This circuit almost certainly has an op amp to multiply the milli volt signal from the shunt up to a multi volt signal to be digitized by the microprocessor. I think they have the gain setting resistor for this amplifier at ten times the required value. Lowering the shunt resistor value just multiplies all the errors.

@judd_s5643

10 ай бұрын

I would have zeroed in on the op amp circuit before I’d mess with the physical shunt. Another area is the A/D Reference

@cambridgemart2075

10 ай бұрын

@@judd_s5643 Probably the same reference for the voltage channel though, so I would suspect it's not that.

@simonmasters3295

8 ай бұрын

Considering it is x10 (precisely) the actual reading I believe it most likely to be a software error, or at least a design error. You should run more tests at different Amps through and get more readings before changing anything... imho

@simonmasters3295

8 ай бұрын

@@cambridgemart2075 what vrf is he, or rather the device, using?

@jpopelish

8 ай бұрын

@@simonmasters3295 Given that current measurements are often only millivolt signals from a very low resistance shunt (a short piece of heavy wire, in this case) amplification is almost certainly needed before there is enough voltage to be digitized. And fixed resistors happen to come in precisely decade steps to set amplifier gain. This is a gain resistor error.

After snipping the shunt in 3 parts, you could have soldered the 3 parts in parallel. This would create a factor 9 drop in resistence. After that you could have trimmed it to factor 10, and saved the puny 10K resistor :)

@deslomeslager

10 ай бұрын

Plus 1, I had exactly the same idea. He stil has a few to go, and I know he kept the pieces so he still can fix it.

I bought 2 units after 6 months of your video and they already fixed the firmware they had to change the formula ;) thanks for your videos Julian

I am seeing suggestions in the comments I don't understand. Can you prove or disprove them in another video? I find this kind of thing fascinating because the (best) answer, it seems, is never obvious to me and when I "get it" I go, "Ooooh, now I see". I just love those moments. Thanks Julian for another great video!! 🤩🤔☺

Good fix. Of course the result is that you are using almost none of the ADC, but never mind - it works. It really needs a software update. That would drop that zero error off the bottom of the scale too.

I'm thinking that since it's fairly accurate but 10x out, maybe just altering it's resistance that goes to the micro controller input so that it accounts for that 10x difference at that point, rather than trying to shorten the shunt? I could be wrong but if it's reading 9.210 amps for an input of 0.921 amps, if you changed that 10k resistor for maybe a 100k resistor it would account for what might be this product's ultimate flaw, which was the wrong resistor being installed for the amp detection circuit? Then you also keep the soft-accuracy adjustment. It may not be 100k but perhaps some experimentation would reveal a much easier fix (swapping out a resistor) than cutting down the shunt.

Interesting, my fix was just using an automotive fuse and added it to the base of the shunt resistor, then trim it in some places, precision is really good!

I got a different revision of this panel and it is measuring the correct values, it looks exactly the same but with a different PCB Layout and has the mark: Deek-Robot Shield Model BL-02, I get the feeling one may be a copy of the other one. I wonder if mine could be an older one than yours since I got mine for at least more than a year now. My shunt has pressed sides on it so that it will be in the same position when you put it in as well, yours are completely not having any determined marks or dimples where it should be placed either which makes calibration even harder

That is such a cool looking panel mount meter. What a shame it is off by a factor of 10. 😥😢

Great work.

SECRET SOFTWARE CALIBRATION MODE: If you power on the device while connecting the two open plated through holes that are beside the large chip, it will enter a calibration mode. I found that tack soldering a momentary push switch made this go easy. Once "0.00V 0.00A CAL" shows up on the screen, reconnect the holes (or click the switch added) to click through the sequence. The display will show 100V. Then after pressing again, it will flash 100V, then show 5.00A. After another press, it flashes 5.00A then goes back into the reading mode. I successfully fixed my unit which had the 10x problem, not needing any other mods; just need the right power supplies to deliver the required inputs.

@pault6533

4 ай бұрын

PROCEDURE (after adding push-button): Have 100 VDC and 5A power supplies ready before starting, as well as a separate 12VDC supply for the unit only. Before connecting, turn device's front Volt and Amp potentiometer approximately to middle position. Connect only THIN red and THIN black to 12V supply while holding button. "0.00V 0.00A CAL" mode starts showing, tap button to set zero offset, then "100V" solid appears on screen. Hook 100VDC to THIN white (or yellow) wire and THIN black wire (shared with 12V supply). Press button, 100V flashes, then displays 5.00A solid. Disconnect both 100V wires, leaving only 12V attached. Connect 5A to THICK red and THICK black terminals only. Press button. 5A flashes, then goes into standard measurement mode. Disconnect 5A. Now use the front panel for fine-tuning closer to the values you want to use the device for, such as 5.00V and 1.00A.

I would just run a wire from the sense point to some point in the middle of the shunt wire. Moving the solderering point you could trim the value to some near the real. Then adjust with the pot. I will watch this item in case they sell them for cheap. 😁 Edit: today it costs 1,79€

There's a small SMT resistor next to the A CAL potentiometer, maybe it would be sufficient to change it, instead of fiddling with the shunt?

You fixed it!! Woo Hoo!! 🥰🥳

I wonder if the design should have had the 10k between the trimpot and the shunt instead of at the bottom end.

A What meter?

Could maybe have used one of the offcut legs of manganin in parallel with your new shunt.

@brucepickess8097

10 ай бұрын

Good comment ,just what I was thinking.👍😏

Love this kind of video. Its educational and fascinating troubleshooting

It is possible that the amplification factor of the shunt voltage amplifier can be adjusted using a different resistor circuit.

@JulianIlett

10 ай бұрын

Maybe. I'd also like to find a zero offset resistor.

@karlbesser1696

10 ай бұрын

​@@JulianIlettWho knows if such a thing even exists.

Informative

You could probably take one of the off cuts and solder it across the gap on the other side that seems l Ike it would be a real short one

nice that you found a solution to the issue, but at the same time, it means the seller has gotten away with selling items that do not work as they should. This does not teach the sellers in China not to sell defective goods.. This was menat to be a brand new item, it shouldn't need modifying by the end user to make it work right.

I notice yours do not have a model number on the pcb...others on youtube have BL-02 which seem to work ,,,,Aliexpress has BL-01 maybe dodgy...they show 48mV drop across the shunt ( which appears to be different) @ 10A.......... 11:48 .......kzread.info/dash/bejne/ZHquxceymKqZZ7A.html

Think it's likely just a sorting error in production for a run and thinking the 10k and 1k resistors are swapped? Course there's no design diagram shown here so it's a guess but likely, I mean the Chinese aren't that stupid to make a bodgy design like that so I bet this is the case. Perhaps look for a design errata somewhere on the manufacturer's site?

@pault6533

10 ай бұрын

I have good and bad units. The resistors are identical and measure correct to their markings

@pault6533

4 ай бұрын

Update: it’s more likely the calibration current was set to 0.5A instead of 5.0A during factory calibration. See my other post about the software calibration procedure for this device.

Awesome, that's proper electronics that is ! analog FTW....cheers.

Do not place the shunt too close to the PCB, the shunt can become hot and scorch the PCB.

You could rewire the tracks going to the decimal points and move them over by one?

@barrieshepherd7694

10 ай бұрын

The thing goes into HHH (Overload) above 1 amp so moving the decimal place won't help.

@____________________________.x

10 ай бұрын

@@barrieshepherd7694 ah, good point

Try putting voltage on shunt holes after it was removed to see what voltage needed for 1 amp. Then put one side of shunt in and solder a wire to the location needed on the shunt wire. Match that voltage to location solder point. That should give your linearity that you wanted.

Just drop part of the snipped shunt in parallel.

Can you parallel the shunt of cuts with the shortened shunt?

@JulianIlett

10 ай бұрын

I guess so, but how would you do that neatly?

With the mod, how does it do with volts and watts? 🤔

@JulianIlett

10 ай бұрын

That's my next project ;)

If you are going to be using quite a few of these, it's probebly worth ordering some 10K, or even 20K, pots and 5K resistors, (instead of just shorting to ground). That way you should be able to keep the accuracy. (?)

could have just modified the divider network, instead of chopping the shunt.

".002 * 10^-3"... Nice.

I would unfortunately vote this error as unacceptable. But your ambitions to fix the problem was well worth watching.

A resistor somewhere is 10x incorrect.

@pault6533

10 ай бұрын

I have an older unit that reports the correct current an one I received today which has the 10x problem. I’ve been all over the unit inspecting the resistors and the shunt and the pots visually and with a DMM. Everything is (to my surprise) identical. I am convinced there is a problem in the little black square chip’s programming.

@johnshaw359

10 ай бұрын

@@pault6533 Maybe try a 10x times to an appropriate op amp gain or hack ln a pot to see if it can be effective somewhere.

@pault6533

3 ай бұрын

@@johnshaw359The factory did the calibration wrong during a production run. I saw the procedure video online. They put the preassembled board in a plexiglas fixture which is wired to a manually set power supply and follow a CAL procedure before putting it in the case. Evidently somebody entered 0.5A instead of 5.0A during setup. I found the CAL test points which enabled me to re-do the factory process on my bench. Works fine now.

They must be cheap to buy so many.

Wow! Almost as if you're shopping from the same seller. Just two days ago I received the very same unit. I'm having some troubles with mine. From the start it showed 1.00A and incorrect values for what, no matter what I did. While showing the first 1, it also seems to give 10 times too high. Haven't finished watching yet... Looking forward to see the rest of the video!

@JulianIlett

10 ай бұрын

It's fixable - kinda :)

@pault6533

3 ай бұрын

Take a look at the secret software calibration I discovered. If you have a 100V and 5A power source you can correct the factory errors.