Fix A Power Supply! Diagnosis Repair And Restoration. 

I absolutely must schedule myself a class time after work to learn from you before my capacitance to absorb wanted knowledge shorts out due to my age. Sure would nice to do something of value before I kick the bucket.

Is it possible to know a bit of your background work/studies etc... ??

@Santa Clause Try watching Electroboom :D

Who was the Mfg. and what was Model # of this Power Supply?

Paul, your skillset and motivation to deliver quality work (and content!) never ceases to amaze me. I absolutely love it. Proud to be a Patreon of you 🙂

Got a chuckle from that too.

Same here

Absolutely no way I could know as much as he. Probably ever.

It freaked me out for a moment, I was convinced it was an editing trick until I spotted the 'Remote Display' text on the readout

That would be a fun trick for an instructor to play on his students, like the magician's trick of sawing the lady in half, this could be sawing the Fluke meter in half.

@@vancouverman4313 We have one of these at work (and a pile of 179s) - at first I thought it was really neat, but I came to despise it, because the AA batteries in the head get drained *very* quickly (in contrast, the 179s simply run off a 9V battery that lasts forever), even when it's not disconnected from the body :/

It may have received a static discharge? As mentioned in the video, I have no idea why they both failed. OH... and I did a video about repairing the board in my fridge.

Thanks for your kind comment Nenad!

Well said, sir!

I couldn't agree more. I watch because it is so impressive to see such skill. A totally absorbing experience every time.

I fall asleep every night listening to his videos. There’s nothing better to dream about than old capacitors and vacuum tubes.

@@michaelh2899 Slightly creepy but saved by it's eccentricity 😄

I thought it was a double exposure or video overlay 😁

That was my thought exactly, cool where the meter face separates from the body of the meter

@@randyelliott9495 A guy I work with has one. dropped it on the floor! $$$$

Same here. I'd never seen a removable display like that before. Cool stuff!

That looked like some amazing CGI. :)

Ha yeh he confused me before on another vid with his spooky headless meter

Still Others: I'm going to order a couple of the original ICs from eBay rather than redesign a power supply that has worked well for 50 years.

@@patlaw53 Exactly, 3 mins on ebay shows it's not exactly an unobtanium component nor what I would consider expensive compared to building an entirely new power supply

@@patlaw53 most of the ebay stuff is either obtained from e-waste or chinese fake parts. I will never trust ebay on expensive components. I would rather pay extra 2 or 3$ on digikey , mouser or lcsc. & A newer circuit with cheaper & more accurate components is way better than waiting 1month to receive fake chips or e-waste chips that will go bad after few months.

@@KuntalGhosh I buy dirt cheap electronic components from China all the time and they work perfectly fine and last a long time, even the non-original ICs. When you buy from those American stores, where do you think those components were manufactured? The only exception is electrolytic capacitors where I must be sure I get original Japanese brand ones. In this case for an old linear bench power supply that is not going to power critical or life saving equipment, I see no issue in using cheap non-original Chinese ICs to get it working. Chances are it was better manufactured than the original decades ago and may last longer.

@@metricmine nope. The fakes r most of the time some different ic that got rebranded.. so the ic never performs to the spec , most of the times they pull it out of old electronics and they die soon .. i have had problems with many AliExpress chips. I regularly order power supply control chips and catv booster chip as i have a buisness of catv and isp & i repair most of the stuff on my own to save a lot of money and power supplies regularly die due to lightning or other environmental damages and using cheap knockoff chips has caused me a lot of trouble and more down time than what i saved by going for the cheap route! ..

it inspired my avatar/business logo...

I used TO-3's as cufflinks once :)

@@dirk-janvandenbroek9947 that is actually not a bad idea. I think I might still have some new 2N3055, military spec somewhere...

I don't like the TO-3, they're difficult to mount and insulate, especially on PCBs. I like TO-247 or TO-264 much more, but I'm also a MOSFET and switching power supply guy. :)

@@stormeagle28 yup, but linear supplies have their niche, especially in ultra-low noise applications. TO3 mounting and insulation is a pain, but it's a very worn road, there's tons of documentation, I just find there's no need for through-hole components in power electronics these days and flying wires all over the place just create unwanted aerials also adding to the capacitance and inductance of circuitry

Thanks for the laugh Jim! :^)

@@MrCarlsonsLab "Try this AT YOUR OWN RISK - take care"

And now you are a family of wealthy surgeons all thanks to your cousin (who is either an electrician or in jail).

We were so poor that when momma let us eat from paper plates, we had to erase the dishes after supper.

Your nose and that of the patient lit up simultaneously!

Thanks for your kind comment and feedback Jim!

I always give a thumbs up before I start watching.

He plugged the thing into the wall the wrong way. ;)

@Dillon Kluksdahl looking at the IC I'd imagine it's the same problem that's killed a few discrete linear bench supplies, zapped Zeners... usually caused by a power spike from the mains, I'd install a surge arrester and a filter in line with it

Maybe they didn't die at the same time.... maybe one died and the previous owner used the other side until it died. Who knows. An interesting thing to me is that this supply doesn't have the range switching setup I've seen in other units with linear regulators. In order to keep the power dissipation under control at low output voltages, they would switch to a different tap on the power transformer. This lowered Vin and kept the power dissipation reasonable, plus it helps keep the output transistors away from secondary breakdown area on the SOA curve. I wonder what would happen if you set Vout to say 3.3 V, current limit to maximum, and then shorted the output.

My question too! I was looking to see if someone else asked the same question. :)

@@davidhunt240 I'm surprised he didn't do any failure analysis of the IC since he had the schematic of the IC.

agreed, I think this was staged to some degree

Mr carlson's extream knowledge and experience already told him where the problem was, he just needed to do some test to prove his theorem was correct. Let's face it, he knows a whole heap more than you or I could ever know

Mr Carlson has been fixing power supplies for a long time and I suspect he made his living for awhile doing so. A linear power supply like that one is laughably simple compared to the switching supplies he was fixing or rebuilding routinely. It’s a pretty good bet that he thought the IC was a prime suspect before he opened the case. This is a guy who understands electronics at a very intuitive level. It is his superpower. Thank God he uses it for good and not evil ;-)

@@jgeorge6081 I agree mr Carlson has great electronic knowledge. But he 's still human, you know..:-)

I feel I also missed something here too, not quite understanding the linkage between both boards,excellently produced video, but left with more questions ,oh well that's why maybe i don't make videos LOL

..but pleasantly surprised. It's a Carlson! ;)

Always leave a circuit diagram inside!

In Soviet Canada, Carlson Fix You, Eh!

@@ChatGPT1111 Soviet Canada? Am I get into a parallel world? :)

Thank You for your kind comment Shahriar!

Can't believe I've never heard of that show before, looks fun! I'm getting some episodes from archive.org now

I do remember Mr. Wizard. You must be an old curmudgeon like myself!?

Why a potentiometer that can switch between them (push in or pull out maybe) or a psu with both and a switch to select between them

They also used the multi-turn to "hide" the non-linearity of the voltage that is inherent in this type of design. The voltage is commonly defined by something like V=1/R, and that inverse creates a logarithmic response to the rotation. Look at 54:20, see how far he rotates the knob to get from 0V to 5V, then about 1/2 as much rotation to get from 5V to 10V, then a further rotation of the same angle takes it right off the dial (over 25V). There's a few ways to deal with it, varying in complexity and effectiveness.

Multi turn vr definitely results in an even more stable output.

@@johncoops6897 I also had noticed the oddity of the high number of degrees of rotation at low voltage settings for a given voltage change and the low number of degrees at higher voltage settings for a given voltage change. I was wondering if Paul had used a log taper pot rather than a linear taper pot for the 10 turn replacement. If you really wanted a single turn pot that was relatively linear in terms of rotation degrees VS voltage change I wonder if you could find a reverse log taper pot to do the job. Back when this supply was built I bet such a component could be sourced, especially if you were building thousands of such supplies.

@@Vincent_Sullivan - I have been battling with a similar issue on the pots for DIY frequency/function generators. I have yet to find a good (low cost) source for Reverse Log pots (they are commonly termed "ANTI-log"), but some come up on eBay search. I did find quite a few webpages relating to the creation of Log pots from linear pots by the addition of series and parallel fixed resistors that alter the linearity of the original pot. Whilst not perfect, my (very limited) initial tests show that there is certainly an improvement, especially in the middle part of the pot rotation. Some of the circuits had strange results at the extremes of rotation, but overall it's a good start. My next tests will be using a digital pot IC, as it seems that some of them can be programmed to log and anti-log characteristics.

I appreciate that!

CuriousMarc: Designing a replacement IC

@@fklabs6813 to be fair, the schematic of the IC is in the datasheet :)

Thanks for your kind comment Matthew!

Could be a limiting resistor on one, but both at the same time is kind of odd.

Given no known direct cause, I'm thinking something external but common to both. EMP? RF burst? Cosmic ray burst? Neighbor experimenting with breeder reactor in shed sees bright blue flash? Let's hope not the latter...

Transient power spike.

Thank You for your kind comment Dan!

A chap upthread (sorry, their nick escapes me) suggests that turning the power off (presumably when the transformer cores are fully magnetized) would / might cause similar to a flyback spike; and because the transformers are wired in parallel, both chips would get hit by the same spike. I'da expected the bridge rectifiers and large caps to have mopped most of that up -- but his still strikes me as a plausible explanation. Not sure how one would prevent a re-occurrence; MOVs or GDTs on the hot side, or TVS diodes on the low-voltage side, maybe? Seems like a only-once-in-40-years problem, though, so maybe not worth too much engineering :-)

I've been wondering about that myself...

That would be a hell of a collaboration! My two favorite RU-vidrs by far

@@midjetville Yes, it's a real pleasure to watch experts at their craft. These two not only restore but improve whatever they're working on.

Thanks for your kind comment Robert!

Hi Martin. The supply's output is only as accurate as the -5V supply. Zeners are very temperature sensitive if not compensated. To see an example of this, check out my "Heathbuilt" video.

That's a valid suggestion Cliff.

@@MrCarlsonsLab What about adding a MOV on the AC side? Isn't this the likely reason it needed repair in the first place?

That strikes me too. I wonder what really happened to this thing. The chips are just bipolar so simple ESD can't have caused it. More likely a huge surge on the mains perhaps?

I blame solar flares.

I agree. This thing is two totally separate devices sharing a case and mains lead. I’d assume the filter caps have an issue caused by age and Paul just happened to turn the thing on at exact mains peak voltage and a transient went through.

In modern designs with low ESR caps you get awful startup transients when the parasitic inductances - including those of the transformer winding - happily accumulate current and then dump a flyback pulse into the circuit. It takes good probing technique and storage scopes to catch those and fix the circuit. But this supply perhaps has died on turn-off, so that’s another conundrum entirely. When the power is removed, both transformer primary windings are coupled together - in parallel - with no mains acting as a buffer. So what you get is two supply boards with their inputs AC-coupled through the transformers. I bet you that there are “interesting” interactions happening there on power-down, and that those particular supplies were basically a problem waiting to happen. The reason they both failed at once is possibly that the slightly stretched flyback pulse coupled to both inputs via the transformers. The bulk capacitor should have arrested that, but somehow didn’t. Such coupled failures are never random - they may be low-likelihood but they are almost always due to latent design faults, and I bet that both chips were fine before they got killed. Alternatively one could see if a chip failing on power down could create a transient that would then kill another chip. These sort of “rare” issues are often not covered in design courses but they do become very important in high reliability designs. Imagine this was a spacecraft with redundant power supplies and one supply could couple via input transformers to another one and kill it. Not good news - and thus interactions of “redundant” or “parallel” modules via inputs are worth investigating.

Maybe there was a static charge on the mains or lightning down the street or something.

"You shall not pass...unless you hook up your isolation transformer and current limited variac first!"

huh. Hope MrC gives the failure reason in the vid

I was wondering that myself. Misbehaving old caps?

Most Likely a voltage spike on the line.

Vancouverman your probably correct but it is transformer coupled so much less likely than a modern switch mode

Full Wave Recked Hello!

@@allthegearnoidea6752 I was thinking maybe a lightening strike or a high voltage line coming down down on a lower voltage one.

Apparently your LM723s (?) weren't socketed, so the pot test would have been a bit harder to use :-)

@@theskett That's correct, but in a production environment, once the chip is removed it is simpler to just replace it with a known good one than to bother testing the supply without it. However, for one-off troubleshooting applications, Paul's technique could be useful.

@@sincerelyyours7538 True, dat. And don't get me wrong, I noticed that Paul's technique was pretty smart! :-)

Yup. 10mins effort, $12 for the two chips, job done.

When I encounter that issue I will post a video. Thanks for your input!

When he starts to talk about happy little VR's....

if he says happy little transistors, im out

Amen brother that little resistor would still be rollin, rollin, rollin.....

Thanks for sharing that Brad!

Yes, 3 would be ok, but 10 is over the top.

I have a couple of old school linear supplies where the voltage/current knobs are strategically placed over the 4mm binding posts... I can see why there's 10 turns, the number of times I've accidentally knocked the voltage knob or current knob without noticing :eek!

This one could use the concentric fine/coarse type of adjustment with gearing between the two so that the 10-turn could be driven faster with the coarse knob.

Thanks!

Great video again Paul. Thank you. I have a power supply that failed. Any chance you would share your circuit board design?

Thanks for watching!

There are exceptions :^)

Those low-voltage soupcans rarely go bad, it seems like. The two/three-section caps in tube radios, however...

This is what I'm wondering. What are the odds of both of them failing close enough in time that it seemed like a single failure? The boards look entirely separate until you get back to the plug side of the transformers. Only other thing I can think is a large shock on the case that somehow made it through to the ground of the ICs instead of getting all the way grounded out by the plug in time. That seems like a pretty far fetched guess to me, but who knows.

Of all the Mr Carlson's Lab videos I've watched, this one seemed unusual in its absence of thorough schematic examination and theory explanation. He didn't actually troubleshoot the problem, step by step, as he usually does, but simply jumped to the conclusion that the ICs were bad. From what little I could surmise, and admittedly I'm a long way from my electronics training and experience, so my opinion is hardly expert, the unit appeared to me to have two virtually identical channels, beginning with their power transformers, with nothing more in common than the line fuse (which presumably was not blown). Your guess is better than anything I could come up with.

logically speaking the only commonality was the line cord that fed both transformers, and maybe a ground plane that was also connected to the negative output post. so either a surge on the line or a back feed n the ground circuit. i wonder if his viarac is ground isolating and if he uses that to power his test equipment

Glad you enjoyed J!

You bet. More zero's, and less drift with temp.

Thanks 👍

So , I'm not alone in thinking that !! The moon is too close , so I'll "just" move it ............

The Darlington driver (MJ1000) has a collector at 44 Volts and will feed current through the base pin of the 3055. This current is supplied to the load through the emitter of the 3055. This keeps the drop across the 3055 low, this results in less heating of the output pass element, as the two share the burden. Edit I should add: The brown lead has a higher value resistor across the the base and emitter of the 3055, so keep in mind, the drop across this junction can only result in .6 Volts. There is little to no current demand through that resistor.

@@MrCarlsonsLab Thank you for paying attention to the comments and replying to a question from a confused subscriber! :) Excellent work here, I just hope you keep doing what you have been doing.

They are a nice supply!

Thanks David!