No.093 - Agilent E3649A PSU Repair - Revisited!

[@jaycee1980]

I get things like this when im repairing old computers (usually Commodore Amigas). I now do a "drop test" - literally i lift the board 2cm off the desk while running, and drop it. If the thing crashes, there's an intermittent connection!

[@IanScottJohnston]

Yep, I used to design electronics for offshore, we even wrote the "drop test" into our procedures.....LOL

[@TheEmbeddedHobbyist]

I hated faults like that, you never know if you're fixed it or just extended the time between faults. in the same way get you get a repair and you have to return it with no fault found, but in the back of your mind you know it will be back. ;-(

[@robbieaussievic]

.... Intermittent faults are the toughest in all repair industries. Our brains love puzzles though, (as exercise).

[@IanScottJohnston]

Yes, its soooooo important to see the actual fault no matter what it is.....and even if it takes a load of time.

[@TheEmbeddedHobbyist]

@@IanScottJohnston Sometimes we did not have the luxury of time, there would be a taxi driver wating outside to take it back. So it would be locate fault, fix, test, put back in box and give to driver as fast as possible. And quite a few would go back with NFF (No Fault Found) labels attached. they might be back a few hours later for a repeat and another NFF label. We would never find where the real fault was, it would just stop coming back. but I did go back with one unit to prove it was not our fault, and got a very nice dinner out of it.

[@davidv1289]

At about 4:40 the display when a little buggy when you wiggled the large ribbon cable. It might be worth double-checking it before returning the supply to service. Those fiberglass brushes are great - first saw one on the Defpom's channel and ordered one right away. Thank you for another interesting video.

[@stazeII]

Saw same thing and thought the same. Would reflow that ribbon connector.

[@IanScottJohnston]

Yes I noticed the analogue signals being abused a little by the loose nature of the ribbon connector. To improve it would need a new ribbon fitted with tighter sockets on the male pins of the pcb connector. Its not a big problem, I wasnt half pulling at the ribbon. Its not the Pcb.

[@IanScottJohnston]

If you like the fibreglass pens, have a look also for abrasive pcb rubbers. Great for wider areas.

[@stazeII]

@@IanScottJohnston oh good. :) was thinking cold joint on the connector. Thanks! Great fix!

[@SeanBZA]

@@IanScottJohnston Probably just needed to be cleaned in the pins and socket. Cable would probably benefit from a half minute in the ultrasonic cleaner, to get the inner contacts dirt free as well. Would not hurt to put all in there at the same time, to get the last bit of gunk out.

[@SalvagedCircuitry]

Hi Ian, huge fan of your work! Glad you got your PSU back up and running. Something to note, when using an ultrasonic machine make sure to have the fluid level up to the level recommended by the manufacturer. It's usually about 7/8 of a unit's capacity. Also, don't agitate anything inside the ultrasonic machine once it is in operation. This can create unwanted stress on the ultrasonic transducers and reduce operational life. I've repaired a crest ultrasonic that was abused by a medical lab that did not use enough or any fluid. All the transducers cracked and needed replacement, and nearly half of the transducer terminal plates developed shear lines. I'll hopefully have a writeup on my site soon about it :D Thanks for the video!

[@IanScottJohnston]

I read about these same issues from other sources over the years, but to be honest i've never had a problem in all my years. Perhaps because i only use the cleaner about once a month. The cleaning fluid I buy is in 5ltr containers, hence its probably not quute enough for the cleaner. Cheers.

[@SalvagedCircuitry]

@@IanScottJohnston Got ya. Its possible the used ultrasonic I picked up was over abused. There were ring marks on the bottom of the tank from where beakers were placed, likely with no surrounding solution. Crazy. Good luck with it!

[@radio-ged4626]

The worst ones are when you find something that seems to be causing the intermittent fault. You soak test it for a week, all's OK so you send it back to the customer then 3 days later it comes back. Sometimes it makes you wonder what you have to do to ensure the fault has gone away.🤦‍♂️🤷‍♂️

[@oldavguywholovesRCA]

Connection issues being mechanical in nature can be detected with heating and cooling cycles. use a little heat from heat gun then a little cool spray to make the connections flex. Works very effectively. Love your videos.

[@BjornV78]

17:30 You can try to place the Piëzo buffer in a vacuum chamber to dry it.

[@allthegearnoidea6752]

Enjoyed watching the fix and seeing the board cleaning. I guess time will tell regards the fix. Thanks regards Chris

[@t1d100]

Thanks for sharing! Glad that you went the extra mile to find the real cause and to repair it well.

[@piconano]

Love your repair videos.

[@GadgetUK164]

Excellent job =D Frustrating when you get intermittent faults like that!

[@TheDefpom]

Thats a nice big ultrasonic cleaner, nice repair, I would be inclined to tack those wires down with some UV soldermask.

[@SeanBZA]

Or just spray the entire underside of the board with a conformal coat, so that both the wire links and the solder pads are protected from corrosion then. I have done a good number of boards with only spray clear lacquer, which works well enough, and provides a good coat with one or two light applications, and holds down that wire nicely.

[@IanScottJohnston]

The entire underside will get sprayed with conformal coating once i know its reliable and which will help secure those kynar wires. Saying that, they are only short and as long as they dont flap away from the board they would be fine.

[@mr1enrollment]

Not always but some times heat + cold spray may also be used too produce difficult faults.

[@NewbFixer]

Thats the most expensive 20v battery charger ive seen in my life 😂

[@IanScottJohnston]

Erm, i have three of them! LOL

[@absurdengineering]

How did that corrosion come about - any ideas? Did the capacitors underneath (if any) spew something, or was it contaminated in production or unrelated servicing earlier, or…?

[@DavidSmith-zx7wz]

Love the repair videos!!! Thanks

[@gbowne1]

I've noticed the same things with some of the other E36xx series with small traces tracks, the ones made in the mid/late 90s. I think they fixed this later on.

[@IanScottJohnston]

Yep, the traces appear to corrode under the solder mask hence the different shades of green. No real way of stopping or preventing it really.

[@simontay4851]

Why did the traces corrode in the first place. Was there a leaky capacitor? Was something spilt on it?

[@IanScottJohnston]

@@simontay4851 Under the solder mask I can only think it was a manufacturing problem, i.e. etchant not cleaned off properly or something like that.

[@SeanBZA]

@@IanScottJohnston Yes I would say probably was poor process control, or there was a section of the large panels with droplets that were left during washing, possibly a clogged nozzle in the washer that did not allow enough flow and pressure, and thus the odd boards with traces of etchant left on them. Then dried off, nobody noticed the slight white residue before they went through the silk screening for the mask.

[@stevensexton5801]

The board is a really nice green once it was cleaned up.

[@IanScottJohnston]

Electrolube Safewash SWAS solution in the ultrasonic tank always does a great job.

[@InssiAjaton]

Interesting how you did it just with your finger! My thoughts went all around using a logic analyzer, triggered from the reset and studying what happened just before. An alternative would be a modern oscilloscope, especially something like my 4 Analog + 8 (or 16) Digital channels. Same idea, trigger from the CPU Reset. But as there is a saying, "Keep it Simple, Stupid!" then, if you have a magic finger, use it. Hi!

[@fredflintstone1]

Woo Hoo!! another video repair 🙂 don't you just hate intermittant faults

[@andymouse]

SQUEAK!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[@fredflintstone1]

@@andymouse Cheese!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[@TMaxElectronics]

which ultrasonic cleanic fluid have you settled on? I'm considering switching away from pure ipa for safety reasons

[@IanScottJohnston]

Electrolube Safewash SWAS05L. I have tried a few over the years, but this works the best. It needs 50degC or there abouts for best results, and the Pcb needs washed thoroughly in water afterwards.

[@Chris_Grossman]

When you got issue by pressing on the chips I had my money on a PCB fault. Good work. What is the solution you use in your ultrasonic tank to clean the board?

[@IanScottJohnston]

ELECTROLUBE SWAS05L, SWAS Safewash Super Flux Cleaner, 5L Bottle

[@ratlinuxgamer2855]

Another one, nice 🥳

[@tyronenelson9124]

That PSU has far too much tech in it to go wrong, in fact it has more microchips in it than a flat screen tv or a DVD player has.

[@alpcns]

Thanks for sharing, Ian. A great fix. These modern supplies are quite complex - what's your experience, are they as reliable as "old school dumb" supplies? Being a dinosaur, I am always somewhat weary of these high-complexity designs with all these miniature tracks and components.

[@IanScottJohnston]

Funnily enough, if I am charging say 25off PDVS2mini's at the same time and especially if I am leaving them over night I use a couple of basic non-processor based CSI power supplies!

[@andymouse]

Pig of a fault, but I bet you got it !...cheers.

[@TheEmbeddedHobbyist]

do pigs squeak ;-)

[@IanScottJohnston]

@@TheEmbeddedHobbyist they do in some people's world i would guess.....:)

[@rogeronslow1498]

A word of advice so you can avoid melted wires and a possible fire. I have the same unit at work and I was charging a 7.2Ah 12V battery. We had a power failure for a few minutes and when the power returned the charger came up with its output disabled, as expected. However, when the output is disabled the pos and neg output terminals are shorted together internally, resulting in a dead short across the battery and smoking cables. The PSU somehow survived it but I never charge a battery from that supply anymore. I latter chatted to a colleague and he said he had the same experience with another model of HP/Agilent charger so I assume its a trait of HP PSUs.

[@IanScottJohnston]

Ouch!......not a great scenario there! Luckily for me my charging circuit doesnt do the same, and i only charge at 240mA max.....so the psu is always set with a 300mA limit.