EEVblog
ฝัง
- เผยแพร่เมื่อ 15 ธ.ค. 2012
- Dave attempts to trace a short "real-time" on the main board of the Lecroy 9384C oscilloscope.
Teardown video is here: • EEVblog #217 - Lecroy ...
Forum Topic: 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
One technique I've used is to freeze the whole board until it turns frosty white, and then apply power to low resistance rail while watching which portions board are the first to defrost.
Dave you need to do that trick WITH the FLIR looking at the board that'd be awesome to see!!!
Surely four-wire resistance measurement would have been more reliable here? I don't necessarily mean with a real four-wire multimeter, just feeding 0.1A into the 3.3V rail and finding the component with the lowest voltage at its 3.3V point? This would remove the reliance on sharp leads, and inconsistency due to probing varying oxide thicknesses, as the time variations as various caps charge and discharge in response to the probes.
Hi Dave! I've noticed lots of dust between IC's. I've had once a situation where small piece of metal was mixed with dust causing short in circuit. After applying some air the damn thing just worked fine. Cheers!
Dave, instead of a better ohmmeter one can use a millivolt meter. Just connect a current limited supply to the shorted rails.
Soren Kuula watch the video next time. He does that after probing.
Adam Besmer watch the video next time and also read the 1st sentence in Sorens comment.
Yes Dave attached a current limited supply but no he did not use a millivolt meter to find the highest voltage drop.
BTW, he does in part 2.
I love your troubleshooting videos. I learn something new everytime I watch your videos. I am in the electronics repair field and find your videos a invaluable resource for information as well as inspiration for advancing my career. Cheers from Canada
YAY! I've been waiting quite some time for this video- I love your repair videos, they're very, very interesting- even if you don't quite get it to work!
Oops. Forgot to add.
When you're cranking in those 11A, measure the voltage across those tantalum capacitors. A low voltage would mean a short nearby.
Dave,we really enjoy all of your video
That static is something out of a horror film. Like a ghost doesn't want you to repair the scope.
This is outragously educational. Thanks for doing this.
I saw a video of a guy finding a short using alcohol, he brushed it over the board and it evaporated first from the hot component. Neat trick. Love your videos. 30+ watts is a lot of energy. Also, my intuition is that if those chips are heating the short might be down circuit from them or they'd have very low voltage.
the problem is, that the psu turns off automatically and with such a delicat device its not a good idea to use that method
I really enjoy your repair videos, more of these please :)
12:45 *Melted looking relay towards the bottom of the board.*
13:10 and 13:55 "[No] obviously blown components."
Even if it has nothing to do with the 3.3v rail, still kinda funny XD
Saw that I was like nooo, Look harder!
I was literally screaming at him!
I right away saw this too. Body burned and the relay really sure will not work anymore :D
This gives me great heart knowing that I have not fixed electrical item yet. So even expert like you failing kinda makes me happy!
This is where a thermal imaging camera comes in really handy. That's what we use for checking "mystery" boards straight out of manufacturing for layer-to-layer shorts.
You should add this to your “repair” playlist!
loved it dave
Thank you for the fantastic lesson. You should do more lessons in fault location and methods. Cheers, again.
Could you try the different way around, heating the board/chips and watch for a increase in resistance of the 3V rail ?
I popped the IR filter off a CMOS camera once and could see a soldering iron glow, but yeah, that's still a couple of hundred degrees hotter than you'd need in this application.
great i am camerounian and i am watching you are channel step by step it very nice what you are doing
If you have an IR-thermal camera, you could use that to look at the board as it is being powered to see where the power is being dissipated.
It is a lot of extra power to be dissipated, so I do certainly suspect the heatsunk ASICs in some way.
I'll have another go at it tomorrow morning...
every time someone says "murphy's law" i think of my actual friend murphy showing up and breaking things.
SWEEET! 43mins of awsome! :D
The point being reading the voltage not the low resistance you don't have to worry about meter lead resistance, the tips to component connection resistance, how hard your pushing on the parts, etc near as much
Fantastic! I have a very similar short on a 12 volt rail on a layered board. Your video helped me realize that my meter doesn't measure low enough resistance to help me find a problem like this one. Are you going to make another video showing that you finally found the short? I also need a better power supply that will go over 10 amps. Confused as to why the short is not heating up, same as my board.
Few days ago I wondered what had happened to this LeCroy, and was going to make a post on the forum asking about it. Murphy:)
I would suggest to feed 0.1-0.5V into 3.3V rail, so that no semiconductors are switching on (they usually start "doing something" after 0.6V) so that only non-semiconductor short is conducting, and measure voltage drop at different points on the PCB.
Fixed or not, this was still a good informative video for diagnosing a bad component.
For the schematics you can try the schematic of the 9374M I think this units are similary. Here there is the namer of the manual to the manual: LeCroy_9374M-L-TM_Service_Manual.pdf
Pretty crazy for sure. Pulling 11 amps when the PS can only supply 6 makes it sure seem like something is wrong and gone to ground. Since the chips all got warm, is there an additional regulator that supplies all of them, or some other common component? Schematic with voltages or resistance measurements sure would be handy right now!
Nice video! thumbs up! I think you will have to remove the heatsinks of these devices and power them up to see which one is the faulty.
Cameras sensor are sensible to near infrared only, which is the kind that can be used to "see" IR controls and see in the dark (with an IR flashlight), the "thermal" infrared is waay apart from the visible light.
Anyway, if any component gets "red hot", maybe you can see it with your eyes, in the dark.
Not sure what the test voltage of your DMM is on the low ohms range. I'd be tempted to measure the short in terms of semiconductor junction drop.
The ASIC's may draw that much current if nothing is clocking.
"... and went KAPUTT!"
as someone who speaks German I laughed way to hard. :)
Hi, thanks for the video. Can you please explain what you are looking for while probing with the Ohm-meter? I don't understand it. Thank you.
Don't know if it would work, but cameras are more senstive to IR than we are. Try hooking up to the high power supply, turning out the lights making the room as dark as possible and looking through the camera screen. It's a shot in the dark (pun, sorry!) but might help.
Great video! When trying to blow up the suspect component aren't you relying on the short being the weakest point? If the short could handle 11A but some other component couldn't, wouldn't you get smoke from the wrong place?
I found this page knowing only what I could remember from my Physics II course I took 3 years ago. Now every day I try to research more and more about electronics so I can better understand your videos. Any books or other material you can recommend for me besides simply searching the internet for more info?
I have a Kef Kit160 and makes a buzzing sound but had it open and can't find exactly what in the power supply is making the buzzing sound. And just because that is buzzing don't mean it fault could be something using more power than it should.
Could you feed in the 11amps then use the non-contact current probe thingy you reviewed a while back to find where all that current is going?
11:10 Between the two P63AD chips, directly in line with the PSU connector, there is a surface-mount component that looks like it sprayed brown gunk toward the PSU connector, covering some traces on the PCB. Maybe that component went "sizzle-pop" and the gunk shorted?
Nice video! Quite strange that at 3.2V it only draws 11 amps. It doesn't make sense since the resistance is only 0.1 ohms!
Voltage should drop on a place where short is present.
When you're applying power supply it sinks more current than LeCroy power supply limit due to short.
This will work only if short has 0 Ohm resistance.
Maybe 10-11 amps is it's normal current consumption? Seems high but not too unreasonable considering there's 4 ASICs on it. Check if the PSU is dead or if it's indeed current limiting and shutting down the 3.3v rail.
You don't thutch with the probe, you just hold close to the device. If there is a dc current the resistance in the probe will go down and you must set the referance on your instrument
Thats a real tough one with it taking the 11A for an extended duration. The extra 5A has to be going somewhere, so probably not into something small like a cap. Perhaps photonicinduction has a beefy enough power supply to find/pop the short
Hi can you recommend the best testing device , fluke etc for iphone motherboards ???
Ever try feeding the shorted buss with a PS limiting the current to a max 1 amp (and 3.33v) and looking for the lowest voltage across the component? The short in this case should limit the voltage to well below what the good ASICs start to turn on. 1 amp .16 ohm .16 volts.
Also I have found shorts like this in the past by spraying a board area with chiller till its really cold and frosts up, Apply power and look for what defrosts first. Those vortex chillers that run on compressed air work well.
I'm not too familiar with Lecroy scope internals, but would the damage to the large brown package in the analog front end on the left-most channel (with BNC's facing forward) have something to do with it? Can be seen in video at 12:37 to 12:50 as others noted.
I would have thought it might cause some problems with that channel even if it's not the cause of the short.
The Scope is not a purely resistive load, and the leads and connectors do factor in, too
Unless some spike on one of the other rails/or the 3.3V rail triggered it, that is.
I assume they are quite much connected to the same places (just different "port number" for digital stuff), so one fault could possibly have killed them all.
Hi, after 30min I'm thinking PSU on it and some IR visual, and controlled current increasing to see where does it go to pinpoint. And then you said it yourself. Maybe it is not so complicated to do after all. And IR camera like that can help you in your next troubleshooting. You said also that your PSU can deliver 20 some Amps, but the board "only" draws about 10. You measured 0.15 Ohm. With 3.2V I would expect 21 Amps.
this is a shot in the dark but...does anyone have an idea how much a used one (working) of these sells for? I've looked around online but can't get a good feel for it.
If you still have the AIM-TTi I-Prober 520 Current Probe as seen in video #296, maybe that could help localize the fault?
One tip is to use an IR (FLIR) camera... faulty circuits pop up like a sore thumb,
have used that method sucessfully several times.
..and a little while later,I see you tried this method!
(That's what I get for commenting before watching the entire video.)
Looking at the layer marks on the PCB, is there any significance to the two being inverted?
The PSU is where l look. Its rated less then 35amps output given that wire size of 18 or 20ga at the MB connection, if not it would have melted a long time ago. Unless the 3.3 splits the load.
The AIM-TTi I-Prober 520 Current Probe might show the route the 11A are following. If you've still got it, that is...
Would a current probe be able to trace out a short like this in a multilayer board?
I had 2 Stratix IV's on a 0.9V rail, that read way
Nothing like measuring the leakage of 0.1billion low voltage mosfets in parallel. Gets you every time :) Good story, I ran into that very thing once as well. Had to pause and think it through.
At 12:44 is that a crack on one of the imput relays because that might be whats wrong with the jumping half of the trace
saw a relay package with melted hole at about 13min...anything to do with that!!
An interesting approach could have been to use a thermal imaging camera. You could hook up an external high current 3.3V power supply, turn it on for a second or two and see where the board gets hot. Crude, but it could work if you accept the possibility of blowing up the board.
I only have a casual interest in electronics and this is the first time I've seen a multi layered board.
Does someone sit there and design it all, or does a computer do it?
Why not try to boot it with your heavy duty psu + the original one, 3.3 from the high power and the rest from the normal. To see if it does boot or not ?
BTW cheapest thermal imaging cameras are like 1500$ (60x60 resolution), which is somewhat affordable - so it's slowly falling out of 'fancy equipped' category.
Hi Dave! What strikes me is that the ”blow the crap out of the faulty component” method didn´t work. I dont think that anything that wasn't designed for that could have survived 11Amps. Maybe you should look closer at the power supply! My gut says that the board is ok... :)
what happens if you bypass the 3.3v rail on the main power supply, then use the high amperage power supply to power that rail (I have pulled 15 amps from a 12 amp 12V rail on a computer power supply). Or if possible, power the 3.3V rail then use the fluxgate magnetometer to see if any of those components that you were concerned about are behaving differently as compared to all of the others. (the short point should have different characteristics)
One of the brown (relays ?) looks a bit weird, have you looked at it?
why dont you use the AIM-TTi I-Prober 520 to find out where the current goes?
With that much current, it might be a short on one of the connectors?
Ha, I was wondering the other day what you did with it. I couldn't recal a repair.
Wouldn’t it be better to limit it to 1-2A and use isopropanol to see where it heats up?
Is it possible the 3v3 rail is ok, and the psu internally is not giving a 3v3 rail. The 11A draw looks about right for all those high current flash converters, as they have multiple parallel resistor parts inside. I would suspect the power supply if the asics are all getting toasty, but all are same temp. 11A on a 3v3 rail is not that much, I see them on servers capable of 40A or more and it is needed. Dying cap in the psu can time?
That does seem to be a bit hot. The asics might be using all of the power. There should be an extra 15 - 20 watts being burned and a cap would be smoking if it was burning that much power.
One of the maroon boxes appears to have a hole melted in it
Question, is a digital 'scope the way to go when in converting from analogue signal to digital necessarily results in the loss of some information in the process of digitising the signal especially when dealing with high frequency signals.?
Yes. Because high bandwidth analog scopes are getting rare (they have not been made for decades now!!), and you need to have plenty of equipment and understanding to maintain them. So if you feel like setting up a one-off scope cal/repair lab then sure - get yourself a historic Tek 7k mainframe with high bandwidth continuous time and sampler plugins and you’ll be set. But in general, digital scopes lose less than analog ones and I’m serious here. The “low noise” of an analog display is just smoothing done by the display process itself. It will hide stuff that will be seen easily on a digital scope. The loss of information is not relevant if the scope is used appropriately. Scopes never were high precision instruments. They are calibrated indicators pretty much. If you want highly accurate recordings of the signal, a scope is never the instrument you want. You need a dedicated digital sampler/recorder with appropriate bandwidth and memory. And that usually is only called for in specialized applications, and if it costs less than $5k/unit you’re looking in the wrong place pretty much. For general purpose use the “loss of information” happens to analog scopes. The digital ones are unforgiving and record everything (as long as - again - you’re not mis-applying them). Generally you want a modern scope with good reviews and lots of memory so you can always sample at full rate. I have had lots of success with the Siglent models.
3.2V at 11A is 0.29 ohms. That's really a really low resistance, even for a short at the end of a trace. You might have a poorly soldered DC socket.
Why don't you power the whole thing up with the 3.3V rail working with the high current supply and all the other rails with the main supply. Something could have daisy chained down all those ASICs and killed something in them. The scope might still work but with a certain function going wrong.
the brownish kind of relay near the bnc at 13:52 have a hole in it. look melted no?
What time of the video was the shown bad part Emmeran?
I agree, but I just could not resist.
Hey friend!
It seens have high current. You can actually get a single head probe that will tell you were the big current is goning.
Well, Ilike you channel very much.
Use the amp probe on the 3.3 volt, see if the PS is the fault or the board.
Sorry mate did not mean to offend was kinda joke. To make me feel better about keeping with electronics. I love your show! Keep up good work.
I've got a method to find your short,and possibly even remove the bad part from the board,all at once!
3.3V at 100A.
There can be some unintended casualties though.
(Not afraid to admit,I've actually used this method before,successfully even!)
The relay may be an unrelated issue. Something blew the ass out of the relay on channel 1
12:38 dave what is that???? the bottom red/brown it looks like a blown componet
The probe detect "big" dc current. You set the threshold.
All these comments about what was obviously wrong and no one mentioned that it was 240 degrees C on the bench?
thermal camera Dave, see it you can get one used, broken or something like that
No, nothing to do with it. Inadvertent damage during the teardown.
I have BWD POWERSCOPE 881A (McVan Instruments) please can you tell me where to get its repair manual please its urgent.
Would suggest Upgrade to the new Metrix Power scope from Chauvin Arnoux.
In India Cyronics in Pune can help you. rahulcyronics@gmail.com
Hook up the 3.3 V and look for a minimum voltage.
By the way, did you brake Murpy's mirrir in the past lol