EEVblog
ฝัง
- เผยแพร่เมื่อ 26 ต.ค. 2013
- More progress (or is it?) on the HP35670A DSA repair.
Part 2: • EEVblog #538 - HP35670...
Part 1: • EEVblog #536 - HP35670...
Forum: www.eevblog.com/forum/blog/eev...
EEVblog Main Web Site:
www.eevblog.com
EEVblog Amazon Store:
astore.amazon.com/eevblogstore-20
Donations:
www.eevblog.com/donations/
Projects:
www.eevblog.com/projects/
Electronics Info Wiki:
www.eevblog.com/wiki/ - วิทยาศาสตร์และเทคโนโลยี
I would trace back from the source output and see why it's putting out +15v, there's definitely an opamp shorted to the positive rail.
Whether you fix it or not I've learned a lot from this series about troubleshooting.
i hope you continue the series , this has interested me since the first one
7:05 Remember this DSA was designed for the northern hemisphere. Ofcourse your voltages are upside down.
Love this series :) even if it's proving a bugger to fix, I've learned much more about troubleshooting than any other videos out there!
This is the best adventure I've seen in years! It is almost like archeology! Those few that can fix it are scattered throughout the world and only Brave Dave (C) takes the challenge of reviving an ancient relic! Best part is that it is not a fantasy fiction, IT IS REAL! There are no cheat codes or walkthroughs. We get real challenges and real emotion. Also, about 10 to 20 thousand people on a search of a forgotten schematic in hope to save HP35670A DSA unit. GEEK PORN
Each time getting closer, let's hope ya get it this time!
love your videos. i love everything that has to do with electonics and computers, and how all the parts work together/
I love these troubleshooting videos.
Don't give up! Just solve the problems one by one; don't focus on the system, focus on the sub circuits. Fix the source output first, then move on from there. Step by step and you'll get there - eventually. I've seen worse than this before and prevailed, sometimes it takes longer than we'd like but the feeling of accomplishment at the end makes it all worthwhile. Identify problems, solve them, move on to the next problem. Repeat as necessary. That's how it's done.
Love your videos, keep up the good work.
I can say, i am learning alot from this video. Being a beginner I am learning alot about trouble shooting. Please consider doing more trouble shooting videos.
That ADC waveform is telling you something- think that, unfiltered, it would be a stepped sinewave between + and - limits. If you still have a -ve reference that is out, you may well see something like that... I think you're very close now! DAC -> anti-aliasing filter -> 'source' output driver: the end is in sight! Hopefully there will be a part 4?!
The opamps are connected in parallel to the power rail, not series. Opamps are generally on a separate supply to digital stuff, so hopefully this means that the ASIC hasn't been blown!
I hope you can fix the Signal Analyzer, even though from the looks of it, it dosent look good. I hate it when countless things are possibly dead with little to no way of figuring out what.
"Will Agilent send you the schematics? Did you ask them? You ARE a pretty big name they're bound to cater to."
I too was thinking if you asked them that they would probably do so, its no harm in just asking.
Hope you can fix the DSA and good luck (you probably will need it).
This is turning out to be a classic case of "be careful what you wish for" hehe.
I used to work in a electronic repair shop, and often didn't have diagrams for stuff, so you have to just hunt & peck around like Dave is doing here to find the trouble.
Watched your just now. I hope you have a "circuit tracker," which what I recall ages ago. Its a small scope where you point your probes on the circuit and the scope will show some sort of signature.
Another great vid!
Really like the series :-)
Doggy dew , I did not realize we were up to part 3 , and I had seen it , bad memory ,
Now waiting for part 4 :)
I can understand Dave quite well. I have a tek TDS540 Scope i've been working on for some time now. After fixing all the faulty caps on fthe boards, there is still an error persistent. And now I don't know wether I should keep on trying or just get rid of it as it is... Dave, I hope you can find the motivation to keep on this nice piece of electronic...
Yes, troubleshooting is pretty much hopeless without easy board access. I need an extender board or cable, or else this on is going nowhere.
Hard luck Dave. If you give up on the fault-finding (and I wouldn't blame you), maybe you'll find some replacement boards on eBay sometime, or pickup another faulty DSA that you can strip for parts. I enjoyed the series anyway mate, keep us updated if you have a Eureka moment.
I'd measure resistance from the VRef test points to the output pins on nearby op-amps. The reference voltage rails are probably driven by op-amps, either directly or through a low resistance. Then use visual inspection or continuity testing to trace the circuitry around the op-amp and try to trace back to the voltage source. Luckily it's all through-hole with everything in plain sight. A schematic would be great but a bit of reverse-engineering would be worthwhile for a piece of gear like that!
Another great video! It will be a shame if you can't repair it but it seems logical (as you've said, and others) that the likelyhood is that numerous opamps will have failed. Still, very interesting. Would love to see more repair videos =)
i like these repair series videos
Have to say this, its been very interesting see you try to hunt down he problem and fix those found step by step... Quite interesting to see a fault that's not easy to repair for a change.
Sure would like to see rep of manufacturer to step up and diggup you one of those schematics. As someone who does small odd repair jobs now and then i do preciate the availability of schematics.
Hang in there Dave, I hope you can source a schematic somehow/where. I'd really like to see a final fix and the "look what this puppy can do" video !
I think only ic's conneted to +/-15V could be damaged by the shorted diode. The stuff powered by the +/-6.2V couldn't because the regulator doesn't provide an higher voltage than nominal.
I think that there are only few parts that need to be replaced. The true problem is find the spare parts if it's castom ic.
The connection of devices on a power rail is like the connection of home appliances in your house. In parallel. If instead of 110 you have 200V it will blow them all. What you said could happen if the failure mode of one of the devices was shortcircuit but usually it's not.
It's great that it's not fixed!!!! This means more videos for it!! :D
Most likely probing is better then replacing all of the chips.
Like comparing output on the opamps (well not that clear data, its in circuit) but perhaps some have no output at all or very unusual values.
Perhaps also trace the swapped -/+ voltages as much as possible - perhaps the block diagram is correct and it has swapped due to some unusual failure somewhere and causing all the problems.
we want to see more videos... you cant stop there dude!
I, too, find the troubleshooting process interesting and educational, regardless of the outcome.
I would check out some of the op-amps in the source-output area and see if there is any signal there. Then at least you have an idea how extensive the damage is...
Dave, did you check the buffer input right at the buffer? Also what happens if you connect a capacitance on the funny 1Hz signal? Is it possible to work it out from the connectors in?
Check the ADC input pins. Seams one or more are failing, hense the bad sinewave on the output. If it was me i'd make a test circuit on a testboard and get all of them opamps out and tested and than replaced by sockets. How many layers is the board?
Could it be possible that the change of the powersupply in any way messed up the + and - 15V Rails???
I second the motion. The worst they could do is say no. Maybe if you point them to the three videos, they will cough up the schematics out of shame.
i wonder is it possible to force a signal into it so that it thinks its own source is working to see how far it gets, doesn't sound simple though, especially without schematic.
you may not have fixed it yet, but you sure made a good video :) thumbs up
It appear more supply stuff.. Maybe some reference voltages for the adcs.. Replace all chips!
I think I would jumper in external + and - 6.5 volts to a representative op-amp with the board removed and look for offsets. Might get lucky.
I think you hit on it at the beginning. The HP procedure was to replace the board level stuff not component level. Somewhere on that board is probably an HP Part number. Just search around on that.
Agreed a little bit of a chore..and cost, but considering the reward. At least do a 'measure out' on the ICs to see what the results are and go from there. You don't need a schematic for that.
In most cases you can test an opamp without undoing it. Just get some clips that fit DIP package and probe +IN -IN and OUT. Not a 100% way of testing them, but rather fast.
Also it may be easy to trace signals from DAC. You have already found that it gives some output from the filter. It is very strange, but probably You can find some chip that gets this signal to it`s input and has +15 on it`s out.
Dont' all the LF412 have a common voltage feed from the 15volt rail? Cant you test the failed op amp and see whats shorted and test the other Lf412 for the same problem either a short or open? I sometimes test for resistance between leeds on chips between a known good one and a bad one. Another thing is if you have multiples of the same chips you can compare resistances between leads on different chips.
Yeah I would probably also try to replace them... Problem is that the ADC/DAC is is probably also shot and they are often expensive to replace. I would try anyways :) I would even buy a desoldering station just for that ^^
Repair thriller!
Also the signal output looks like there are common mode noise.
Hi Dave,
Too bad you couldn't find the problem. What exact type of extender do you need? I have some various that I don't use anymore Mostly 300mm europa of varying pincounts. Leave a comment if you need one.
Hope someone finds the schematic, would like to see this get further!
Easiest way would be to find someone near with a similar unit. Exchanging boards would at least point out the defective one(s) for a start.
Shame it wasn't repairable, still great videos.
Can you use sig. gen. and scope as signal tracer to check opamps?
Short the output to ground for about 30 seconds. Quickly kill power and remove the board. Feel for the hot chip.
Why does the "Half" LED light up when you push on the CH2 connector? Bad connection, hint at something further?
As for op-amps, the pinouts should be pretty typical -- doublecheck any numbers you can identify to be sure. Check if they are linear or saturated: if the input voltages are equal (within a mV or so) or different.
Looks like data is all kinds of upside down though. Not a good sign for the gate array...
Was the oscilator readjusted after the first video? Maybe the frequency setting is causing problems on a serial bus or something.
Looks like you have sprayed some cold spray @ 11:43 location U-120 in small video format : )
Given the state of the DAC output, I'd drop a logic analyser on the DAC digital input and plot that (X-value vs. time or sample number) and check that the input to the DAC is sane.
I wonder if you've got a failed DAC or a TTL buffer or similar upstream which drives the DAC.
It's not inconceivable that the DAC output being wrong (in other words: the test data) could cause the selftest to fail.
I'd also get an edge connector and put together an extender card. Would help immensely with debugging.
To see if an op-amp has failed could you could measure the inputs, if they are equal the op-amp is fine, if not the op-amp is kaput or not in feedback mode. The latter should be easy to rule out with an quick look. :)
You sure you got the reference right when measuring the test point? Cause some ADC gnd reference is NOT signal gnd but ADC gnd for various design requirement
Bit hard to say without schematics.
I really feel more anxiety about a new episode of this repair series than what i used to feel when i was following House M.D.
I'm no engineer, but I would take a look an see if there is any traces of that grey goop on the board connector pin sockets.
The absolute maximum voltage rating on the negative rail of the LF356 (too bad they aren't the LF356B, their absolute maximum rating is higher), AD845J and the LF412 is -18 volts. Most likely all of these are gone.The DAC 702 can handle an absolute maximum of -22 volts. The MC14053 would most likely have been powered by the +15 volt rail. So it probably survived. Not sure of the voltage rating of all the bypass caps. Those three legged tantalums can be a problem. I'd still keep plugging away.
He already checked the power supply pins in one of the previous videos.
I am not an EE but a software developer for embedded systems so my knowledge of electronic troubleshooting is very spotty...especially analog side of things but I have a question.. If one of the power rails had a surge of some kind that took out the op-amp(s) can you briefly explain to me how it would be possible to take the majority of them out as well? I thought as soon as it destroyed one in the signal path wouldn't it prevent the surge going further in the circuit and spare the rest?
Thanks
Your HP calibrator friend cannot help with scematics or panel? Ask him. But check the power and the reference clock of the DAC. One time I got same output noise in one DSP card. I fixed with a cap and one zenerdiode replacament. Thumbs up!
The designer of this product HP35670A DSA may be enjoying watching Dave troubleshoot. Otherwise he would have helped by now. Anyway I love watching too.
what is a 'Dynamic Signal Analyzer" (this HP thingy) actually?
(I know I could google it, but I doubt everybody knows that so...)
That looks to me like you have a stuck latch or data line feeding the DAC.
could the issue be with the actel chip also ? :o
mm yeah toasty opamps. It's a shame they didn't socket all of them :) The funny looking 1hz sine wave looks like it's missing it's MSB - not sure what's going on there.
Did you ever get this thing working?
Call me pessimistic but I don't see the schematic helping tremendously. Odds are that's not the only bad board either so I don't blame you at all for stopping at this point.
if he made a flex cable from board to case so he don't have to mount the board to power it up i don't see a way he could do that
a peek under a hand held thermal image device maybe
Like that pioneer amp... Look for the thing looking u right in the face haha. You should start referring to fluctuating voltages as bouncing around like a fart in a bottle. Always made me chuckle to hear my old man say that ;)
Your question on the Agilent Forum has been rerouted to the power supplies section. It's not an Agilent power supply I would say. Whatever...
can some one help to find solar power garden light problem?
it does not light up, battery are new contacts looks clean
light only flash one time when turn switch to on, when solar sensor is cover
or when is not cover and then cover it flash one time
1. I haven't accused him of anything.
2. Somehow I had in my mind that the review for it was not really that good. Hence I was confused somehow.
But it seems he was happy enough with many things of it - still.
Dave, Use a FLIR thermal camera to detect which op-amps are non functional.
Faulty transistor most likely, or the LED.
OK, I deleted the question. I was just confused a little bit. He obviously just likes that thing after all.
Too bad you don't have the schematic. I really hoped to see it working this time :P
Enumarate all op amps. Make a simple C++ random generator and begin to replace.
Whit a pair IC leg hooks he could try quite a few at a time. :)
isn't there a way to test op-amps on the board? pushing a brown ball uphill with a stick....lol...
Get an intern to swap out all the op-amps :) They work for free and get the experience in return.
I'm sure at this point the schematic must be somewhere on Wiki Leaks, lol. C'mon Agilent give it up! Sorry Dave in, case NSA was eavesdropping on your otherwise excellent blog.
I searched the internet like hell and can't find a schematic to help u. It's gonna take a different tactic. Like emailing someone who would work at HP and possibly have access to the data and be willing to share because it is outdated tech
Crikey! a service manual and no schematics! what are HP playing at!?
I guess there is some truth to calling HP products HoPeless
Please Dave repair this sucker, it deserves to have a second round. Pleaasseee 00
What's your point? Dave gave it a good review too... If he used the 87V in every episode would you accuse him of being sponsored by Fluke?
You seem to give up to easy. Changing 20 dip ics is maybe 2-3 hour job and cost of maybe 50$. Not too bad for expensive dsa. No schematic is problematic, but little bit of reverse engineering could go a long way:)
This is a good series. BTW even if you can't fix the thing, it has in no sense been a waste of time. The idea is to make good vids, which you've done, not to repair a DSA.
What if I don't have an uncle Bob? :P
Not front panel but an internal
Is it just me or do I see ATTEN at 12:28 above TP7
If you want some stuff for repair videos I have a lot of audio equipment that you can have that is not working.