stuff that you probably already know but might not be applying. All those that weld aluminium ( yes I said it) know that the reverse current portion is to clean the weld area of oxidization so more negative results in more “cleaning” action. Well in that case we could just reverse the leads of a DC welder and do it on the cheap-no?- Well, no.... as it happens there is more heat transfer in one direction than the other and therein lies the probable answer . For a truly balanced output you have to have an asymmetric wave as one direction gives more heat than the other . The likelihood of the scope or welder being off is low- My suggestion is it is to level out the power differential. Thanks for the videos- always entertaining :)

Fun fact, I just toasted my oscilloscope 5 min ago...AGAIN SIR YOU ARE THE BEST KZreadR ALIVE

The current clamp is acting as a secondary of a transformer (with the ground cable of the welder as a "single turn" primary). It is the induced current in the secondary, via a resistor or load (within the clamp) that produces a voltage which is what the scope sees - the scope has NO need for a ground as it has a potential difference to look at via the clamp and its load resistor. Square waves are a pretty harsh signal for a coil so you may want to first try experimenting with the clamp on, say, a 110v live conductor (with mains being a sine wave) into a small load - perhaps a lamp or such. You'll need *something* where the live and neutral conductors are split (rather than within one overall twin cable) - this will give you a verification that the clamp meter is balanced. The other thought is the welder being grounded at its supply connection *and* at the welding clamp negative via the bench - what about ensuring that the negative has no ground loops by isolating the clamp and "weld to" material from the bench. Another thought is to reduce the frequency and amplitude on the welder to as low as possible.

Hey Tony, First i would rule out a dc offset generated by your Scope. To do that check if the coupling is set to AC. Second thing you can do is build a one turn CT with your probe to rule out a fault of the Clamp-On-Meter. Just connect the negative Probe-Clip to the tip of the probe whilst looping it around the Torche's cable. The output signal will be weak, but 30A should be plenty to be yet visible. At last, your scope is not really trust worthy. Make shure the probe is compensated properly best would be a cross check with an other scope. The strange behaviour you saw with the menues and ''u-disk'' is caused by the hf stuff coupling in places in your scope where it really shouldn't. Hope that helps!

I know this link is old but I just got into your channel...very funny and very cool...when I grow up I want to be this old tony, or maybe a reasonable facsimile. Anyway, I digress, Your viewers who suggested turning the clamp around are correct in so much as it will tell you the clamp is truly showing you what's there but I believe the wave form is always going to show a reduced current on the positive half as this is the "cleaning" portion of the AC cycle in TIG and as such does not require the same amount of current as the actual "weld" portion of the cycle even though you are welding in both portions of the wave. This is a great example of why you have to use the "AC" form of TIG to weld Aluminum and not ferrous Steels. Aluminum requires the "cleaning" action to deal with the material changes and oxides developed as the Aluminum liquefies. Great Videos!

Good job on narrowing down the problem with the power of pure reason 👍 My money is on the clamp, too. In the video, the clamp is set to 1mV/10mA. The CC-65 datasheet says that for this setting, the "efficient measure range" is only up to 2A. So I'd expect waveform distortions for any output signal exceeding 2A/200mV. => Try changing the range to 1mV/100mA and seewhat happens. If you want to go down the rabbit hole, build a 50Hz variable current source (e.g. variac + resistor; rate for 1A; wind xy turns of wire in your probe to simulate xyA) and calibrate against your multimeter. The asymetric waveform is caused by the probe, if you clamp it on in the inverse direction of current flow, but the output signal does not invert. Also, the claimed bandwidth is only 20kHz. The power semiconductor switching frequency of modern welders should already be in that order (2...20kHz). If you want to visualise the actual waveform, your measurement devices should have about two orders of magnitude higher bandwidth. A clamp with 1...10MHz would be ideal. Below that, you have to assume that your probe constitutes a lowpass filter of unknown characteristic.

Your pixies are angry

I'm sure people have covered everything else (i.e. ground ref and / or shitty clamp on meter) so I'm going in with a real long shot. You're measuring the "return" current in the ground lead, now your big metal weld table could (read "probably is") providing an alternative path to ground, so you're not measuring the full current coming into the torch. One would assume that the leakage would be symmetrical either side of the waveform and wouldn't really matter for the measurements you're taking, however metal oxide layers can act as ersatz rectifiers, in which case the welding table could be providing a path to ground that conducts preferentially on one half wave compared to the other.

You can check the "balance" of your scope setup by looking at the current profile of a fully resistive load. A incandescent light bulb works great here, or a water heater.

I was taught in collage that the in ac tig welding one half of the wave was "cleaning" and was less effective at fusing the metal together. Just like the dc electrode positive/negitive positive being power flowing from the material to the electrode causing the surface impurities to evaporate but not heat the material and heat the electrode more vs negative being current flowing from the electrode into the work piece pushing surface impurities in possibly but going deeper with more heat into the material and less in the electrode. As AC is just the swapping of those two "phases" it would make sense to off set the "zero point" to make the phases equaly powerful to the average person

@SuperDeeredriver

6 жыл бұрын

But I just went to collage and relaying info that I learned anyone who actually welds for a living is happy to correct me, not saying my memory is spot on

The HF start is needed to establish an ionized electrical conducting path in the air. As the breakdown voltage of air is ~25 kV/cm in the HF starting phase your welder acts like a radio transmitter with voltage spikes of ~50 kV (from 2 cm). The torch and the cables are the antenna. This is what your scope don't like. Do the manual of the welder say something about the HF start? Like don't try to strike an arc from a distance bigger than 1 cm? The Hantek CC65 current clamp has a limit of 65 Amperes. The peak of a sinusoidal current of 65A is ~92A (65 * sqrt 2). It is ~184A peak-to-peak. According to the manual of the current clamp currents bigger than 2A should measured in 1mV/100mA setting. As others say, it may worth to try to clamp the current clamp backwards (rotated by 180°). -If the oscillogram turns upside down (mirrored horizontally), the measurement setup is faulty. If it remains the same then that is the actual waveform.- Sorry, I was wrong. This way the problem cannot be determined. But another idea. Your workbench is steel and though it is electrically conductive it's not the best conductor. You could use it as a shunt resistor.

To check te probe, you can try and flip it so that the current goes the other way around (refering to the positive current) through the probe. If it is indeed the welder, you should see your wave form flipped on the scope. My money is on the probe/scope being wonky though.

The third and very real option that we really need to consider is that I have no idea what I'm talking about and can't positively contribute to this conversation. But after deep soul searching, I've decided that I am still a genius. But I won't offer a suggestion here for "reasons." Seriously, good luck with that. Looking forward to the followup.

"I know I owe you a follow-up to the gears video; I promise that'll be next" Ahem. Just kidding, any video of yours is interesting. Or funny. Or both.

That waveform is from the current probe correct? If you rotate the current probe on the wire do you get an exactly inverted signal? If not, definitely a setup problem. Have you tried DC mode and see if it agrees with the probe. Flip the probe to make sure it responds equally to each current direction. If that works I think your setup is clean and your asymmetric waveform welds better theory comes into play.

@jhudrlik

6 жыл бұрын

Pondering this a bit more I suspect the voltage waveforms are symmetric and the physics of the EN vs. EP makes the current asymetric.

@JGnLAU8OAWF6

6 жыл бұрын

But welder supply current, not voltage.

Have you talked with HTP?

Someone call Uncle Bumblefuck and tell him there's a problem with the Smelloscope

@kalijamies

6 жыл бұрын

My thoughts exactly!

@dattepo7534

6 жыл бұрын

TjDolHaus86 I commented mumblefuck lmao

@leocurious9919

6 жыл бұрын

As much as I like him, this is way out. He is not a electronics engineer. This is more like for kzread.info Or kzread.info/dron/2DjFE7Xf11URZqWBigcVOQ.html etc.

@ArmchairDeity

4 жыл бұрын

He’d just say say the magic pixies been drinking too much Moulson’s...

A year later, I know. Old Tektronix engineer here. The hantek is not an inductive pickup, it is a hall element that can measure DC as well. The clamp does *not* block DC. (Sorry old electrical engineer Deathlok67). You still probably don't have a good ground, but that should only be an issue between the probe and the oscilloscope. Anyway here is what I would do to test if you are getting good date: - Install a shunt in the ground return wire, and measure the voltage across that. Maybe just some length of heavy wire with 50mOhm resistance or so, or maybe just the ground wire itself. - The AC frequency is 100Hz, and the RF start is 27MHz (my guess), so to shunt out the RF, just install a capacitor in parallel over the shunt, and if you like, a choke in series with your meter. - Twist all your probe wires to make sure you do not pick up any RF or other inductive currents. The arc itself is also very noisy RF wise. - Now, First, just measure voltage across the shunt with a multimeter, and multiply with 20, and you have your amps. Measure AC and CD and RMS. - Now hook it up to your scope and see if the waveform matches the multimeter. Please tell me how it worked out.

I suspect they deliberately tune the inverter to produce more negative current for more heating with the same amount of overall power. It would also reduce the heating of the tungsten.

@mairmatt

6 жыл бұрын

carabela125 - That's what is in the back of my mind also. But I'm not sure at all.

@bjarnivalur6330

6 жыл бұрын

I find this to be highly likely. bump! to make it more visible.

@TheAmpair

6 жыл бұрын

Nah. Positive is for heating and compressors, negative is for cooling and vacuum pumps. AC is for traffic lights, televisions and fans (that blow as much as they suck). I fear your electricity provider is not supplying you with Atomic or coal fired power which is ideal for welders as it has been preheated. Solar is similar, but only intermittent at best as it stops at night or anytime you go indoors. Wind or hydro power is obviously inherently unsuited to anything as demanding of heat as welding. There's scope for research here.

@bjarnivalur6330

6 жыл бұрын

Shit, fam, I didn't think of it from that perspective [scratching chin emoji]

the manual shows symmetrical waveforms: www.usaweld.com/v/vspfiles/pdf/HTP_221_Manual.pdf try mounting the current clamp the other way around and see if the asymmetry follows if you are feeling adventurous take the lid of the welder and see if you can find the place it measures current, I'd also expect that to be before a filter that keeps the HF start out of the inverter

@theymg3

6 жыл бұрын

In page 11 of manual, they are showing current and hoe to adjust it En and Ep.

@fuzzy1dk

6 жыл бұрын

on page 10 they show asymmetric pulse width, on page 11 they show asymmetric currents

@ThisIsAVeryBadIdea

6 жыл бұрын

I found the Linux nerd! ...hello, friend!

@TheMrAwax

6 жыл бұрын

That’s is one nice piece of factual data severely lacking from most of the other comments

@nf4x

6 жыл бұрын

Do you folks really think that TOT would be asking about the ep+ and ep- voltages if he had twiddled the AC ep+ and ep- settings? It seems pretty unlikely to me.

This kept me awake,you're problem with the waves in your electric, in England we don't have this problem, our electric doesn't have waves, its dry..I was left an old ARC welder,from the 60s I think.was owned by British Rail,huge thing,really heavy,big dials,I tried to weld a vespa floor,blew a 2" hole in it...I digress..no waves, good ole dry English electric... Crazy Americans

Like a lot of people here are suggesting, I say turn the clamp around on the wire (so that it's on the same side of the wire, but upside down), if the waveform is still the same (clipping on the positive side), it's the clamp or the scope, if the waveform is inverted, it's the welder. I would also suggest looking into what the welder is doing in terms of voltage.

Maybe you could avoid the Hantek clamp, just put a single loop of wire around the ground and measure that. It wouldn’t be calibrated but it should show asymmetry.

@LesNewell

6 жыл бұрын

Don't do that. First of all a coil of wire will have a very poor low frequency response and Tony is trying to measure a DC offset. Secondly you could induce very high voltages in that coil, especially when the HF is on. There is a pretty high risk of killing the scope with this setup.

@BenMitro

6 жыл бұрын

Bill suggested a "a single loop of wire" - very unlikely to kill anything, even with 200A + peaks and inductance of a single loop is very low, so HF response shouldn't be affected overly.

@KallePihlajasaari

6 жыл бұрын

Not going to work. A loop of wire around the ground cable will only pick up stray electrostatic leakage. If you want to make a current pick-up the primary and secondary wires have to travel in the same direction. Passing the ground lead through a toroidal core and placing a 60 turns of wire for the pick-up will give you a 60:1 current transformer, do not forget your 1A rated burden resistor of you have a DIY Tesla coil instead. Same as the clamp meter but bigger.

Turning the probe around is a good suggestion. Also, where is the ground wire from the welder going? I would watch the scope screen while moving the two leads, and while changing the current.

At 4:40 freeze the frame where the welder is turned off. Then you can clearly see that the positive signal is clipped. (The positive flyback spike is truncated) So I would as others have suggested flip the clamp. If it still is the upper part of the screen that is clipped, the clamp is overdriven and not balanced. You can also try to measure the current in a 12V 100W halogen lamp. When it turns on it is almost a complete short so some of the inrush current should peg the clamp. If not, make 2 or more turns through the clamp to make it happen. Do this with clamp turned both ways to check they are mirrored images. The bulb will not have a HF start, which just may upset the clamp. Being a DC clamp, it is not a current transformer, but an electronic circuit. It is also a single pulse that is not rich in harmonics and very repeatable.Also try turning off the HF start and scratch start on a piece of heavy copper to not tax the EMC capabilities of your scope and probe. You did remember to zero the clamp and snap the jaws open/shut a few times?And no, do not ground clamp or scope as som write! Your thinking is right, a fully floating setup protects the instruments and does not introduce any ground related errors.

Your scope is picking up erratic EMI as serial or command communications, and it tells it to do all sorts of things, but the first thing you need to do is shield, make a faraday cage (not like a luxury resort condo, just a cage) to either put the scope and as much of the electronics into that you can, or put the welder and torch into. Then you may need to insulate your leads by using twisted pairs causing interference to cancel itself out, and lastly, you may need to add some ferrite chokes to the cables to help filter unwanted noise that still might make it through. Your probe and cables act as antennas that collect the EMI from the welder.

Did you try turning it off and on again?

@evilfish456

6 жыл бұрын

Bjarni Valur give this guy a medal

Tony, what happens when you turn the CC-65 clamp 180° to the cable and redo the measurement?

I vaguely recall you get most heat on the welders' negative side, so it would make sense that's where the most current is. The positive part is just to break the oxide layer. The more cleaning action you need, the more positive you need, but the hotter your torch gets. It's been a few years since I did the course though ;)

Perfect video's to watch when trying to sleep

If it's an AC current clamp, then an imbalanced waveform (=DC current) can saturate it making measurements wildly bogus. This value should be in the data sheet ("maximum DC saturation" in mine). If it's an AC clamp that goes below ~200 Hz (likely), then it almost certainly has an amplifier built-in (probably requires batteries). In that case, the amplifier can clip and give you the impression of a very clean square wave at lower frequencies (at higher freqs clipping tends to produce very noticeable artifacts in many amplifiers). Only the very expensive AC/DC clamps can take large asymmetries without giving you bogus measurements. As others suggested, flipping the clamp around and checking if the trace inverts as well is a good way to check for issues related to the clamp. I don't think it'd be a wise idea to try to measure the output voltage of the welder directly, not with a regular old x1/x10 probe. I'd only do that with (1) isolated scope (like you have) (2) a x100 probe which can take a beating (3) a scope I don't like very much :D

@leocurious9919

6 жыл бұрын

The direct measure of the output should be no problem as long as he lift-starts it and not with high-voltage stuff that exceeds the limits of the scope times.

According to the datasheet of your currentclamp, its just rated for 20A in the "1mV / 10mA" range which you showed on the screen. (the exactly 20A look very suspicious to me) So first i would switch it to "1mV / 100mA" to get the full range (maybe its clipping there). Also i can't really see what your settings on the oscilloscope are, but i guess it's set to 1mv/DIV like it's supposed to. My second suggestion would be to flip the clamp. so its showing inverted current (negativ is positive and vice versa). Just like everybody else is suggesting. Ground is in your setup absolutely not needed. Because the current probe just puts out an differential signal, which is not related to ground. I hope this helps. If not feel free to contact me.

Plug in and turn on some big-ish load such as a heater and put the clamp on the AC wire and check if what you see on the scope is shifted downwards too, if it isn't then most likely what you're seeing is correct, except, perhaps, if the current clamp is being overdriven by the welder because it can't handle so much current and the signal is being clipped, for some reason, assimetrically, when the current sense clamp is fed with out of range currents.

As we say on the Mod Squad we definitely owe you a solid !

I'd expect the top of the square wave to be as dirty as the bottom (there is some definite ringing in the old welder and overshoot in the new one, typical square wave fun). You may have something there with the clamp limiting the positive peaks. Do you have a different clamp to borrow or make(!!!!) one by just wrapping some wire around the ground wire with a limiting resistor(series)? Simple is always good.

@DCFusor

6 жыл бұрын

I twigged to the same symptom. That perfect flat top "jus ain't right". Even if the scope or the probe is on the slow side - there should be all sorts of fiddling around on the arc on-set.

@TheAmpair

6 жыл бұрын

The bottom of the square wave is below ground, of course it's dirty.

@DCFusor

6 жыл бұрын

! Tony in disguise, right after I make a comment about his superior humor?

To eliminate any error of the probe just use your ground connection of the welder to the table as current shunt. Just connect the ground of the oscilloscope as near as possible to the welder and the other tip to the ground clamp on the table. You can not measure the absolute value of the current with this method. But you will see the waveform which is the interesting part for you. Thats the way I would troubleshoot this problem. If it's a problem at all or if it's just as wanted.

@IanSlothieRolfe

6 жыл бұрын

This is absolutely the way to do it. The ground lead will make the voltage small so you wont damage anything, and youll be able to see the hf start. Try and keep the scope as far from the spark as possible to reduce interference from the spark.

You were just measuring a well know physical effect. I wanted to do this by myself, but I'm too lazy. Thanks for your effort to prove it! At the heated tungsten electrode it is easier for electrons to be emitted at the re-ignition when the electrode is a cathode. If the workpiece side (eg of aluminum) is poled as a cathode, the electron emission is very low. The re-ignition is harder and a higher field strength is needed and this creates a spike during arc ignition. The arc voltage after re-ignition is higher than reverse polarity, since the thermal emission of the weld pool is lower, resulting in a resulting DC component of the voltage.

Tony the pixies will escape from the circuit board directly through the scope plastic case. Put the scope on a plastic shelf that is several inches above the welding table or better yet a wooden bench nearby. If you have any clip-on ferrite common mode chokes in your junk box put them over the scope cables. You might flip the current probe around and invert the current waveform. This might show clipping if it is happening. The advanced course involves aluminum foil shielding of the scope case and possibly the current clamp, hats anyone… -Peter

Seeing as you can change your welder's waveform amplitude to a forced asymmetric square wave, why not try that and take another reading. This could provide you with a method of elimination for the issue, as if the waveform simple shifts higher then the reading setup is correct, but if the waveform is truncated further then your reading setup has some form of clipping/ clamping device.

Might be your clamp. Although you are measuring AC, you should check that the DC reading is zero with a multi-meter. From www.hantek.com/Product//CC/CC65_Manual.pdf "Before measuring the DC current, it is needed to turn the zero adjustment knobs on the clamp until the multimeter reads zero. That means pushing down the zero adjustment knobs on the clamp until the multimeter reads zero. Then let go hold of the zero adjustment knobs.When using a scope,set DC coupled mode to measure DC current. "

When I was training as a machinist, my welding instructor told me that different polarities distribute heat differently, one is 1/3 in the work 2/3 in the electrode, and vice versa. It's possible that these fancy inverter machines modulate the current so as to distribute heat in a more even, or otherwise more desirable, fashion.

To take any reading of a wave, sine or square, triangle whatever, you need to know where 0v is. Usually a scope has the black ground lead of it's probe connected to it's chassis and thus mains earth. The DUT (device under test) amplifier, welder ground or whatnot is also usually the same. therefore, the ground on the DUT and scope is the same. You may have a DC offset where your welders wave OP is sitting at 5v DC say. One way to check is to use your DMM on DC volts, one probe to welder ground, one probe to scope ground. If you get anything but 0v, you have an offset. You can use another channel on your scope, no IP to show where 0v really is. It's hard to be more precise here as I can't see your test setup.

You should run the scope in DC coupled mode, not AC. I could not see from the screen if you were doing so but it sure looked like it.

@JGnLAU8OAWF6

6 жыл бұрын

It's already in DC.

BAMM : operation manual: Application Notes: 2. In the case of DC current measurement, the current clamp can not possibly zero properly, due to the hysteresis effect. To eliminate this, open and close the jaws several times and then zero again. I think thats the Problem

@JGnLAU8OAWF6

6 жыл бұрын

He measures AC

Quick test, as many have mentioned, would be to flip the clamp and see if it still clips on the top half. Another thing to try might be moving the clamp up and down the wire, since that big steel table has the potential to add some hysteresis to your measurements if the clamp is too close, but that might be a longshot, I'd expect the modified side to be a ramp instead of a hard clip. The clamp is probably maxing out somewhere, either the hall effect sensor or the preamp.

My typical response is if it works leave it alone. You have a fantastic welding machine with features that most of us would love to try just one time. The arc characteristics of the HTP especially with the pulse features is a welder's dream. Take a newbie learning to weld and put him on a TEE plate with pulse after you demonstrate keeping the wire in the puddle and he will look like a PRO. You have a machine that does all the figuring for you, just drop your shield and go to town. Seeing the HF Arc shoot over an inch to the work table is amazing. Try using my Miller 330 A/BP. It's like brand new with maybe 100 hours on it but it's not forgiving. You need to be skilled and understand how to dial it in then apply your skills. This is old school and when you started out you didn't dare think of grabbing a up a lead to weld with until you mastered Oxygen/Acetylene gas welding first and that means you weld all your steel, cast iron, stainless, and Aluminum with a torch. I love your videos Tony and your sense of humor is hilarious, even my little grandsons like watching you cut steel with a chop of the hand. But sometimes you can't measure things and these new machines adjust for all parameters, arc length, voltage, amperage, rod angle, position and so forth. When you go to lite up and nothing happens then make a similar video and get the Electrical Engineers or experts to comment and we can all learn how to fix it.

What has this got to do with the gears Tony?

Hey Tony. A few suggestions from others. Test on a A/C line. Flip the clamp. Here's mine: what is the channel coupling setting? Channels can be coupled for AC or DC. AC coupling means the scope will filter out DC Even though you are looking at a square wave, it is not a perfect square wave so it may have a net DC offset. If it is AC coupled, try DC coupling. If it is DC coupling, I don't think you want to switch it to AC coupling because the scope will filter out the DC.

@JonathanMeyer1986

6 жыл бұрын

wgm4321 this, it's an issue about AC or DC coupling, try the DC coupling to see the real value, AC coupling will shift the 0 line that both sides are equal, with the not perfectly symmetrical waveform that translates into an offset

@wgm-en2gx

6 жыл бұрын

Fixed some typos and here is a little more info. AC coupling removes the DC so the area above the 0V is the same as the area under the 0V. In the case of a 50% duty cycle of a clean pure symmetrical square or sine wave, the area above will be the same as the area below. However, if the duty cycle is not 50% or if the wave form isn't truly symmetrical, AC couple will cause the wave form to appear higher or lower until the area above equals the area below. Pretty sure you want DC coupling if you want to see if the welder is putting out a DC component.

@wgm-en2gx

6 жыл бұрын

See page 23 akizukidenshi.com/download/ds/owon/SDS6062.pdf However, After watching a second time, it might not be coupling problem

@JonathanMeyer1986

6 жыл бұрын

wgm4321 interesting.. what makes you be unsure?

@wgm-en2gx

6 жыл бұрын

Hey Jonathon, I looked at the video again and the wave forms don't seem symmetrical around zero with respect to area. It could be that the zero isn't set or some inversion option is on. Haven't used a scope for some time so I can't remember the exact impact of AC vs DC coupling. However, I'm pretty sure AC coupling will make the wave form symmetrical around the zero with respect to area and that doesn't seem to be what is happening here.

The simple way to figure out if it is a setup issue is to try the clamp both ways on the same cable. If the top of the waveform is always flat, they the clamp meter is limiting. Also make sure it’s in the high range, at 6:29 it’s seen as in the low range which is only rated at 20A.

@RichChant

6 жыл бұрын

Also at 6:50 when you turn the welder off of flashes up with a symmetrical waveform briefly, albeit with the spikes on the positive side clipped at 20A. Try a test at less than 20A.

I guess there's something to be said for the old analog scope like a Tektronix 465. Here's an idea. Remember how the cathode on a vacuum tube works. The heater heats the cathode to assist it in releasing electrons that make their way to the plate. The plate can get pretty hot with all these electrons bombarding it, same as the base metal with electrode negative. The tungsten gets quite hot on it's own, so it is always going to be a good source of electrons, but when the welder switches to electrode positive, the base metal is not such a good emitter, so the current flow, when the base metal is negative, is limited more the voltage between the negative metal and the positive tungsten than the constant current control in the welder. The welder can only put out so much voltage, so the low impedance condition of the electrode negative will deliver a much higher current than the higher impedance condition of electrode positive. One way to test this is to arc to another tungsten with about the same shape as the one in your torch. Presumably, at 50% duty cycle, they should be at similar temperatures. The ramblings of an old EE.

@gancicida

6 жыл бұрын

George Carlson Exactly !!!! I designed ac tig welder with 300A full bridge and arc rectification is responsible for assymetric waveform . I commanded 50:50 both currents,but output was assymetric.Look at my channel,you can see videos of that arc at 2 different frequencies.Also arc rectification makes ac tig low amperage arc starts and stabilisation much more difficult.Respect,old EE !!!

It's acting like your AC balance control is set to 75% or something. I would first eliminate measurement tool problems. Using your current probe, check one leg of something plugged into the wall outlet that's under load and see if the sine wave is also symmetrical. If not then it's your probe or scope. If it is symmetrical then it might be a clipping issue although it would be tough to get 65amps out of the wall without involving the fire department. Do you happen to have an old school transformer type stick welder that you could check the output? Have you tried flipping your current probe around? That should cause the wave form to be inverted which would tell us that it's probably the welder's design although I don't know why it would be that way. Good luck on finding the answer.

@MrBobcat975

6 жыл бұрын

You're correct. I was thinking asymmetric ac which changes one leg's amplitude without changing the duty cycle or frequency. So you can still have an AC balance of 50/50 however one leg is shorter than the other.

Trying to send the message again..You have a RFI problem. Un ground the scope to isolate it. The inverter is generating a lot of high energy RF on the power line And most likely the welding cables are transmitting a large signal through the air. Get a large toroid and put 6 to 10 wraps of the scope power cord through it. Do a similar thing with your scope probe cable. The starting voltage is likely kilo volts and maybe a different frequency. Check scope setting for that. Keep welding cables away from scope power cord and probe.

@emilerichier

6 жыл бұрын

The scope is battery powered

@Ballanux

6 жыл бұрын

He has an RFI problem for sure while the high frequency start, but that wouldn't create an stable measurement like he has after that, probably the clamp is either not correctly set to zero or has and internal problem and the positive range is not working properly. The best way is to test the current clamp to the full range with a known working current waveform

@MichelPASTOR

6 жыл бұрын

His scope is running on batteries it is not a ground or power line issue.

@watsok

6 жыл бұрын

Power cable is only one entry point. The RF radiating off the welding cables and the power cord are usually large amounts. A scope, especially if a battery powered one , are not designed for high energy commercial welding environments. The clear look on the scope, that was with the older welder and the setup with respect to sensor cabling may be different. Something is going on and resetting electronics is a sure sign of RF getting into them. Easy to kill equipment as the RF can be many times a higher voltage than the welder is set to supply. Also, magnetic fields enlarge as more current flows through a wire..Something a welder outputs in large amounts. Might be worth while to test at lower currents than are required to weld anything and if that level works ramp up the current.

@johnpossum556

6 жыл бұрын

I tend to agree with Kenneth's size-up of the situation and would add/refine a few points: a) Doesn't matter if the scope does not have a line power supply because RF (or in this case High Frequency noise) can still be the issue. The HF noise travels right through the air to the sensitive Hall Effect Sensor in the "current clamp" which is EMF based. My response to this observation would be to try to first shield the Hall Effect Sensor (also could not hurt to take a look at it's specs of operation paying particular attention to the HF aspects of the HES's "current clamp") and secondly to try a different way to pick up the signal. (What do welding companies use when/if they do this in house on their own machines?) At the moment the biggest thing that is bothering me is the video's reference to calling it a current waveform. I know this was probably just a slip of the tongue or temporary oversight by Tony but I see the scope IS set up for VOLTAGE measurements. _So it's improper to think that waveform is the current._ It's still the voltage even though a "current clamp" probe is being used. Current could still lead or lag that visible waveform. What the waveform's position below the line is telling me is that there is a clear negative voltage DC offset. IF this is accurate to what is happening at the welder's electrode I would posit that it is that way because this is some of the technique the design of the welding company installs to make it work as good as it does. Balancing it more towards the negative than the positive. So Tony, you're wrong when you think you are getting less positive current than negative current. I suppose it's because you don't have any formal electronics schooling. It's a voltage representation not a current representation as you point out at 1:22 in the video. When you adjust your thinking as such maybe whatever 'problem' you see will clear up. But knowing that I have to question if you are not making one of the simple errors of beginners to scopes. Like is your DC offset on the scope set wrong? I am not really sure I see any "problem". At 2:16 you clearly refer to the voltage amplitude as current amplitude. I think the erroneous way you are thinking about it is the problem. 6:30 There you are onto something. If you want more waveform data I would use a normal scope probe and tap it off at the pre-amp stage inside the welder itself. You have different stages of amplification inside the welder. Some like common emitter amplification will amplify voltage and current and invert the signal 180 degrees while others will amplify current. OK, so there is what you are doing wrong here. ha! Yes, you're learning electronics the hard way. In an _oversimplified_ way this might help you as this will most likely be in the welder's output driver: www.allaboutcircuits.com/textbook/semiconductors/chpt-4/common-collector-amplifier/ The important thing here is the set up of the amp. Your welder will most likely be using MOSFET technology to deliver all the Amperage. The key thing to remember is when a transistor is set up as a Common Collector Amplifier is that it will amplify the CURRENT and has a VERY LOW IMPEDANCE. You know why the current is important, it does the meat of the work of heating the metal. The low impedance is important because the maximum power transfer theorem( google this for more details ) says that the more you can match the impedance the more efficient and better POWER transfer you get. You see this the most often in Audio Amplifier specs where they might say you can get 100 Watts driven into a 8 ohm load. It's impedance matching for maximum power transfer. A welder is going to do the same thing and though its a total different beast to have a live arc I'd imagine the oversimplified circuit would illustrate the welding as almost 0 Ohms or a dead short. Very interesting to see the HF action of your welder. We are often taught that the higher the frequency of the signal the more it rides on the outter diameter of the conductor and at microwave freqs it rides off the conductor. So much so that waveguides are then used at those frequencies. It's fascinating to see the penetrating power it has when coming out of the torch.

-Put the scope on DC coupling -Put the clamp meter of 1mV/100mA -Open and close the jaws a few times to get rid of any hysteresis. -Connect the clamp meter -Zero the clamp meter -Disconnect the clamp meter -Start the arc -Connect clamp meter -Take measurement. If you have doubts about the result: -Use a known good set-up (clamp meter!, the scope is fine) -Test your setup with a known AC (grid) and DC source (battery charger)

I am pretty sure the symmetry problem is called "Hantek" and the "going wild" with HF issue is the combination of that and the "WON" plastic box pretending to be an oscilloscope. My Tektronix 465 (plugged into the socket next to the welder) with a passive AC current clamp shows the HF just fine. As always, you get what you pay for and using the right tool for the job is essential. ...OK, I hear the experts shouting "but that's not a storage scope..." ... no problem, in this case i'd use my Tek 464 which certainly does just fine because there is no computer junk in it as well. And not to forget, both got a proper metal case !

Does the waveform change when you turn the clamp around? Is the waveform symmetric when you clamp on some random ac line? And the classic, did you check the battery?

I think i know what is going on. I'm not a professional electronic engineer but i practiced on my scope a lot. I think that waveform is CORRECT. I have read the current clamp manual and you are not exceeding it's parameters (max freq at that amperage is ~20kHz) I know that this waveform is ugly but it is VERY difficult to create very clean and perfect square. Almost every welder uses forward transistor configuration: that means it can produce good clean waveform on one side but the other side it will be distorted (because it uses energy stored in transformer to make negative half). I doubt it in this case, that welder have a lot of settings so the only configuration for it is probably FULL BRIDGE, i couldn't find any informations about it internals unfortunately. I think welder is doing that. Don't worry in electronic world you only need to make something work. If it's ugly but it works then the job is done. BTW. If you want to make PROPER measurements you should put your scope AS FAR from welder as possible to reduce interference. (Maybe put something like sheet of metal between scope and welder too). Clamping the ground clamp is bad idea to get realistic measurements! Some current can go somewhere around that current clamp. You should probe the torch but i would be scared to not fry my scope. Last tip for you: You can disable HF start option in that welder. I have read it on official page: "Select the touch start mode for welding aluminum or any other material without High Frequency. This is important if your application takes you near sensitive electronic equipment, such as on site repairs in a hospital or repairing aluminum vehicles with on-board computers." I hope it helps you in your future projects. I hope you will see it. Greeting from self taught electronic, mechanical hobbyist.

I would try to make a really simple current shunt. Like a heavy piece of steel angle. Attach(weld) one end of it to whatever your ground is, then attach the probe ground near where the angle is attached to your ground and attach the sensing part of the probe around halfway up the angle(this would be adjusted to change the sensitivity). Then start an arc at the other end of the angle(first try starting the arc then attaching the probe)

The good thing is you took the correct first step by believing in the pixies in the first place.