Funny how I always hope there is another fault so we can watch Marc solve the whodunnit.

@medicman4444

Жыл бұрын

😂 damn it now I've got visions of Mark dressed like Sherlock Holmes and an R2D2 next to him

It's a "quantum" flip-flop. The act of observing it, stops it working 😉 😂

@ovalteen4404

Жыл бұрын

On a more serious note, it makes one wonder how the original designers got it to work, if they couldn't debug it in process.

@TheDimsml

Жыл бұрын

Given that it is some sort of neon tube logic, it might be affected by ambient light. So yeah, watching it work outside the case does not mean it works inside the case.

@Broken_Yugo

Жыл бұрын

​@@ovalteen4404I'd guess active probe.

@subwooferbone

Жыл бұрын

Sounds like a perfectly normal Heisenbug! Although in those cases, observing usually just moves the problem.

@audiodood

Жыл бұрын

Schrödinger’s flip flop

"it's neither flippin' nor floppin' " nothing doing anything, that moment got me 😂

Nice to see full, clear, and correct service notes. I've just finished calibrating something from a manual that was a) badly translated, b) missing vital information, c) wrong in a few places, and d) conflicted with the schematics. Fun times! And always a pleasure to watch someone else do it with less swearing, at least on camera :D

@lambertax

Жыл бұрын

Maybe a HP schematic is without error...

HP instruments always count, even when they don't count. :)

Magnificent job!! What challenge to build a counter with only transistors to 50 MHz. Early counter ICs years later would struggle at 10 Mhz. Wonder what a 400 Mhz transistor cost in those days. How many transistors are there? I am shocked to see tin-lead edge connectors. They are notoriously bad. The photocell decoder is quite a piece of work, They used radioactive neons (to get them to fire quickly in the dark) and photocells to make remarkable chopper amps in meters like the 410C. These might be radioactive too. Bulb had either green uranium glass envelopes or a bit of KR-85 (Krypton gas) to keep them partially ionized so they would fir quickly in the dark and the cold. I remember in the early days (my early days) using neons as legit multipurpose components. The high voltage was a pain. Another great video.

@Richardincancale

Жыл бұрын

Some gear using neons as memory had little incandescent lights next to them to encourage them to fire! Of course when the bulbs burnt out the counting stopped!

In case anyone is wondering how they made a frequency counter in the GHz range in the 60s (example for the hp 5255a plugin): They are using harmonic frequency multiplication to get 200MHz from the 10MHz reference. They use this 200Mhz Signal for harmonic downconverion of the input signal to a 1-200Mhz IF signal. That signal is then put thru a frequency divider and fed into the actual counter. Since they are using harmonic downconversion there are multiple possible input frequencies for a given counter counter display. You have to follow a procedure to "find" the correct frequency or you have to know the input frequency approximatly. HP used this principle for a long time after this unit was made, but they automated the "frequency finding" process with computer logic. I have a HP 5340A which does exactly that.

@jxh02

Жыл бұрын

Nice! Can you explain why neon, here? And "memory"? I don't quite get it.

@WolfmanDude

Жыл бұрын

​@@jxh02 I am not really familiar with the neon stuff since both my counters already use normal logic ICs for the counting. I think they are using neon lamps as binary memory elements for the counting process. The memory always holds the current count and also defines what will be the next number. They also use the emitted light for the binary-to-decimal conversion. Dont take my word for it, you can seach for "neon ring counter" online for better information on how it works.

@absalomdraconis

Жыл бұрын

​@@jxh02 : They probably used neon because it was both smaller and cheaper than the alternatives. This was long before cheap memory.

I wonder if the engineer who came up with the neon and photo resistor nixie decoder idea had to spend time convincing their peers or if everyone was just instantly like, that's genius. They sure had to make a fancy and weird little module to house it

@paulmoir4452

Жыл бұрын

I'm picturing a blackboard with pros and cons. On the con side, "We have to make a weird little module and house it." On the pro side, "We have to make a weird little module and house it."

@RS-ls7mm

Жыл бұрын

Used to work with brilliant engineers. It was high praise if someone said "good idea".

@absalomdraconis

Жыл бұрын

A lot of this stuff was "well, the easy ways won't work..."

@RickBaconsAdventures

Жыл бұрын

@@SPINNINGMYWHEELS777 I don't think you understand what I meant. Some singular person on the team had to have some kind of early idea of the concept and present it to the other engineers so they could all work together on completing and executing it. At SOME point, somebody had to share the idea they had to do something a funny way.

The engineering is fantastic; you can tell this thing oozes quality. Beautiful example some of the best technology of the era.

@martino6172

9 ай бұрын

It is impressive to me how well calibrated this device is, even after tinkering inside, replacing parts, and decades of oxidation. Timeless device

I don't know what is stranger. One is that I like watching you reason through this, or that I understand what you are doing. Thanks anyway.

these neon/photoresistor decoders are totally genius! I am constantly amazed at how they found ways to solve requirements at this time. Like the damn AGC that looks and feels like Doc Brown had to make an FPGA with 1950s tech. Ideas, totally located in the future, solved with current available tech and processes! And, again: the curve tracer ist such an amazingly cool tool!

It's very important to count your Nixies. Any discrepancy between Nixies in and Nixies out must be accounted for. Stray Nixies have the potential to cause serious damage, even to the extant of magic smoke egress. Common places for Nixies, after checking on and under the bench are behind the sofa and in the fridge. Once more, it cannot be overstated the importance of counting Nixies.

WOW! that nixie driver/decoder was sexy! Never seen something like that, but leave it to HP to come up with this! Congrats on getting it working and the troubleshooting! Someone really did a number on that thing in the past.

I love your repair-a-thon vids Marc. Explaining us the work of true magicians and restoring a gem to its glory while doing so

Talk about oozing quality... this beautiful gear was designed and built so well. What a wonderful piece of equipment to own.

Makes me want to drag out the not-working 5245L in my closet and get started fixing it up! I'm very jealous of Marc's collection of plug-ins, I must say!

Wow. This turned out to be more detective work than EE. Congrats that you got it working! :)

Great fix Marc. It shows the value of systematically starting at the input end and working through finding and fixing problems as you go. Cheers Malc.

@marcwolf60

Жыл бұрын

These should be textbook videos for upcoming techs re fault finding

GREAT OLD COUNTER. Repaired many of these during the 10 years I worked for HP. I have a 5345 and a 5340 in my equipment rack today.

These old nixie tube counters were a marvel. I ran across a similar one but it used the voltage drop of the nixie filament that was 'lit' in some sort of memory/ decoder fashion. Just the fact that a particular filament was on, 'steered' the next 'flip' of the thing in some fashion. I can't remember it all, but at the time I was just amazed how they designed it such that the display tubes actually took part in how the thing actually counted.

Your troubleshooting skills are amazing... and your video's are to.

If only that counter could talk. I'd sure like to know what happened to it. Your sleuthing keeps me at the edge of my seat. Congrats and thanks!

I just love seeing this old tech being brought to life again. Those engineers building a better mouse trap back then is interesting to see how it was accomplished. I'm the type of learner that seeing it done is much simpler than trying to understand from a book or text. Thanks for making these videos.

I used to work for HP from 1975 to 1985 in their repair center in Rio de Janeiro and later in São Paulo, Brazil. Worked on many of these 5245 counters. The weak point for us in Brazil were the Nixie Photo resistor decoder drivers, these often failed due to humidity eroding the tracks or resistor patterns laid on the ceramic backplane. Yep, once they had TTL BCD decoder chips for Nixies snd 7 segment LEDs, made these works of Electronic Art redundant, although those Nixie Counters with the high stability time base hung around for years in many Labs. Luckily HP made excellent manuals that still enables capable technicians/ engineers to this day repair them. Wonderful to follow your troubleshooting, brings back good memories.

Oh, what a beautiful decoder implementation!

Schrodinger flip-flops, if you probe it, it does'nt work XD

16:46 so cool to see the neon tubes :)

Marc, thx for taking the time and sharing your journey through the repair of this counter. Upon further consideration it begs the question as to how a single unit could have all these problems simultaneously. I'm wondering if this wasn't a hanger queen where the cards were swapped out of it to keep other units operating leaving this one with many problems. Eventually funding came along and these units, 5245s that is, were phased out for newer types of counters. Just my 2c worth. I love this level of troubleshooting. Employing block diagram guidance, signal tracing, crkt level troubleshooting, component failure verification, component replacement, then viola repaired section, followed with moving on to the next malfunctioning section. The best part of these units are that they don't use ICs with unobtainium, that is they don't exist any more. All the units that use ICs with unobtainium will be moving into the recycle bin long before these units will be unrepairable. Of course these units don't have Bluetooth, data analysis etc. internally. However, these units do have data ports that can be connected to data analysis units. Then you've got units that can do all the wizbang things the newbies can but for a long time into the future. Marc, again thx for sharing.

Incredible inventive minds of the 60’s… 😮

@projectartichoke

Жыл бұрын

No doubt about it! The neon lamp/photo resister stuff is pure genius, and the speed and sensitivity of that one discrete transistor flip-flop is incredible.

Such a pleasure to work on equipment with component-level schematics! Designed to be fixed.

I've had one or two of these older counters but never this one. I think I had one of the non-plugin versions but cannot recall the model number. I never did anything with it and eventually sold it to a friend. Great repair video, looking forward to the plug-in troubleshooting!

it's so impressive, as you said, it's like diagnosing an IC but with discrete components! :)

Now that's an excellent analysis and repair! I have an old Berkley Nixie counter/timer very similar- but not as old.

Fantastic video as always.... Thanks!

Looks as if working boards from this unit were swapped with bad boards to repair other units. Love the large visible neons of my 5245L. When this product was introduced in 1963, I was using an HP 524 counter with columns of neons for display.

@CuriousMarc

Жыл бұрын

I have come to the same conclusion, that this may be a board swap unit that got all the bad boards. The gate PCB isn’t even the same color and material as the others, apparently from a much earlier unit using phenolic boards.

@chefchaudard3580

Жыл бұрын

I did that kind of repair for an HP competitor in the 80's. It was very common for large companies to stuff all the faulty boards in the same device and send it back to the supplier for repair when there were no good board to salvage. We received counters, multimeters, and other devices which took hours to repair, going through every stage, from the input to the last one, with almost every one of them faulty. I remember that, for multimeters, we always forgot to test one or two functions. the faults were found during the final calibration process, usually at the final stages, with the guy in charge of it mad at us, as he had to fix the fault and redo the whole calibration again.

@CuriousMarc

Жыл бұрын

@@chefchaudard3580 Oh no! That's funny. The ancient version of the Amazon return scam :-)

Supreme troubleshooting skills have you Master Marc

This tech is amazing and creative. But so is the troubleshooting and repair! Excellent work.

There was a build your own counter project in the ELV magazine in the end 70's .... that damn thing drove me crazy !

You've truly made that song yours by now. Love it :)

love the methodical work through of the issues !

@jurjenbos228

Жыл бұрын

It's the only way that works with problems this complex.

Gotta love doing the computation *in* the blinkenlights

this is very cool.. looking forward to the next stage

I also have one with plugins. I had to repair the heater circuit in the TCXO master oscillator which has been calibrated to GPS and has not moved frequency after running for 10 years. This is a fine frequency counter. I appreciate your nice troubleshooting techniques.

That transistor testing thing is super cool

That is amazing! In all my experiments with photoresistors, they have always been cheap CdS cells, and all were painfully slow. (tenths of seconds!) I have always thought you needed photodiodes or phototransistors if speed is required. (Of course, 1960's HP would have special photoresistors, burnished by grey-bearded masters, washed with water from the river Styx.)

What a wonderful video. Both educational and just plane fun to watch. Great job Sir.

Great video and interesting process of finding failures. I'm in the middle of fixing my HP 5245L :)

What a beautiful instrument!

@projectartichoke

Жыл бұрын

The build quality and serviceability are truly outstanding. That stuff was built to last forever.

@dufflepod

Жыл бұрын

Mainly the ingenuity of the wholly analogue design, and the genius of using the neons with the photo resistors, it simply leaves me agape.

The reset line was a case of unplug it and plug it back in

@zebo-the-fat

Жыл бұрын

dirty connector, unplugging it fixed it?

Awesome video. What a master troubleshooter you are! You’ve certainly forgotten more than I’ve ever learned. I love seeing your masterful videos. Thanks for sharing.

Maybe a solder whisker causing a short, that melted away?

Nice progress Marc

Nice work my friend

Anything with swappable plugin slide chassis of that era remind me of the TOS Trek when Scottie replaced some burnt out stuff with moduels from a space station "they fit like a glove captain"

loved that red tracked circle!

Very interesting repair and a good demonstration of the use of a 4 channel scope. I think I have some of those nixie tube modules in a junk box somewhere.

I am absolutely clueless but addicted 😁👊👊👊👊

The king is the HP5248M. It counts up to 150 MHz, using the same technology, and has the better ("M") OCXO. Mine ages at

Nice work!!!!

C'est fantastique. Même si la technologie utilisant des photorésistances est super intelligente, la réparation finalement simple parce qu'utilisant des composants discrets, me rappelle tellement mes débuts d'ingénieur en électronique que c'en est émouvant. Merci pour ce partage qui une fois de plus est passionnant. 😊

LOVE IT! Thanks!

Une hécatombe chez les transitors. Félicitations pour cette réparation.

Very interesting thanks a lot marc.

It blows my mind that a garden variety dirt-cheap 2n2222 is a higher-quality replacement for the transistors that are in this ultra high performance counter. Nice job troubleshooting, and be thankful you have a decent set of schematics to work from!

I had one of these about 20 years ago, with the intention on using it for some sort of project. I ended up just giving it away with lots of other stuff in a basement purge. :)

looking for next part :D

I think that the short circuit on the reset line is caused by the electrochemical migration of the tin alloy, it create a tiny conductive filaments with low impedance that short low distance pins. Gerry

A 1960’s isolating opto-coupler! What a great way to control 115v.

It was a Schrödinger flip flop, was in both state until you was looking with the probe.

If you're having issues with 2N709's, think of the poor guys at the Computer Museum. Their CDC 6500 computer has about 70,000 of those! Perhaps surprisingly, they used to have it running for weeks at a time between failures.

Aha! Schrödinger's transistor 😊

I used to have one of those frequency counters a long time ago you see the one with the m and the dial control on it same one identical

According to one manual, the best substitute for the 2N708A is a 2N2222, but for the 2N708 (no 'A') it says the 2N2369, which is a common transistor used in high speed circuits such as video amps. It says the European "Pro Electron" substitute is the BSX20. The 2N2369 or 2N2369A are common and still available. I found that the ubiquitous 2N3904 will work well but it may have too high current gain. In one circuit I put a 470 ohm resistor in series with the base to reduce the current gain slightly. The 2N2369 is found in video amps, sometimes in old TVs or monitors. Thanks, Marc for keeping us curious!

HP - Quality Kit

You may have a look at Usagi Electric's video about an older HP freqency counter. "Restoring an HP 521A Frequency Counter". Looks like they recycled the basic design with the discharge lamps for the bcd couting. It uses tubes instead of trans, but the schematic looks very similar (at least to me).

Did you know that Hewlett Packard is an anagram for The Lewd Packrat...?

I don't know if you spotted it, but the 2nd LSD jumped from 1 to 7, then later back to 1, at around 12:00

@ferrumignis

Жыл бұрын

Marc mentions this behaviour a bit later.

Schrodinger's flip flop 😂.

I got one from an estate sale and too my surprise it worked, and also it reads that accurate. So I'm not sure anyone calibrated yours

@CuriousMarc

Жыл бұрын

Thanks, good info. You are right, I can't know for sure, it might just be very stable. But for the power supplies at least, they don't stay exact to two decimals for very long, so that had to have been calibrated very recently.

@clytle374

Жыл бұрын

@@CuriousMarc either way thanks for another great video. I wish I had time to watch all your stuff. I have these bookmarked in case I have to work on mine someday. Anyway, mine looks great with a 8640b sig gen on top of it. I have a 333 distortion analyzer I need to fix to add to the stack

I found with my 5245L that some photo-resistive decoders for the Nixies were not working reliably. This was due to there being considerable variation in the firing voltages of the neons which, I believe, was caused by neons being operated on DC for long periods. With DC operation the neon glow is only around one electrode and this electrode is damaged by ion bombardment over time, considerably increasing the firing voltage. I restored operation of the decoder by reversing the faulty neons so that the undamaged electrode became the active one, reducing the firing voltage. Hopefully, this will give a few more years of operation before I have to find some of the special "dark effect" neons to replace them.

Makes me wish I had kept my 5245 with the 18 GHz plug in, but I was young and foolish...

@ 11:20 - Would it be worth running a SPICE model of the original T-FF? I seem to recall that transistors used in a bistable had to have gains matched to a fine tolerance - is that also part of the problem?

I wonder if the person before had several of these units, and this was the one that donated it’s working components and received the failed repairs. And while they had the equipment to keep it calibrated, they didn’t have the expertise to keep it working

@CuriousMarc

Жыл бұрын

We don't know for sure, but it certainly feels like that is exactly what happened. HP units usually do not come with that many faults stacked up on top of each other. I know for sure the previous owner had several of these units, and this one was a freebee, so I betcha I didn't get the best one...

Possibly a bad solder joint creating noise on the reset line? That's a wild guess.

You make it look so damn easy! It isn't, at least for me anyway.

With the amount of reworked boards on this I wonder if this was an unit that was scavenged for spare parts and just collected the failed repair attempts from other units.

@ShainAndrews

Жыл бұрын

Unlikely...

Is there a modern equivalent to the instrument you use to test those transistors ?

@ulwur

Жыл бұрын

SMU's such as Keithly 2400.

Wizard repair on the reset line...

19:05 what?! Made it worse?! Been there, not at this level, scale, expertise, but in my little world in other areas (like repairs on plumbing 😸)

Hey is that a Roland Juno-60 in the background there?

@mudi2000a

10 ай бұрын

Yes, or Juno 6!! Was scrolling through the comments to see if anyone else spotted it ...

what about rebuilding that divider board using BFR960 transistors which work at 960Mhz

Beautiful machine, when ECL IC’s were just a twinkle in the eye

How much power does that use and could it have been a power surge that caused 8 failures?

Relay logic with light instead of relays. Much faster.

I swear that 98% of youtube electronics repairers assume that someone fiddling with a knob to get the calibration on a machine *just right* must be boring, should be cut out.... NO! NO NO NO! I wanna see it line up to all 0's in real-time as well! Watching everything come to beautiful alignment is basically electronics **rn to me... Second hand satisfaction. Like seeing a cake decorator do the final assembly step (cherry in just the right place on top) to get the perfect result, YES I want to see that last process. Surely I cannot be alone. On the reset line, maybe things just needed some more deoxit / wiggling around to get good connections?

Is it actually BCD ? My 5275A is 5x2 (biquinary), then the same neon/ldr decoders and 1-of-10 neons in a faceplate instead of nixies

@CuriousMarc

Жыл бұрын

This one is equipped with the 1224 BCD counters, the default. There were also options for 1248 BCD and biquinary.

Hey, it must be the World's first Quantum-counter… If you look at it, it stops working!

"its an IC in discrete form". Also called just a C ;)

Strange so many transistor failures, you would expect them to turn it off after the first failure. Someone must have used it as an accurate oscillator source until that stopped as well.

👏👏👏👏👍