Author Topic: EEVblog #398 - Lecroy 9384C Oscilloscope Repair  (Read 42148 times)

0 Members and 1 Guest are viewing this topic.

Offline Electr0nicus

  • Regular Contributor
  • *
  • Posts: 131
  • Country: at
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #50 on: December 17, 2012, 11:32:22 pm »
Maybe the worst possible thing happend. Maybe there is a intermittent short between the 5V and 3.3V rail. As the 3.3V rail is weaker than the 5V rail, it would have easily brought the voltage on the 3.3V rail up to nearly 5V.  If that is the case then this 3.3V chips would have been fried internally. This would explain the increased current consumption, and nothing else except these 8 chips with  heatsinks getting hot.

Just another possibility to think of.

Cheers Gregor
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6680
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #51 on: December 17, 2012, 11:42:24 pm »
What about the PSU being mislabelled? "1" rubbed off, is actually 16A?

I mean, 11A at 3.3V seems high, until you consider that over 8 ICs, that's only 4.5W each... which could be pretty normal for high performance devices, they certainly have a good heatsink and would normally be air cooled.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37664
  • Country: au
    • EEVblog
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #52 on: December 17, 2012, 11:48:25 pm »
What about the PSU being mislabelled? "1" rubbed off, is actually 16A?

It is scanned from the service manual and is very clear, 6A.

I'm amazed at how many people seem to think that 0.1ohms both directions on a multimeter could possibly be normal on a 3.3V system rail  :-//

Dave.
 

pir

  • Guest
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #53 on: December 18, 2012, 12:33:53 am »
What about the PSU being mislabelled? "1" rubbed off, is actually 16A?

It is scanned from the service manual and is very clear, 6A.

I'm amazed at how many people seem to think that 0.1ohms both directions on a multimeter could possibly be normal on a 3.3V system rail  :-//

Dave.

stop posting and fix that damn old horse !!!...
we are all waiting the part 2
 :phew:
 

Offline Razor512

  • Regular Contributor
  • *
  • Posts: 156
  • Country: 00
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #54 on: December 18, 2012, 12:53:46 am »
While not an oscilloscope, I tested 4 different old ATX motherboards and the in going from ground to the 3.3V rail, they were all around 3.2-3.8 kohms

The only system that was lower was a non ATX design compaq system from the early 90.s, the 3.3V rail was at 7.5 ohms (it still worked the last time I used it but have not used it in many years.

I wonder what if you power the device suing the stock power supply but then bypass the 3.3V raiil on the power supply and instead use your other power supply to power that raid, will the device still work?

A power supply can give above it's rated power. (eg I have an old 300 watt power supply that is rated for 11 amps on the 12V rail, but I am pulling at least 40-15 amps from it (though it is running no other load on any of the other rails (I use it to power an old car radio/ sound system since the system can do 200 watts at max volume (connected to 4 speakers, and I can max it out and none of the lights dim on the radio or the amp. (it is a cheap $12 chinese ATX power supply that came with a $20 ATX case that tiger direct was selling (and also had free shipping :) )

I have been using it for a while with no problem (I keep it away from anything flammable just in case since I do have it wired up in a sloppy fashion)

(I bet if I wanted, I could push that china power supply even further above it's limit (not sure at which point it would start smoking, but what ever the real limit is, what ever protection it has (if any) has not kicked in yet :) )

edit: Pictures yay (has been had been used in a PC for a year back in 2003, then retired until about 2 years ago for use with a speaker system)
« Last Edit: December 18, 2012, 01:53:47 am by Razor512 »
 

Offline Greg323i

  • Contributor
  • Posts: 29
  • Country: au
  • Resistance is futile! (If less than 1 ohm)
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #55 on: December 18, 2012, 12:54:55 am »
Hmm, suspecting the power supply is an interesting one, but consider this. If 11A is actually a normal current for this board, that means EACH 3.3V pin on that connector is carrying 6A a piece. That sounds a bit much, don't you think?

Yes, that is why I don't question the manual's mention of it being a maximum of 6A capable 3.3V rail. 6A per pin does not make sense.

Dave, didn't you say that you don't have the exact service manual for this unit? Could it be that the unit you have actually has a higher capacity 3.3V rail?
 

Offline Dread

  • Regular Contributor
  • *
  • Posts: 122
  • Country: us
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #56 on: December 18, 2012, 01:03:58 am »
Mike beat me to it in that I knew you were wasting time with using that meter and trying to find milliohms of difference, that’s futile

No, it's not. This is a very common and successful technique to quickly locate shorts. It is also safer than powering the board up, so is a common first option.
It just happened to have not worked in this instance.

Quote
and compounded by the fact that that you say  the meter has a high resolution but as you stated it's accuracy is in question! So at those resolutions it begs the question what is it's repeatability of reading at the same point twice or more, it seemed to be really poor.

Watch the video, the repeatability of the meter and probes is excellent.

Quote
Which goes to the point that a lack of accuracy does not mean it will always be off by plus X or minus X but can drift all over the place.

The accuracy of the meter has nothing to do with repeatability, that's not how it works.

Quote
My second piece of advice is don't be afraid to use that exacto Knife and cut a surface trace or two.  I use to do that when the voltage ran all over the board.  I could then isolate which side had the problem.  A second or third cut can isolate it down to a very small exact area

Once again, a last resort method. And something that is not possible with multi-layer boards.

Dave.

Dave I never had any real success doing it that way, you may get lucky but on a complex board like that one it's a long shot.

Cutting the board can be done when the 3.3V line comes up to the top or bottom, it will often snake around the place then go back down to another layer.  You can cut it at the surface when it snakes around and that will isolate a section of the board.

I must confess that my feeling is that the power supply is the issue and not the board. I suggest you do what some previous posters have said and use your external source for the 3.3 and put the unit back together and see if it works.   The fact that you had odd behavior before makes me think the 3.3V was going south from when you did the Teardown and now it has failed.
Yes it defies the power consumption you expect but I have a feeling that consumption is going to drop once you hook up the other rails.


The Optimist says the glass is half full, the Pessimist says its half empty, an engineer only see's a glass that’s twice as big as it needs to be!
 

pir

  • Guest
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #57 on: December 18, 2012, 01:16:23 am »
hey people...you are missing the point here...

3.3V pinout:
pin1 ----> good
pin2----->sorted gnd
pin3----->sorted gnd

as the man said in the video they use many pins to get the watts in...
with your logic pin1 is powering something else say low watts...and pin2,pin3 is the super-duper 10+ amps
 :o
 

Offline Electr0nicus

  • Regular Contributor
  • *
  • Posts: 131
  • Country: at
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #58 on: December 18, 2012, 01:21:24 am »
As Dave and some other forum members pointed out, that 0.15 ohms on the 3.3V rail in both directions can't be the normal state. You can think through it back and forth, it is just not possible in my opinion. So the real question to ask, is not if the powersupply can deliver 6 or 16A, but why the resistance is the same in both directions. To measure the resistance the multimeter has to apply a low voltage across the 3.3V rail, so if the voltage is reverse polarity (probes reversed) than there should be at least a little bit of change in the resistance value (due to reverse biased semiconductor junctions). But in this case it is exactly the same, disregarding tolerances.  So my gut feeling says that there is maybe something fishy going on in those heavily heatsinked devices, because no other device (Chip) on this board can withstand that high power dissipation for a long time without emitting magic blue smoke. The only devices which can be faulty without actually blowing up or bursting in flames when x times the rated current is passed through them are those heatsinked devices.

I know I sound very pessimistic, but that can be a very propable scenario.

I would also recommend powering that 3.3V rail separately via the lab supply. This way Dave may be able to rule out a defect of the ADCs before digging deeper into the problem and maybe wasting hours for nothing.

Cheers Gregor
 

Offline Greg323i

  • Contributor
  • Posts: 29
  • Country: au
  • Resistance is futile! (If less than 1 ohm)
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #59 on: December 18, 2012, 01:30:14 am »
My suggestion about the service manual was mainly because I found it odd that the resistance you measured and the current that you measured don't correspond.

I watched this video again and I agree with the people that think that screw hole is probably the cause of the short, as it doesn't look deformed. Easy enough to check resistance between pin 2, 3 and the screw hole.

[edit]

I'm an idiot...

Of course you're going to measure the same short anyway if that screw hole is grounded.

[/edit]
« Last Edit: December 18, 2012, 02:17:46 am by Greg323i »
 

Offline Monkeh

  • Super Contributor
  • ***
  • Posts: 7990
  • Country: gb
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #60 on: December 18, 2012, 01:51:29 am »
What about the PSU being mislabelled? "1" rubbed off, is actually 16A?

It is scanned from the service manual and is very clear, 6A.

I know the service manual (didn't you say it's for a different model?) says 6A, but.. is it just me, or does the label on the PSU in the teardown video at 19:37 say 3.2V 13A?

Quote
I'm amazed at how many people seem to think that 0.1ohms both directions on a multimeter could possibly be normal on a 3.3V system rail  :-//

I know, it smells funny to me, too. Still, if that label says what it appears to..

E: Here, have a screen cap.
« Last Edit: December 18, 2012, 01:56:47 am by Monkeh »
 

Offline Anks

  • Frequent Contributor
  • **
  • Posts: 252
  • Country: gb
    • www.krisanks.wordpress.com
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #61 on: December 18, 2012, 03:23:17 am »
Get your AIM-TTi I-Prober 520 Current Probe out and see if you can find the short. Used to do this sort of thing with a HP 548A current tracer. Should point to the places that are eating the current without having to resort to thermal imaging. I know its not really a run on the mill everyday joe type of tool but if its at hand it would be nice to see it used.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37664
  • Country: au
    • EEVblog
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #62 on: December 18, 2012, 03:32:30 am »
My suggestion about the service manual was mainly because I found it odd that the resistance you measured and the current that you measured don't correspond.

On the contrary, you would expect something like that with the difference between 11A and naff-all.

Dave.
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 8973
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #63 on: December 18, 2012, 04:25:24 am »
Connect a function generator across the rail and set it for some low frequency like 46Hz. (Keep it distinct from the mains frequency so you don't get confused. In fact, if you keep it near the mains frequency and then trace it with an audio method, you'll probably get a very distinctive beat pattern.) Then you can use a magnetic field probe without getting false readings from magnetized bits of metal. And it's possible to use a MP3 player instead of a function generator and a hacked tape player instead of a magnetic field probe.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline G7PSK

  • Super Contributor
  • ***
  • Posts: 3859
  • Country: gb
  • It is hot until proved not.
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #64 on: December 18, 2012, 04:57:41 pm »
What is limiting the current to 11 odd amps as at 0.16 ohms the draw should be 20 odd and as that power supply is able to output 40 amps something should have given, is there a current limiting device on the board, I did think I saw a TO20 device in the region of the 3.2 volt area Dave was looking at. Also as Monkeh pointed out the power supply does look like its rated to 13 amps.
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #65 on: December 18, 2012, 06:31:20 pm »
The video from this LeakSeeker 82b page gives a good idea what Dave was doing except with much higher resolution and audio feedback.

All the videos HERE

The one on the link to the site
« Last Edit: December 18, 2012, 06:55:36 pm by robrenz »
 

Offline KedasProbe

  • Frequent Contributor
  • **
  • Posts: 646
  • Country: be
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #66 on: December 18, 2012, 07:55:53 pm »
What is limiting the current to 11 odd amps as at 0.16 ohms the draw should be 20 odd and as that power supply is able to output 40 amps something should have given, is there a current limiting device on the board, I did think I saw a TO20 device in the region of the 3.2 volt area Dave was looking at. Also as Monkeh pointed out the power supply does look like its rated to 13 amps.

Dave did you check the voltage on the board connector?
There will be loss over your cables but we don't know how much.(we assume not much but if calculations are done it has to be right.)
Maybe you can try a voltage below 0.7V, see what it gives.
« Last Edit: December 18, 2012, 08:00:08 pm by KedasProbe »
Not everything that counts can be measured. Not everything that can be measured counts.
[W. Bruce Cameron]
 

Offline Mr Smiley

  • Frequent Contributor
  • **
  • Posts: 324
  • Country: gb
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #67 on: December 18, 2012, 08:29:51 pm »
And how to do it with a battery and multimeter  :-+



 :)
There is enough on this planet to sustain mans needs. There will never be enough on this planet to sustain mans greed.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37664
  • Country: au
    • EEVblog
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #68 on: December 18, 2012, 09:21:27 pm »
What is limiting the current to 11 odd amps as at 0.16 ohms the draw should be 20 odd and

It's the heating up of the short (and changing of it's resistance) due to the current. This is quite common and generally expected.

Dave.
 

Offline G7PSK

  • Super Contributor
  • ***
  • Posts: 3859
  • Country: gb
  • It is hot until proved not.
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #69 on: December 18, 2012, 09:35:17 pm »
If the short is in one of those chips then the heat sinks could keep the short from blowing out forever then. So what happens if the board is frozen would the current go up enough to blow the short.
To paraphrase Enrico Fermi, I used to be confused but now after watching the video I am still confused, only at a higher level.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6680
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #70 on: December 18, 2012, 09:49:43 pm »
Freeze spraying components could help in locating shorts.
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 3686
  • Country: us
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #71 on: December 19, 2012, 05:28:41 am »
I definitely prefer the "pass constant current  from a supply through the short and trace it down with a volt meter" method to using 2-terminal resistance measurement.  Using an ohmmeter may be a standard technique, and certainly can work, but I think it is categorically inferior.  Saying that you are more likely to damage your board is a red herring.  You can use whatever current you want.  Ever meter I have seen has the same or better resolution on mV as ohms.  If you use a small current on the order of what your multimeter uses, you will get the same resolution.  You have the option to crank up the current for more sensitivity, but are in no way required to.

Even with good sharp probes, you definitely had to fuss around to make sure you were getting a good contact.  With a 4 point measurement, this is essentially not a problem.

If I were really frustrated or just being careful, I would use two meters.  One connected right near the power supply leads that would show changes in the resistance due to mechanical changes, and the other that I would move around looking for the fault.
 

Offline PuterGeek

  • Regular Contributor
  • *
  • Posts: 88
  • Country: us
    • SolutionsPLUS engineering
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #72 on: December 19, 2012, 06:11:59 am »
The video from this LeakSeeker 82b page gives a good idea what Dave was doing except with much higher resolution and audio feedback.

Too bad they don't make those any more.

The fourth paragraph from the bottom of the page said the last production run was in January and they have none left. They do say they might come out with a new design.

I'm guessing some part(s) they used were obsoleted.
 

Offline PuterGeek

  • Regular Contributor
  • *
  • Posts: 88
  • Country: us
    • SolutionsPLUS engineering
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #73 on: December 19, 2012, 06:41:46 am »
I'm amazed at how many people seem to think that 0.1ohms both directions on a multimeter could possibly be normal on a 3.3V system rail  :-//

I agree it would be strange but the symptoms aren't really adding up.

It was only drawing 10 to 12 amps and you didn't find a hot spot. I was very surprised when the bench supply was able to maintain 3.2V and apparent hotspot.

When I run into strange situation I try to think outside the box. I try to eliminate anything I can, starting with things that are easy.

Verifying the 3.2V supply is OK with a visual inspection and load test would be quick.

Since bench supply can manage the current, using it for the 3.2V and powering up the scope might show a fault in the POST.

If not, you will be able to use differential voltage measurements which as several people have said, should be more accurate.

I would also try a cheap infrared thermometer to make spot measurements.

Then would be pulling chips if you want to keep at it. In particularly I would focus on the devices with heat sinks (I'm guessing they are DSPs). Since they're getting hot, removing them would keep them from masking a short or they could be the short.

I would probably scrap it if all that wasn't fruitful. That is unless I 'needed' to get a short tracer then this was a good excu reason to buy one now.   ;D
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37664
  • Country: au
    • EEVblog
Re: EEVblog #398 - Lecroy 9384C Oscilloscope Repair
« Reply #74 on: December 19, 2012, 06:52:07 am »
I agree it would be strange but the symptoms aren't really adding up.
It was only drawing 10 to 12 amps and you didn't find a hot spot.

That's because I haven't found it, yet!
It will be there. It's almost certainly not evenly spread across all the ASIC's. So you can't beat the laws of physics.

Dave.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf