Author Topic: EEVblog #398 - Lecroy 9384C Oscilloscope Repair (Read 42378 times)

IanJ · « **Reply #75 on:** December 19, 2012, 07:29:43 am »

Hi all,

Reading all this it's turning out to be on of those repairs that HAS to be done irrespective of how useful the final product will be. Pure satisfaction over usefulness......

Ian.

PuterGeek · « **Reply #76 on:** December 20, 2012, 01:46:36 am »

Quote from: EEVblog on December 19, 2012, 06:52:07 am

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.

I'm sure you will find it Dave, we're just trying to be helpfully impatient!

Hopefully we haven't crossed the line to annoyingly impatient (yet).

-Joe

EEVblog · « **Reply #77 on:** December 20, 2012, 02:19:35 am »

I have a few extra tools at my disposal today....

Dave.

Monkeh · « **Reply #78 on:** December 20, 2012, 02:59:54 am »

Those should be pretty revealing..

While you're at it, check the label on that PSU.

PuterGeek · « **Reply #79 on:** December 20, 2012, 03:09:17 am »

Quote from: EEVblog on December 20, 2012, 02:19:35 am

I have a few extra tools at my disposal today....

Dave.

Oooh! Pretty!

And there will be a contest to give away two of them, right?!?

SeanB · « **Reply #80 on:** December 20, 2012, 04:45:08 am »

Now you can see the power planes in the board.

albertr · « **Reply #81 on:** December 20, 2012, 04:48:40 am »

Out of curiosity, I've cranked open LeCroy 9374L scope, but unfortunately the power supply connector has completely different layout (has a single row of pins) as well as main board and ADC chips are not anything like of 9384. I still could try to measure the resistance on 3.3V rail if anyone let me know the pin-out of that power connector. Anyways, speaking of 9374L these ADC chips get quite hot and I wouldn't be surprised if they dissipate some 30-40 Watts...

-albertr

TehPepper · « **Reply #82 on:** December 20, 2012, 07:07:58 am »

Quote from: EEVblog on December 20, 2012, 02:19:35 am

I have a few extra tools at my disposal today....

Dave.

Thats the spirit!

lewis · « **Reply #83 on:** December 20, 2012, 12:25:12 pm »

Quote from: EEVblog on December 20, 2012, 02:19:35 am

I have a few extra tools at my disposal today....

Dave.

Awesome. Teardown! Teardown! Teardown!

JoannaK · « **Reply #84 on:** December 20, 2012, 06:31:18 pm »

I'd expect some spooky videos coming up.

deephaven · « **Reply #85 on:** December 20, 2012, 06:53:50 pm »

What I have done successfully before now is to spray the underside of the board with freezer so that it frosts up, then quickly apply power. Sometimes you can see the hot spot with the rapidly disappearing frost around the fault location. It's a sort of poor man's infra red imager.

robrenz · « **Reply #86 on:** December 20, 2012, 07:22:38 pm »

Nice tip

jerry507 · « **Reply #87 on:** December 20, 2012, 09:22:59 pm »

I can't help but think he's going to fire up those cameras and find 6 hot spots above the asics. It's worth keeping the possibility that all the asics are blown in the back of ours minds...

xDR1TeK · « **Reply #88 on:** December 20, 2012, 11:16:26 pm »

What would be the worst case scenario? Taking the components one by one and testing? even the chips?

Psi · « **Reply #89 on:** December 20, 2012, 11:56:46 pm »

I think there's one thing left to try. Try and run it anyway.

The device was working, it just had memory issues.
So the reason it stopped altogether may have been the 3.3V power supply dying.

If that's the case then it will probably run again with the 3.3V wired to the bench supply.
Also, if the memory issues were due to 3.3V ripple (because the psu was being pushed to far) then using an external psu which can handle the high current may fix the memory issues.

I guess my point is that it may go just fine (even with the fault) if you can supply the current.

My data projector has a very similar issue. One of the 3 LCD drivers has a shorted output pin inside the chip.
Luckily the pin was a simple DC 5.5V output to control LCD bias (brightness adjust) so i forced the pin back to 5.5V and that made the projector work again.
It still has the short and the chip does get hotter than the others but it runs fine, I just cannot adjust the green channel brightness from default.

It's been running like that for 2 years.
The only other mod i did was to add a bigger heatsink to the faulty IC to get rid of the extra heat.

HammerFET · « **Reply #90 on:** December 21, 2012, 12:43:04 am »

Quote

I think there's one thing left to try. Try and run it anyway.

I was going to say the same thing but to be honest if one of those ASICs have gone wrong internally, you might as well not trust the entire chip anymore. It would be riddled with massively dropped internal voltages and the whole thing would be going bananas inside.

It was a bit gutting to see all those other devices heat up when the 3.3V rail went down. I think I saw that IR camera peak one of them to just over 100 degrees at one point

Another problem could be that the drop on the 3.3V rail as soon as it hits the board and spreads out is bad enough to stop things working all together.

Maybe put it back together and set the voltage so that it's 3.3v at the connector. Might just have some life left in it.

JoannaK · « **Reply #91 on:** December 21, 2012, 12:49:36 am »

yeah.. that new video #401 didin't show much promise. When the other chips (5V powered) got heated over 100C at IR camera there's high likelyhood that there's a lot broken on the board. Not just those 3.3V powered Special AD/buffer asics

That said. Taking the hottest Asic out of the board and force-feeding the 3.3V power lines real hard (so that it keeps within 5%), could *in theory* work.. Unfortunately, not likely.

pir · « **Reply #92 on:** December 21, 2012, 01:07:35 am »

D_E_A_D....

long live to us

.....
blog suggestion.....take it apart with A HAMMER

mikeselectricstuff · « **Reply #93 on:** December 21, 2012, 01:18:17 am »

Certainly worth unsoldering that hottest ASIC.

Mike Whitenton · « **Reply #94 on:** December 21, 2012, 01:21:58 am »

I'm sure you called it right Dave. You've demonstrated the ASICs are shorted. Since it isn't repairable at this point, how about sucking them off the board and seeing if the short goes away? That would settle it.

So what do you think the most likely failure mode was on this thing? PSU over volted the 3.3V rail on the board damaging the ASICs. Then the shorted ASICs blew the PSU 3.3V circuit?
If this is the case, there may be 2 components bad in the PSU at this point. One that went open at the tail end and one that let the voltage go above spec.

It certainly isn't worth opening the PSU other than to satisfy curiosity about what started the failure.

Unrelated:
What's up with one of the 5 brown boxes on the underside of the board near the BNC connectors? Looks like there is a hole blown in the corner. Not related but just curious.

Monkeh · « **Reply #95 on:** December 21, 2012, 01:24:06 am »

Quote from: Mike Whitenton on December 21, 2012, 01:21:58 am

Unrelated:
What's up with one of the 5 brown boxes on the underside of the board near the BNC connectors? Looks like there is a hole blown in the corner. Not related but just curious.

It's a relay. He broke it during teardown.

Mike Whitenton · « **Reply #96 on:** December 21, 2012, 01:31:53 am »

Quote from: Monkeh on December 21, 2012, 01:24:06 am

Quote from: Mike Whitenton on December 21, 2012, 01:21:58 am
Unrelated:
What's up with one of the 5 brown boxes on the underside of the board near the BNC connectors? Looks like there is a hole blown in the corner. Not related but just curious.
It's a relay. He broke it during teardown.

Don't know how I missed that. Thanks! It's been bugging me.

EEVblog · « **Reply #97 on:** December 21, 2012, 01:40:54 am »

Quote from: Mike Whitenton on December 21, 2012, 01:21:58 am

So what do you think the most likely failure mode was on this thing? PSU over volted the 3.3V rail on the board damaging the ASICs. Then the shorted ASICs blew the PSU 3.3V circuit?
If this is the case, there may be 2 components bad in the PSU at this point. One that went open at the tail end and one that let the voltage go above spec.

My best guess is that it's a two stage failure.
There was something screwy with the 3.3V rail which was causing the issues that we saw, and then it finally whent kaput and took out the ASIC chips.
Although given that I never measured the resistance of the 3.3V before it died, we will never know.

Yes, I'll probably desolder the 2nd ASIC for kicks.

Dave.

jerry507 · « **Reply #98 on:** December 21, 2012, 03:09:19 am »

So now that it's obvious the asics are busted, remember back to the first video when you were looking at the display and 3/4 of the display were solidly drawn and only the right quarter of the display worked? Kinda makes sense now doesn't it?

peter.mitchell · « **Reply #99 on:** December 21, 2012, 08:18:29 am »

Quote from: EEVblog on December 21, 2012, 01:40:54 am

Quote from: Mike Whitenton on December 21, 2012, 01:21:58 am
So what do you think the most likely failure mode was on this thing? PSU over volted the 3.3V rail on the board damaging the ASICs. Then the shorted ASICs blew the PSU 3.3V circuit?
If this is the case, there may be 2 components bad in the PSU at this point. One that went open at the tail end and one that let the voltage go above spec.

My best guess is that it's a two stage failure.
There was something screwy with the 3.3V rail which was causing the issues that we saw, and then it finally whent kaput and took out the ASIC chips.
Although given that I never measured the resistance of the 3.3V before it died, we will never know.

Yes, I'll probably desolder the 2nd ASIC for kicks.

Dave.

I'm surprised you didn't freeze the ASIC with the dust buster, eg, freeze it before power on (really freeze it and wait so the chip gets cold all the way through) and see if the current was different?

That 100+ degree device? I immediately though probably powered between the 3v3 and 5v for 1.7v, but when you take out the 3v3 and it is shorted to ground it gets a full 5v - probably bullshit, but hey, it makes sense to me so i don't care what you think!


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: EEVblog #398 - Lecroy 9384C Oscilloscope Repair (Read 42378 times)

pir

Share me