Rigol DP832 - smoked channel 1 [FIXED]

[Skimask]

Pulled a dumb-ass move this afternoon.  Needed to top off a 12v Pb batt.  Plugged it in backwards.  Got some nice sparks at the alligator clips and that wonderful odor of smoke escaping from somewhere inside the '832.
Don't have time to pull it apart for at least a few days.
Who's got the schematics and/or a good idea what part I have to try and put smoke back into?

[c4757p]

Probably, a chunky diode whose entire purpose is to sacrifice itself to protect the rest of the PSU. Should be easy.

[robrenz]

Dang... the one time I have a clue and you beat me to it

[Skimask]

Odd...doesn't smell like a diode 

I'll try to make a cheesy video of the repair when I get to it.
I can't possibly be the only one that has invented such stupidity...or could I?

[bronson]

Given how the 832 puts positive on the left, I'd guess you're far from the only person who's done that.

[Howardlong]

Given how the 832 puts positive on the left, I'd guess you're far from the only person who's done that.

Absolutely, not the smartest design decision ever made that's for sure. The first time I used the DP832 I operated it the wrong way around despite the red and black leads I use. Luckily I usually put in series Schottkies pre regulators as a matter of course on my breadboarded stuff to save me from myself, clearly it worked.

[jadew]

Given how the 832 puts positive on the left, I'd guess you're far from the only person who's done that.

Absolutely, not the smartest design decision ever made that's for sure. The first time I used the DP832 I operated it the wrong way around despite the red and black leads I use. Luckily I usually put in series Schottkies pre regulators as a matter of course on my breadboarded stuff to save me from myself, clearly it worked.

I'm not sure what supplies you guys have been using, but I think that's the standard. I have two power supplies from different manufacturers on my bench, both with the positive on the left. Agilent has them like that too, which would probably account for the majority of lab power supplies.

So... what brands have them the other way around?

[Circuitous]

All of my power supplies, from 5 different manufacturers (Keithley, Rigol, Array, Maynuo...) are all Positive on the left, same with the DMMs and electronic loads.

[tautech]

All of my power supplies, from 5 different manufacturers (Keithley, Rigol, Array, Maynuo...) are all Positive on the left, same with the DMMs and electronic loads.

Interesting  seems counter-intuitive to a right-handed person and certainly not what I have used or am used to.
Must just be the Asian stuff I have here.
Fluke 15B
Vichy VC99
SDM3055 Bench DMM
MCH PSU's

All positive on right
Siglent PSU's also.

I wonder where this apparent LH positive standard originated from and how far back it goes?

[Nerull]

I don't know about other companies, but HP started in 1958. Their PSUs of the era used a three-connector triangle configuration, which generally was setup with the positive and negative outputs oriented vertically, and the ground beside them, like this, from 1955:



In 1958 they rotated it, with positive on the left.



By the mid 60s they looked like something you might see on a PSU today:

[jadew]

Personally, I used positive, negative and ground (from left to right) on my home made PSUs and that's before I even saw a professional PSU.

I remember I gave this some thought and it felt like the right way of doing it, so there might be more to it.

[c4757p]

Personally, I used positive, negative and ground (from left to right) on my home made PSUs and that's before I even saw a professional PSU.

I remember I gave this some thought and it felt like the right way of doing it, so there might be more to it.

IMO, there should be a direct line from ground to both of the output terminals, so that you can ground either one with a short bar.

[jadew]

Personally, I used positive, negative and ground (from left to right) on my home made PSUs and that's before I even saw a professional PSU.

I remember I gave this some thought and it felt like the right way of doing it, so there might be more to it.

IMO, there should be a direct line from ground to both of the output terminals, so that you can ground either one with a short bar.

Yeah, I don't really like the ones with the ground in the middle. Can't put my finger on it, but I think the benefit is too small to have to deal with that annoyance all the time. It just feels like it's in the way.

I never ground the positive terminal anyway, if anything I tie it to the negative terminal of a different supply, which may or may not be grounded - makes a lot more sense to me. If there was a situation where I would use it as a negative supply by itself, I wouldn't mind using a longer piece of wire, if that meant that I can keep the ground lead out of my way the rest of the time

[c4757p]

I adore the early HP triangular configuration. We need to bring that back.

[jadew]

I adore the early HP triangular configuration. We need to bring that back.

I agree, no compromise!

[tautech]

I adore the early HP triangular configuration. We need to bring that back.

Yes it is smart.
Does it need a different colour scheme?
Red   Black
   Green
And in a binding post configuration of course. 

[ivan747]

By coincidence I was meditating on this exact topic just before visiting the forum, and in fact, I came to the same conclusion that the early 50's HP triangle is the best solution. It's the most intuitive, has positive and negative just as you will find it on an schematic (positive is the uppermost, negative is on the bottom), it has 3/4" spacing between all connectors, you can use 3/4" spaced dual banana plugs as well. Binding posts accept bare wires, U terminals, banana plugs, alligator clips, clips, short bars, actual bananas...

[ivan747]

We kinda derailed the thread 

Have you opened the power supply yet, OP? Show us photos of the board, there should be visible damage it the thing smelled like you said.

[Skimask]

Working on cleaning up the shop, first time all year and it's a pit.  I got a bench cleared.  Gonna work on it in a bit.
Don't expect much though.

[Skimask]

https://www.youtube.com/watch?v=f2hQiE-G_f8&feature=youtu.be

Found R52 on the top board smoked.  Current sense resistor for channel 1.  Think it's a .020 ohm, 3W, 2512 (or is that 1225?) sideways mounted SMT resistor.

http://www.digikey.com/product-detail/en/RL3720WT-R020-F/RL37WT.02FCT-ND/567370

http://www.mouser.com/Search/Refine.aspx?Keyword=66-LRF2512-LF-R020-F

May as well buy a small bundle of them, if they're the right type.

The trace on the PCB is broken/burnt, but just barely.
A chunk of wire, some scraping on the coating, solder it up, coat it with a bit of epoxy...Good as new!

I'll make a better repair video after I get the parts, and maybe a better camera that'll focus up close a fair amount better.

Note:  Uploaded the wrong video.  The right video is cut up a bit, captioned, etc.  Will handle it later.

[Skimask]

The two links above aren't even close to correct.  They're for the sideways type SMT resistors...mount the long way rather than the short way.  "Wide" type packages.

This is the closest thing I could find at Digikey
WSL2816R0200FEH
.02 ohms, 2 watts, 75ppm, 2816 package.
I measured the original at about 7.4mm x 4.2mm.  This package is 7.1mm x 4.2mm.

Thinking since I'm swapping out 1 of the 3 channels current sense resistor, I should probably swap out all 3 while I'm in there...to make all 3 the same.
What say the masses?
Also going to buy a bunch of them.  Maybe I can hook up somebody else in the future...

[ivan747]

The two links above aren't even close to correct.  They're for the sideways type SMT resistors...mount the long way rather than the short way.  "Wide" type packages.

This is the closest thing I could find at Digikey
WSL2816R0200FEH
.02 ohms, 2 watts, 75ppm, 2816 package.
I measured the original at about 7.4mm x 4.2mm.  This package is 7.1mm x 4.2mm.

Thinking since I'm swapping out 1 of the 3 channels current sense resistor, I should probably swap out all 3 while I'm in there...to make all 3 the same.
What say the masses?
Also going to buy a bunch of them.  Maybe I can hook up somebody else in the future...

I think you should leave the other channels intact. That way you have 2 pristine channels as references if the CH1 fix turns wrong.

But, yes, order about 4 of those so that the Digi-Key guy who that 1 resistor from the top shelf of the bottom of the storage room doesn't go suicidal 

[Macbeth]

Blimey, I wouldn't have thought a polarity reversal during a battery charge could do that to a DP832. It's my second favourite PSU.

Thankfully I use my 6632B as a battery charger and it works a treat no matter what scummy cells you chuck at it. There is something to love about that old VFD beast.

[Skimask]

@ivan - Good thought...about leaving channel 2&3 alone.  And I ordered 20 of them.  Not that I plan on smoking more channels or anything.

@Macbeth - I would've thought the same thing.  Maybe there's a DP832 "reverse polarity input protection mod" in there somewhere.
And, yes, I should've walked the 50 feet over to the garage and grabbed my small battery charger to top off the battery, but, that was over there...and the DP832 was already right in front of me

[Nerull]

Blimey, I wouldn't have thought a polarity reversal during a battery charge could do that to a DP832. It's my second favourite PSU.

Thankfully I use my 6632B as a battery charger and it works a treat no matter what scummy cells you chuck at it. There is something to love about that old VFD beast.

Most lab PSUs - including the 6632B, do not support negative voltages across the output terminals. Bad Things can and will happen. HP recommends putting a diode in the circuit when charging batteries to prevent the crowbar circuit from blowing the internals out of the PSU - it shorts the output terminals together, through the PCB. Also be careful about never turning the supply on or off with a battery connected.

[Macbeth]

Blimey, I wouldn't have thought a polarity reversal during a battery charge could do that to a DP832. It's my second favourite PSU.

Thankfully I use my 6632B as a battery charger and it works a treat no matter what scummy cells you chuck at it. There is something to love about that old VFD beast.

Most lab PSUs - including the 6632B, do not support negative voltages across the output terminals. Bad Things can and will happen. HP recommends putting a diode in the circuit when charging batteries to prevent the crowbar circuit from blowing the internals out of the PSU.

Oh my old war-horse isn't as good as I thought. Well, now I know. Thank you.

I never would have tried charging reverse polarity batteries anyway - that seems pretty retarded - now I know I shouldn't do it ever ever ever.

[tautech]

2 things you can do in an attempt to protect your PSU
Diode in series, but allow for the V_f.
Dirty big Reverse bias diode across the terminals.
Maybe the sparks might alert you just in time. 

[c4757p]

Seems to me it shouldn't be prohibitive for them to put a reverse diode and a bloody fuse on the outputs, and sense the voltage past the fuse...

[codeboy2k]

Seems to me it shouldn't be prohibitive for them to put a reverse diode and a bloody fuse on the outputs, and sense the voltage past the fuse...

Every lab PSU should be done like this, and I don't know why it's not. It does add parts and cost, but it's worth it on a lab PSU. You also have to sense before the fuse too, otherwise the feedback loop goes through the roof trying to bring the voltage up when the fuse is blown.  Sensing before and after the fuse gives you a bonus: a fuse blown indicator too

[Skimask]

I never would have tried charging reverse polarity batteries anyway - that seems pretty retarded - now I know I shouldn't do it ever ever ever.

I didn't do it on purpose! 
I put red on red, black on black....at the battery anyways.
For some dumbass reason, I had them reversed at the PSU jacks.
And I did get sparks, but they weren't that "big" and I figured the battery was just that discharged.

Quick look at the upper PCB shows D6 on channel 1 is a reverse diode.  Have to break out the meter to verify it.  Will handle that later.

[Nerull]

Seems to me it shouldn't be prohibitive for them to put a reverse diode and a bloody fuse on the outputs, and sense the voltage past the fuse...

Probably so, but lab PSUs are generally built to power DUTs, not things that will dump hundreds of amps back into them if the power fails. Battery charging is an instance where getting leads with an inline fuse might be a really good idea. The crowbar exists to protect the device being powered from a failure or voltage spike from the PSU, not the other way around.

[devanno]

@ivan - Good thought...about leaving channel 2&3 alone.  And I ordered 20 of them.  Not that I plan on smoking more channels or anything.

@Macbeth - I would've thought the same thing.  Maybe there's a DP832 "reverse polarity input protection mod" in there somewhere.
And, yes, I should've walked the 50 feet over to the garage and grabbed my small battery charger to top off the battery, but, that was over there...and the DP832 was already right in front of me

If you have the spares, you won't need them.  It's written in the rules somewhere. :-)   

[Skimask]

Resistors showed up today.
Digikey part number is WSLK-.02CT-ND.,  .02ohm, 1%, 2Watt, 2816 package. (manufacturer part number is WSL2816R0200FEH )
Fit just right, work good, last long time, as long as I don't hook any more batteries up backwards.

I was going to do a quicky video on the fix.  Decided against it.  Just so dirt simple, if a person can't figure it out for themselves, probably shouldn't be messing around with it in the first place.

Ordered 20, got 19 left if anybody else decides to let some smoke out of their DP832.

[Skimask]

Well, so much for "last long time".
I thought it was working just fine yesterday.  Ran some power thru a couple power resistors.  Got volts, got current, got heat in the resistors after a bit of pumping a few watts thru them.
Then again...I was a bit tired.  Maybe I was in channel 2 thinking I was in channel 1.  Who knows...

Point is...  Channel 1 is still dead.  Shows "UR" on the display, varying voltage showing on channel 1 of ~.11-.15v, zero for current.
Thought I looked at everything well enough.  Inspected all the traces I could see top and bottom, etc.  Metered and inspected all the parts I could without a schematic, etc.
Guess I get to tear it down again and have another look.
Either that or just buy another "top" board for it and be done with it.

[Skimask]

Bit of an update...

Emailed Rigol looking for some info, parts, troubleshooting, etc.
Got the standard response of "proprietary design", blah blah blah.  Ya no kidding.
I expected something else?

Anyways...and I feel like a bit of a dumbass.  F3 (right next to the J12 input), a 5 amp fuse, is blown.
...DOH...

Thou shalt check voltages...any voltages...the presence or the lack of any voltages.

Guess what?  Blown fuse, no volts.  No shit right?

All I got for PCB mouse fuses is a handful of .3amp pico fuses laying around which were originally meant for rear projection TV convergence chip swaps.
Maybe one of those will work long enough for testing purposes.  Maybe not.
We'll see in a few minutes...

[Skimask]

And not 15 minutes later...

A properly working DP832 with all 3 channels up and operating.

So, to summarize...

A- Wanted to top off a 12v Pb battery with my DP832 (maybe not a good idea in the first place?)

B- Hooked up the battery, red-to-red, black-to-black, but had the leads on the DP832 red-black, black-red (dumbass mistake #1, wasn't paying attention)

C- Smelled the smoked, unplugged everything.  Test and found channel #1 inop, channel #2 and #3 working just fine.

D- Pulled it apart, found R52 (.02 ohm current sense resistor fried), and trace between D6 and positive output lead broken/lifted.

E- Replaced R52, fixed the broken trace with a chunk of 12ga solid copper wire laid over the broken trace, after removing some of the mask of course.  No fix.

F- Profit???

G- Got pissed off at myself for not being able to fix it the first time, took it apart again, and measured the PCB mounted fuses.  Found F3 next to J12 blown.

H- Replaced it with a .3amp PCB mounted fuse for testing.  Worked good.
I- Found and swapped in a 5amp PCB mounted "PicoFuse".  Found a 10ohm sand resistor, loaded it down as best I could, max'd out the voltage across the sand resistor.  Resistor got good and hot (only a 10 Watt resistor sitting in a glass of previously cold water).

DP832 survived.

I'm happy.  How 'bout you?

[tautech]

Yeah well we've all done some dumb shit at times, thanks for writing it all in your closing (we hope)  post.
Just what we all strive for: a happy ending. 

[devanno]

Glad it's working. Congrats and thanks for letting us know the resolution!

The "Profit " line cracked me up... reminds me of an episode of South Park... heh

[macboy]

And not 15 minutes later...

A properly working DP832 with all 3 channels up and operating.

So, to summarize...

A- Wanted to top off a 12v Pb battery with my DP832 (maybe not a good idea in the first place?)
...

I use my lab supplies for charging SLA batteries all the time. They are perfect for this purpose: current-limited, fixed voltage. That's the definition of a proper lead battery charger.

[Skimask]

A- Wanted to top off a 12v Pb battery with my DP832 (maybe not a good idea in the first place?)
...

I use my lab supplies for charging SLA batteries all the time. They are perfect for this purpose: current-limited, fixed voltage. That's the definition of a proper lead battery charger.
[/quote]
True.
Maybe I should have stated I've got a few much cheaper options to top off a Pb battery sitting in the garage next door.

[nctnico]

Better put a diode in series when charging a battery. I got my HP 6002A cheap because someone didn't. The PSU overheated and because there was still more than 0V it decided to activate the crowbar. The owner believed it was beyond repair but the damage was just a few evaporated PCB traces.

[macboy]

I use my lab supplies for charging SLA batteries all the time. They are perfect for this purpose: current-limited, fixed voltage. That's the definition of a proper lead battery charger.

I'll revise my statement to say that the supplies I use do not have a crowbar, but do have a reverse protection diode.

[Skimask]

For future reference in case somebody else blows up channel 1 by doing the same thing I did...(mainly because you can't really order 1 of each of these parts at a half-decent price)

I've got 19 more replacement resistors for R52, and 9 replacement "pico fuses" for F3, sitting in my bins.
If anybody else needs one (or more), PM me and I'll slip them in an envelope and send them out for a few bucks.

[andrija]

Isn't "crowbar" just another (EE slang-ish) name for the diode in series, preventing reverse polarity? That's what I believed for the last 20 years (English isn't my first language though).

[rob77]

Isn't "crowbar" just another (EE slang-ish) name for the diode in series, preventing reverse polarity? That's what I believed for the last 20 years (English isn't my first language though).

no it's not. crowbar is a "trigger-able" dead-short across a supply rail. it's triggered in case of need to protect the load/supply. mainly used to avoid over-voltage - the idea behind is to rather blow a fuse than destroy the load with over voltage.

[SteveyG]

IMO positive should always be on the right. In the UK at least, TTI power supplies are pretty much the defacto standard equipment in most labs which have them this way,  but actually every power supply I've used have the output this way around.

[Howardlong]

Well, I managed to fry my Ch 1 today when showing my nephew what happens when you put an electrolytic in the wrong way round.

PSU to 32V, 3A. 10V 100uF cap, wrong way around. Pop, magic smoke. But PSU starts saying it's overvoltage on Ch 1.

Then nothing, just a weak voltage slowly increasing, no error.

Took it apart just now, F5 was open circuit, replaced with 20mm 5A fuse soldered in. Back to the over voltage error - I measured it, about 50V - oops.

Took aim at the biggest transistor on that board, Q5, a CET CEP80N15. Pulled it off the board, all pins direct short to each other.

Had a look what else I had in stock as N Ch in TO220, all I could find was an IRF510, the Vgs looks alright, Vds a bit lower but the PSU won't be anywhere near 100V. On the negative side, the IRF510 is a mere wiper snapper at 5.6A continuous drain compared to the CEP80N15's 76A. But current isn't everything, it's power dissipation too, particularly in linear regulators like this. The IRF510 is still pretty bad at 43W compared to 300W for the CEP80N15.

Anyway, put it in and all seems hunky dory for now, until I get a proper replacement MOSFET that channel will be on emergency use only.

I am not sure if there is a hardware crowbar on this PSU. The programmed OVP and OCP still appear to function perfectly well, but I am not sure if that is a real crowbar or it just switches the transistor off. Either way, it is a bit of a nasty failure mode, 50V slapped right on the output.

[mcinque]

PSU to 32V, 3A. 10V 100uF cap, wrong way around. Pop, magic smoke. But PSU starts saying it's overvoltage on Ch 1.

I fear that some 832 could be defective: an intentionally reversed electrolytic on the output can (if I'm not wrong) be compared to a short, and I've tried that my 832 can handle that kind of issues at full power (as we all expect)...

What's your serial and calibration date?

[Howardlong]

DPC8C161250xxx
Digital version 00.01.11
Cal Time 2014-3-25

[electr_peter]

Well, I managed to fry my Ch 1 today when showing my nephew what happens when you put an electrolytic in the wrong way round.
PSU to 32V, 3A. 10V 100uF cap, wrong way around. Pop, magic smoke. But PSU starts saying it's overvoltage on Ch 1.

Sad to hear. Was the cap fully discharged?
Cap may presented close to short circuit situation for a PSU. How can a short circuit bring down lab grade PSU? It is a complete fail

[Howardlong]

Well, I managed to fry my Ch 1 today when showing my nephew what happens when you put an electrolytic in the wrong way round.
PSU to 32V, 3A. 10V 100uF cap, wrong way around. Pop, magic smoke. But PSU starts saying it's overvoltage on Ch 1.

Sad to hear. Was the cap fully discharged?
Cap may presented close to short circuit situation for a PSU. How can a short circuit bring down lab grade PSU? It is a complete fail

I would assume it was discharged, I pulled it from a bag of caps that I've had for quite easily 10 years. The end state of the cap was as expected, loud pop, a nice load of smoke, bit of a red flash, plenty of fibrous stuff knocking about in the vicinity with a bit of capacitor offal hanging onto the croc clips. I'd tried a 47uF 16v one a couple of minutes earlier, but that just got a bit fat, I wanted some action, and I got it, just a bit more than I wanted and not quite what I expected.

The more I think about it, the more I think that this is a fail. I took the opportunity to do a firmware update this morning, to 1.13, but irrespective, I won't be trying the same stunt anytime soon.

[mcinque]

I popped a cap with my 832 with a crappy usb charger here using 13.8VDC 3A with no damage to CH1, so I guess it could be a defective unit. Otherwise me too would have a smoked channel, don't you?

[Howardlong]

FWIW, I couldn't find locally a CEP80N15, so I used an STP80NF12 which while slightly lower on the Vds specs (120V vs 150V) is 80A rather than 76A on the Id, max power dissipation is the same, and otherwise spec-wise it appears good enough. Just tested it at full smoke on Ch1, seems fine, but I'm not intending to do a reverse cap test anytime soon.

[giosif]

Sorry to revive this old thread, but I thought I'd add this information for future reference: the 0.02 Ohms SMD current sensing resistor is an Isabellenhütte SMT Series and its size is 2817.
More details can be found here (also attaching the .PDF, in case it gets pulled from the manufacturer web site at some point in the future):
https://www.isabellenhuette.de/fileadmin/Daten/Praezisionswiderstaende/Datenblaetter/SMT.PDF

[tangram]

Thanks Skimask.  Just did the reverse voltage on a partially charged car battery and had exactly the same result.

I think the diode is still intact spookily enough, but the sense resistor, the fuse and the track blowout is consistent.  I'll swap out the diode anyway, 'cos it must have taken a smack, even if it's hanging in there.

Was foxed by the resistor, so glad you were able to confirm the value, otherwise I would have bought a range of 'em and spent hours swapping out and testing.

Cheers 

Followup:

Managed to get everything up and running again.  There was enough of the fried track leading to the diode that, with a little judicious prodding, more or less got it together again and soldereed the two broken sections together.  I guess the bad news is that with 2-3mm of solder bridge here, that section of track will be able to carry a huge load.  Something somewhere else might crap out if I'm stupid enough to reverse polarity a car battery again.   I needed to replace three components:

The Diode (D4?) I replaced with SCHOTTKY DIODE 45V 8A AXIAL - 80SQ045NRLG
Blew a PCB (F3) Fuse, which I replaced with FUSE BRD MNT 7A 125VAC/VDC AXIAL F6123CT-ND

both of the above from digi-key.

The sense resistor (R52) was a pain in the arse to replace.  I found a supplier in the UK who would sell me the part, but he wanted me to pay £99.99 per unit for 5 off, or buy 5000 of the bloody things at a much reduced price.

Found a suitable replacement eventually:  Shenzhen LCSC Electronics Technology Co., Ltd. had the parts.  LCSC#: C466590  Mfr.#: SMT-R020-1.0  $1 USD each, so not too bad, though the UK postage was fierce!

Having fixed the track,  sweated the new resistor in place and swapped out the other busted bits, my DP832 is back up and running.  Needed a recal' unsurprisingly, but it's back in full trim :-)

Thanks again to all the peeps who've added their tuppence worth to aid in the repair.