gerrysweeney.co... This video is about an Agilent E3634A PSU Teardown and Repair I done recently. I purchased this PSU on e-bay - it turns out there was a lot wrong with it, in the end though its working and serving on my bench.
Indeed it was, I thought it worth sharing and as it turns out someone else had an almost identical problem in an almost identical PSU, his comments are on my blog page that contains this video (link in the video description above)
Hopefully working hard on making more awesome videos! You are about the only guy on RU-vid that makes videos of my favorite aspect of electronics... modding test gear!
Hello James, thats a good question. I guess a combination of stupidity and experience. The voltage they are operating at a max of 50 VDC in this particular PSU. 50V is not enough to hurt you, unless had wet hands and open wound so there was no shock hazard. However, there is a risk of damaging the unit, or your probes/screwdriver if you discharge the cap into them. However, there is a bleed resistor so the caps discharge to a safe level of 2-3V in about 20 seconds, so I was not bothering.
Assuming your simplified circuit diagram @33:00 is accurate enough I think there might be another (much smaller) design flaw: the secondary voltages use the big +-rail as GND but it's connected after the shunt resistor while the main switching mosfet(?) is bevor the shunt. Wouldn't it be better if both were on the same side? - so if/when the shunt fails the PSU doesn't kill itself because most/all parts are still referenced/connected to the same GND (only when an external load is connected)? With the failed shunt it's kinda split-GND isn't it?
If you hold the output switch down while powering up you will get an error code that helps point to the problem. There are also many test points on the bottom of the board which should be checked before tearing down. But in this case there is an obvious burnt component so...?
I honestly will probably never understand why precise equipment like that is frequently battered so much. Are people using these as stools? Are these counterweight when lifting bricks up on a construction site? Are these being used to measure/power anything while somebody throws rocks at it? Are people in general so clueless to not be able to move a piece of equipment from place to place without hitting it from every possible side? I mean, honestly, you set this on a bench or a shelf, you connect power and leads and you press buttons to operate and you usually do this in a context of electronics repair which, at least for me, never required anything that could even potentially create that kind of dent in a front face of the instrument.
Oh OK, now I see, its a square head screw - no that is not at all common here in the UK, in fact I am not sure I have ever seen one. I have seen hex and torx but never the square head screws, I must look out for some....Gerry
I cannot remember the error code 609 but I have certainly repaired a few with ADC problems, they are easy enough once you set yourself to be able to diagnose the problems and you have a schematic to hand.
Come on Gerry, to use one of your UK words, I was being cheeky! Notice the wink at the end! Tenma used to be "decent" years ago. I think they are under new ownership or something because most of it is crap now. Some items have same model numbers but cheaper parts and horrible workmanship compared to one manufactured before. Obviously they were never in the same league as Agilent ever, not even close :-)
Nice video and explanation of the fault. I was watching 'Thesignalpathblog' the other day and he had the same model with a much easier fault it turned out. He used the self test power on utility that he said saved him a lot of time. Were you aware of this and if not, to use it, you press and hold the 'Output on / off' button while switching on the unit until it goes into self test. I'm only saying this as you never mentioned it in the video. going to watch the video on E3646A now.
Michael Hawthorne thanks for the comments, I was not aware of the self-test utility at the time but am now, I have fixed a few variants of these things since. This PSU ended up with another fault, it exploded and I had to fix it again :( I done a video on it, as it turned out I was unfortunate the first time... ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-zQKkrC5VwHY.html Thanks for watching. Gerry
Just another echo... Great repair. My nomination for Queen's Award for Getting Stuck-in. Top marks. Are you possibly assisting Shahriar under the alias of Gary? :)
964tractorboy Hi, thanks for your comments and nomination. I quite enjoy repairs like this that have challenges. Shahriar is a great engineer, so assist might be a string word, I sent him an IC that most would find impossible to find to help out with one of his repairs, not really assistance, just a favour for a fellow blogger from who I have learned a lot. Thanks for watching. Gerry
I tried to find a circuit diagram of the device to get some more details of it. But in the Agilent Service Guide i found at the Agilent Document library on their web site i could only finde component locator, block diagram and such things. I am assuming you had a circuit diagram for repair - where is the place to find/get?
Just watched this video (better late than never). Good working repairing such a hard to find fault. I was surprised, though, at the comment that the current shunt resistor wasn't available. I recognised the logo on the resistor in the blog photo from as beloning the German brand Isabellenhüte. Type PBV R050, 0.5% should still be available, e.g. from Conrad electronics (search for item 447382-89). Thought I'd let you know just in case you decide to put back in the "proper" shunt resistor.
+Jitter 000 Thanks for the info, I did eventually find the part on-line although i have not bothered replacing it, the PSU has been flawless - if it ain't broken don't fix it :) thank you for your comments. Gerry
Really interesting repair video, where the repair is a bit more involved than a loose connector. What an interesting journey to analyze and debug that circuit! I can't imagine how you'd be able to do that without a schematic to analyze the function of the power supply.
yes, the schematic is most definitely a must have - would be very complicated to figure out whats what without it. Thanks for watching, and for the feedback. Gerry
It was not that much work really, I admit the secondary fault had me stumped so it took me three couple-hour sessions to sort it, I was convinced the thing was oscillating under its own steam...the schematic convinced me. Sorry about the meters, its a long story. I do make use of them more often than I ought to be able to, I like to hook all the measurement points up in one go, I can often have 4 or 5 on the go at once. I won't tell you how many HP PSU's I have :)
it depends on what you are checking for, and what the circuit is doing. In this case, checking a transistor for a dead-short while in the circuit is a perfectly good technique, but you do need to properly understand the electronics and circuit around the device to make that judgment.
@@gerrysweeney2127 Yes. I've been thinking about it. I guess the idea is this - if the meter shows no short, then the transistor is probably ok. But if the meter shows a short, then either the transistor is shorted or a path it is connected to is shorted. So it would then make sense to isolate this transistor and check it again. That the idea ?
@@youpattube1 that's great. When doing this kind of test you are looking for clues, you might not want to go to the trouble of removing the transistor if it's not short, you might choose to look elsewhere first, so if the meter reads a dead short, then you know for a fact the transistor has to come out. Then you test it out of the circuit, and also cross-check the circuit is not short in the absence of the transistor, just a process of elimination
A huge amount of work, both in fixing the power supply and making the video. I hope it was cheap! Very clear and well presented work - but having matching 87Vs *AND* 289s is just rubbing our noses in it. In future, please use a motley collection of different DMMs.
Couldn't the designers stick cpl of Zeners on the op07 input ? it would save all the hassle in my opinion. Gr8 stuff Gerry ! Surely I'll stick around. Rgds from Edinburgh.
Hope you dont mind comments on early videos. That looked like a bloody nightmare gerry :-(, ive been repairing things for years but i think i would have given up on such a task, "Overwhelming" is an understatement. Over complex designs causing complex failures, brilliant repair :D I think ill stick to crude diy L200/LM723/LM317 based supplys with a hand full of 2n3055's chucked in on heatsinks ha ha :D
I Mark, I am pretty sure using Schottky diodes instead of standard silicone diodes would have made a difference. I guess though in reality they probably did not anticipate the problem. Thanks for watching from Edinburgh - a fantastic city.
It cost $90 plus about $50 shipping to the UK (and £18 in customs and *handling* charges......agggrrrr) so not that cheap really. However, it was interesting to fault find on and learn some more electronics along the way, and its always nice to end up with something useful at the end of it. You may have noticed that I am also messing with video so different segments had different lighting and stuff as I was experimenting - so it was a good actor subject too :)
I am afraid I will have to disagree with you on that one, as PSU's go these Agilent ones are pretty darn good, well built, well designed and work nicely too (when they are not broken of course). I have something that *looks* like a Tenma PSU badged Skylab and its the worst piece of crap ever
Thanks so much for this video. Do you happen to still have those photos you took of the connections from the transformer to the power connector? I thought I took enough, but it appears I didn't.
@@gerrysweeney No problem Gerry I eventually got it sorted out after a few blown fuses haha. Mine had almost the same issues as yours just not as fried. I found the original shut resistor in china which works well. Just had to replace that, the 33 ohm resistor, the fet (q5 only) and an OP7 op amp and I’m back in business!
Oh, the head type is called Robertson here in Canada. It is the most common screw head we have. All screws you get from the hardwares store use them. Except for drywall screws, they are still Philips (cross).
Enjoyed the video. I'm a relative novice to electronics but I noted you did not discharge the large capacitors before starting your tear down. When would you and when would you not?
I know this video is wary old, but maybe you still can use my advice. To undersold components, placed that awful, I do wind some thick copper wire around my soldering iron and use the copper tip to heat the tin.
I have never repaired anything on this, but now I am required to fix the front panel which is broken /bent at points. The interior is working fine, is it possible to repair the control knobs