TSP #11 - Teardown & Repair of an Agilent E3634A 50V 200W Power Supply

[@Thesignalpath]

I dedicate a portion of my salary for my lab. Many of the equipment you see were purchased and repaired.

[@UrElectronicsFanatic]

Well done! I always like it when people take you step by step through the process of elimination. Now I want to find a power supply to fix myself.

[@NebukadV]

I know this video is pretty old, so chances are, nobody is gonna read this ever :) Since nobody has pointed this out yet, I will^^ This whole repair worries me a little bit. There is a 2.4 V Zener, that is supposed to bring the 17.4 V down to 15 V (exactly 2.4 V difference). Now the 17.4 rails is reading high (about 4% higher than normal), which could just be due to a slightly higher mains voltage coming in. This is of course no problem, since the 17.4 rail is not regulated and does not need to be precise. The 15 V rail can also be a little bit off, since it's also not regulated. But what bothers me is: The 17.4 V rail is reading about 4% higher than it should, but the 15 V rail is reading almost 3% low. In fact, there is now not a 2.4 V drop over that Zener, it's almost 3.5 V. This leads me to believe, that the zener diode it operating above it's rated current, which is not a good thing, especially in such a tiny package. Either something is faulty on the 15 V rail and drawing excess current, or the Zener diode is constantly overloaded by design. Considering the voltage drop of about 3 V also on the negative rails, the latter might be the case. This might have also been the reason for the "death" of the original diode. I would have at least measured the current, that is now provided by that zener to compare it to it's specs and maybe put a larger package in, to keep this fault from ever occuring again.

[@NewScope8]

You totally nailed this one. First of all, the 17.4V is not 17.4V to begin with. It comes out of an LM317. If you calculate the LM317 output voltage with the feedback resistor values on the schematic (which is consistent with the real board by the way), it comes to 18.2V. And then the 15V is not really 15V. It's 14.6V as our dear professor measured. A 2.4V zener has poor IV curve as a zener diode, perhaps because it is at the low end of the process that a zener can be made. It supposed to drop about 2.4V at 10mA. But in the circuit it is passing 110mA and drops 3.6V. Guess what, the poor little guy is rated for only 250mW at 25C ambient. It must have been slowly cooked to death even with forced air cooling. The similar situation applies to the 2.4V zener on the -15V, where it is passing about 67mA @ 3.5V, not grossly over spec but marginal. The current is not unreasonable, considering that the +/-15 rails power about 27 opamps total, each drawing 1.4mA(typical) to 2.5mA(max) quiescent current alone. If it is by design, I can't imagine how this series of design oversights happened at HP/Agilent. Well, we know the professor's fix works, until the zener diode dies again. In the mean time, I'm having trouble finding a zener diode that has high enough rated power, low enough thermal resistance and small enough footprint. There is definately not enough space for a 78M15 here.

[@Thesignalpath]

I appreciate your advice. I am well aware of the dangers of breaking a component in half. Unfortunately given the amount of space I had to work with, I decided to break the component. Furthermore, the unused pad had no trace on it, therefore it was easy for it to peel off.

[@gamccoy]

Good job troubleshooting. I'll give you 5 out of 10 stars for the repair, however :) You really took me back. I used to repair comm gear in the Marine Corps for 20 years. I miss it.

[@Thesignalpath]

Thank you. I ensured everything is discharged. Your advice is very important.

[@thewii552]

ahh, ok. I love your lab- It's just so nice... Both the fact that you actually keep it tidy and organized (which I cannot manage) but everything is just where you need it... And you have pretty much everything you'd ever need.

[@TheCrazyInventor]

I love video's where I can see other people trying to diagnose and solve a problem on a piece of hardware. :)

[@Narwaro]

Agilent / HP hear has historically been a pain in the ass to troubleshoot on quite some occasions, at least in my experience. Not only did they produce boards WITHOUT ANY silkscreen on it in the past, but they are also often dicky about schematics. The 90s and early 2000s were quite of a "dark era" IMHO. Repairing gear before that is a joy forever.

[@Thesignalpath]

Yes! I saw that too. I placed a drop of glue on it before closing it up. Thanks!

[@Thesignalpath]

I try to keep it as clean as possible. I find I always get results faster with less frustration in a clean environment. :)

[@Sahko123]

imagine watching this video in 2021 cause you got two of these free and luckily you have the same issue

[@GeorgeTsilikas]

Two great videos on a single day! Fantastic, thanks Shahriar.

[@elisacollosi]

great walk thru on diagnosis and repair. details with practical application. love it thanks!

[@TheDefpom]

I just picked up a E3646A which is faulty (I purchased it to do a repair video on it), they are electrically very similar to this unit, seeing the non changing display was a reassurance as I wasn't sure if that meant a controller fault or not. On my unit only one channel is effected, and both channels are electrically identical which means I can cross check between them both to track the fault down so I will be very surprised if I cannot fixed it, I haven't started it yet,I'm just doing some research before I start it.

[@xhighalert]

That instrument is so beautifully documented... I wish it was always like that.

[@maxjacoby3048]

Sorry to be off topic but does anyone know of a method to get back into an instagram account?? I was stupid forgot my account password. I appreciate any help you can offer me.

[@maxjacoby3048]

@Juelz Colin I really appreciate your reply. I got to the site on google and Im trying it out atm. Takes quite some time so I will get back to you later with my results.

[@maxjacoby3048]

@Juelz Colin it did the trick and I finally got access to my account again. I am so happy! Thanks so much, you saved my ass :D

[@juelzcolin3646]

@Max Jacoby No problem =)

[@Thesignalpath]

My lab has been 16 years in the making. It didn't become like this overnight :)

[@mcsniper77]

I love your videos. Your doing such a great job that broken instruments are selling for more than good ones. People tend to think there garbage is worth a fortune this is nothing new but in today's market it is worse than ever. Just look at guitars a '59 Gibson Les Paul in mint condition by now is probably worth around $600,000 ridiculous. I fear many suicides when the bottom falls out of that market.

[@johnhenry1222]

It is a really good tutorial. Thank you for your time. I have a slightly the same issue you have. Mine passed full self-test but still does not give out voltage or current. Do you you have any clue for it? thank you

[@Narwaro]

For such small components a reflow gun with a small nozzle is a godsend.

[@octavmandru9219]

I do not have a problem with the removal procedure, given the circumstances. The pad broke most likely because it was cooked by the diode. Congratulations on the successful repair

[@mbaker335]

Breaking that diode up made me wince. Congrats on the repair but I would have thought desolder braid might have worked. Or a hot air gun and angles=d tweezers. Hot air desoldering guns are cheap enough and useful for a multitude of purposes.

[@JoeyaoChou]

Great Video. You're lucky that there's just a single broken diode. P.S. Glad to see that you're using RIGOL Instrument, which is a Chinese brand, and made in China.

[@JorgeHernandez-ji5gb]

You forgot something in the list of things you needed to repair it: a Ph. D. Lol. Awesome videos, Shahriar.

[@pychneag]

Very nice! I learn so much from your work.

[@Thesignalpath]

A good place to start is eBay. But don't expect to be able to fix every broken equipment you find. I as mentioned, this was almost a lucky fix.

[@thewii552]

Love these videos. Also, I wish I could set up a lab like yours...it looks wonderful.

[@snipersquad100]

Excellent repair video.

[@BGaborr]

well done ! fixing something broken escecially if its expensive is awesome :)

[@PaulHigginbothamSr]

Shahriar: how can you brighten the display? Not as easy as with fluorescent tubes and leds. I bet you can do it without changing the display out, right?

[@phuang3]

It's a design flaw. Maybe you should solder a 1W zener on it.

[@vaualbus]

Where you find the schematic circuit? I need to repair mine alimentator.

[@Thesignalpath]

Thanks! It is a toy shop!

[@NuggetNapper]

I just noticed that the display at the beginning is not showing the current to be 8.16A but 8.16kA.

[@juopotski]

Nicely done! More repair videos, please! :)

[@tfinmoraes]

Hi, Shahriar! Thank you for the video. Could you tell me what is the light source (brand/model?) you used when you were replacing the diode?

[@igwood71]

Beautiful! Thanks for sharing!

[@oBsRVr666]

You really notice that this PSU was designed to be repairable. Agilent / HP documentation is usually excellent and widely available. That makes it relatively easy to fix. I've recently bought a broken Fluke 87-IV hand held multimeter on a HAM fleamarket. Documentation (Schematic) available? Nope. Silkscreen legend on board? Nope. Custom parts available? Nope. Any chance of repairing this? Nope (most likely). I bought a second working Fluke 87-IV to do comparative measurements ...

[@drabara]

I love the repair video's ! More ! More !

[@brig.4398]

could you maybe have tried some Chip Quick to remove that part?

[@oBsRVr666]

... on both units. But it will probably still take me forever to find the error. And if one of the custom ICs inside the DMM is broken I am screwed. Cheers from Germany, Florian

[@thewii552]

Just wondering, your lab always that tidy? I know mine never is but I usually tidy it up for video...

[@user-py6ji6jv4u]

I have repaired more than 100 E363X series power supply ,this series psu is under-quality, often with some common failures such as wrong +-10V ref voltage\ and broken risistors with green color .

[@w6by]

It's easy to remove SMT parts with two soldering irons...

[@arcadeuk]

I love your lab, it looks like a toy shop :)

[@ajw6715]

Hi, Very good video! I have a question about a Tatung power supply I bought on ebay. It is a PS 3336 and it is suppose to put out 60 volts and 2 amps. When I turn it on and turn the volts and amps up it will go the 60 volts but the amps only goes to .16. When I put a load on it the volts drop back to 2 volts and amps stay at .16. I have inspected the boards and I don't find and burned out components. There is a big cap that charges to 72 volts and it is rated at 100 volts so it is fine. The power supply has 1 transistor in the back and it does not heat at all. It has a large transformer and it looks fine. Any ideas what cound be wrong?Regards,Arnold.

[@rotlerin]

I agree. Also have a look at some magic stuff called CHIPQUIK.

[@Az00ra]

Great video. Thanks.

[@dirkovic2112]

Or he could set up an other camera or iphone to film the multimeter and display it simultaneously in a corner of the video.

[@percih70]

Does the e3634a sink current with no offset current? Some Agilent supplies allow you to sink current with the same resolution as you can deliver current to a load, but some of the early models have a large sink current offset (250ma or more) so can't be used to accurately sink current Does anyone with access to the e3634a know the sink current spec? I couldn't see it in the user manual? 73

[@thewii552]

yeah, haha. I find that too...but for some reason it really never stays tidy...

[@jtootall67]

great video thank you

[@Neolith100]

This was great!

[@hla27b]

I love it when it works :)

[@MrMac5150]

Will you direct me to the right place to find good equipment like yours, at reasonable prices, have a low budget...

[@K2DO]

Can you please post the schematics for download, Thanks mate,

[@ronaldlijs]

Keep up the great work Shahriar, you get my full support and will watch all your videos! I like Eevblog, but David is losing it nowadays, too many viewers and too much money I guess! :-) Anyway, great stuff!!!

[@felipeceglia]

Hi, I have a e3634a power supply, which fails the fan test. All teh manuals I could find on Agilent site miss the schematics. Maybe you could forward me one copy? Thanks!

[@yoramstein]

Hello signal path bloger-May i ask which country are you blogging from? Where is such nice broken equipment to buy?

[@gacherumburu9958]

👍👍

[@ElectronicsPubVideos]

Afarin Afarin

[@dunaihorgaszok]

You can find the schematic of the power supply here at the end of the pdf: gerrysweeney.com/wp-content/plugins/download-attachments/includes/download.php?id=1890

[@fantasticoadidas]

Hey, can you also repair the user interface of this thing? Man, that has been implemented sooo badly ;-)

[@RussellTuan]

cool!!

[@thewii552]

How do you get all the money for your gear- Save up and repair stuff like this? Or do you get companies like Rigol sending you stuff like I know Dave from the EEVBlog does...

[@kevinhirt]

24:36 finally something you dont have

[@DLTX1007]

Man ... i love your lab, chock full of orgasmic equipment i wish i had the cash and space

[@razean22]

new video, yesss :>

[@MasterNiva]

Hard to believe that a respectable company as agilent dont fix mistake done it at the board and decided to install a nice and nasty jumper, instead resolve the problem properly, this is not a 20 dollar chinesse power supply that doesnt have even a brand name, this is a very sophisticated and expensive power supply, I never seen jumper like that installed from factory in a Fluke or HP equpiment just for mentioned some

[@TheProCactus]

why are none of your videos working

[@Porto88]

low audio

[@kevinhirt]

3rd comment on the vid sorry but HOW MUCH DID ALL THAT EQUIPTMENT COST YOU I WOULD GUESS < $20,000 USD

[@FooBar89]

ugh, poor pcb ;(

[@kevinhirt]

23:31 WOAH BRO WATCH WHERE YOUR HANDS GO THOSE CAPS COULD STILL BE CHARGED , I HOPE THE WERE DISCHARGED. I DIDNT SEE THE uF READING BUT I IMAGINGE ITS A LOT BE CAREFUL

[@user-fp3bf4mv2b]

Хоть бы не позорился КОСЯЧНИК, дорожки поотрывал, можно сказать зафаршмачил хороший БП.