Oh yes. I meant to add: It was you that gave me the courage to seek down the reasons why a "junked" LCD screen went poop. 2 blown Caps and 50p later for new ones from our local Leccie shop cured it! Ripper! - and that's from a London Pommie. Thank you.
Keep at it Dave! You will have it narrowed down in time. With every session of troubleshooting you seem to make progress. I'm thoroughly enjoying watching this repair series.
Hi dave You did pretty much did what I would do so far(man I enjoyed that). Voltages, then frequency, touch next. You know for all those errors there is a common theme. Perhaps another rail, ripple, 15 year old electros? Tantalums normally go s/c, so not that.
I love learning how you tackle these faults, it gives me more knowledge and things to try when I have issues aswell. I know the feeling where you replace stuff and it still doesn't work right, but on the plus, it was completely dead, you got it to turn on again and now it's not got as much wrong with it! Good stuff, can't wait to see what else It could be
Please don't give up on it Dave. We all want to see you outsmart the thing. One day Sagan will watch these videos and you don't want to disappoint him do you? There, is that enough guilt yet?
No, I've easily tracked down and found three faults so far with no circuit diagram. The circuit diagram will make it easier in the long run for sure, but there is still a lot I can do without it that isn't "shooting in the dark". My main issue right now is (lack of) PCB access, not the schematic.
I was rooting for you on this one... unfortunately, you can't fix them all on the first try - maybe next time :) But it's always fun trying - especially when you finally get it working - it makes it all the more worthwhile and satisfying! Great video.
Usually, Tantalums are my first suspects that I measure and lift legs on.. compared to normals elcaps they have one annoying specialty: they tend to fail to a short circuit which not only messes up your further measuring, but can also easily take out your recently fixed power supply again...
I think the main problem is an input overvoltage (measurement input). All the things you found is related to analog circuitry. Maybe the user exceeded analog input voltages blowing up the input stage, then, this stage fail short to the power rails, and then blow up the regulators. Good luck.
Love these vids. Please keep the the process, successes, and fails coming. Very nice to see the whole process. Make me feel better for all my fails ;) Cheers Sir
Here is what I do in this situation. Get a large schematic or board layout of the system printed. put it on a dart board and throw 5 darts, check the parts or circuits that the darts land on.rinse and repeat.
But the power supply does not give out more (or less, how do you say that for a negative supply) then about -19 Volts. The maximum voltage across the opamps will have been -19 to +15 Volts, or 34 Volt in total. That could be stretching it for the opamps, but many have an absolute maximum rating of +-18 to +-22 Volts with a nominal of +-15, so depending on the part, they could have survived. Also, I don't think the opamp really cares if it is symmetrical, so I'm guessing they're all fine.
The negative rail going "high" probably wouldn't have mattered if it was single rail, but being split rail you have to take into account the positive rail and the total rail to rail voltage which was over 34 volts and many opamps (at least the 324) can only handle 32 rail to rail
In a production lab we pull boards and swap till it solves the problem. The original suspected board then goes in for repairs. Allowing the unit to go back in service during the repair.
When you say "ADC Gate Array" I keep hearing "ADC Gatorade". Nice sport drink! Anyways, I really hope you'll be able fix that Murphy-possessed bugger! Waiting for Part 3...
As you had mentioned structural corrosion in part 1 of this series, I've got this impression that the original power supply went mustang, causing power surges and spikes, hitting just about every sensitive component in this unit... I think you are going to be chasing after more faults and ghosts in this hunk of steel.
love you repair videos and your soldering tutorials had lots of great info and showed me that though i havent been doing it long i have been doing it wrong, keep up all the great videos and great score with the plotter found a cheap one online for myself! cheers!
All of them, it seems that there will be at least -35V on them according the specs as the specs tells you that that is the max voltage this regulator can handle. Total voltage over the OP-AMPS power lines was at least 50V in total. Most OP-AMPS can't deal with that amount of power. Hope to see a follow up on this.
Cool video. I admire your skills and thanks for the teachings. Looking forward to finding out if u can fix this bugger or if it will ultimately be too much of a pain in the rump.
In the first video you pointed out that a rubber foot or cushion had melted and spread over the top board that would take a lot of localised heat. Have you managed to track down the heat source yet ?
No, since he´s in the right track, but it at least depends on the ASIC healthy.... I´ve repaired some synths successfully without having any kind of documentation for it. But when you know how it works and its main electronical structure as well then it´s possible. But if it´s someones first repair ever than it would be quite hard...
I know the feeling. Spoiled by the easy fixes. With some experience in most cases reparing stuff comes down to really simple problems. But maybe 1 out of 10 "patients" drag you deeper with every step you make. Hardest thing is declaring something beyond repair after pouring hours into it.
I feel for ya Dave. I hate it when you think your sure you've tracked something down then get that sinking feeling when power-on doesn't go as planned :( There's always next time though. Great video.
not necessarily. the clipping LEDs being hard on gives some hope that it might just be an opamp (or many opamps...) that have failed and are railing their outputs.
you know.. the input power from the lab supply seems awfully high. ~ 90Watts. Im guessing the opamps are fried( maybe even the ASIC). The next step should be to test the op amps out to see if they're shorted out.
The benefit of this is that he can keep getting videos out of the repairs, so there's less incentive to write it off as uneconomical to repair. Proprietary/custom parts being broken will stop things, but otherwise it's a potential video.
Maybe look for a component that heats up. Mostly faulty semiconductors heats more than normal when powered, as it is known to fail into bad short-circuit.
A good test a this point would be to check if any of the out of scale rails connect to that ASIC, cause if its failed its probably beyond economical repair. On a other note, the op-amps could probably tolerate that -19V, a lm741 can take +-22V but an tl081 cant take over +-18V. If its +15V -19V the voltage across the op-amp is less than 36V and probably wont kill it. But don't take my word for it, maybe the spec sheet of the op-amps say it cant take over 30V total or something.
I'm wonder how high was voltage on these rails during power supply failure. It's not so easy to kill -15 V regulator... But shorted regulator means -19V on ASICs so ASICs probably died too... Maybe they're protected. It should be in such expensive instrument, but who knows ?
The other observation is that you moved the LO varicap. Has that changed the freq of the I2C? Was it 400kHz or 1MHz? Was it that time critical that can't be measured by 1 significant digit of a DSO?
I think, the ADC has been blowed away. Since if it´s rated +/-15V it may tolerate 10%or so and -19 are much to low...As i know that many synthesizer hardware (with curtis chips) dies because these chips rated +/-15V and do NOT tolerate +/-16V. Likewise some of the OpAmps witch tolerate +/-18V are dying when one rail goes beyond 19V.
if you want thermite you need iron oxide and aluminium. not aluminium oxide.. the aluminium takes the oxide from the iron which releases loads of energy, so you end up with aluminium oxide (white solid stuff) and liquid iron :P
Hi, Could it have damaged the ADC? I'm fairly new to electronics but Dave's videos really help learn. I know he always says that the USB scopes are pants, but does that include the Hantek and Rigol ones. I'm on a real budget and have most parts but a scope! Would be neat to have it all PC based (I'm going to build a USB controlled signal generator), any thoughts?
I'm betting on opamp or other analog failures. The common 32V PS differential limit would have been busted with a +15 and a -19.x. Look for an opamp with a railed output. I'm betting that will lead you to the input issues.
It is a tad worrying, as we will never know how bad the PSU went to screw up the pessimistic regs yet knowing the diode let what ever it was through, Grrr. Would the gate arrays rely on the I2C? Hazarding a guess, I am no genus. I won't bug you about the the time circuit anymore, whilst your in repair mode, however, if you get it working we could just pop back in time, just before this unit failed and hook up some test leads to see what happened ;p. All the best
Dave, toss the solder wick and get a Hakko 808 solder SUCKER... It really SUX the big one. I use mine for pcb work and even point to point stuff. Best $200 spent in a long time.. Check youtube videos for Hakko 808... No, I do not work for Hakko, just love this product
But now we have... sorry Dave has working diagnostic procedures. and overload lights continuously... I think that's the point something is wrong on the inputs, and that's now good point to continue
If there an input overload my gess is to check the input section, probably they transistors that can be shorted or amplifier (strange because the -19 is on adc board and not on the channel boards). The instrument uses dac that could fail if too much voltaghe gon in it. The adc gate array wish is not gone away. I think that the gate array error is linked to the i2c error. What ther error logs said thi time? Some error go away?
Oh yes. I meant to add: It was you that gave me the courage to seek down the reasons why a "junked" LCD screen went poop.
2 blown Caps and 50p later for new ones from our local Leccie shop cured it!
Ripper! - and that's from a London Pommie.
Thank you.
Keep at it Dave! You will have it narrowed down in time. With every session of troubleshooting you seem to make progress. I'm thoroughly enjoying watching this repair series.
I am mesmerized by your analytical approach. Great to see a maestro at work - "Lead on McDuff!"
PART3? A cliff hanger . . .
Hi dave
You did pretty much did what I would do so far(man I enjoyed that). Voltages, then frequency, touch next. You know for all those errors there is a common theme. Perhaps another rail, ripple, 15 year old electros? Tantalums normally go s/c, so not that.
This is getting better than most TV series I watch. I felt fear, then anger at some point, I even cried a little, and now it ends on a cliffhanger.
I love learning how you tackle these faults, it gives me more knowledge and things to try when I have issues aswell. I know the feeling where you replace stuff and it still doesn't work right, but on the plus, it was completely dead, you got it to turn on again and now it's not got as much wrong with it! Good stuff, can't wait to see what else It could be
Please don't give up on it Dave. We all want to see you outsmart the thing.
One day Sagan will watch these videos and you don't want to disappoint him do you? There, is that enough guilt yet?
Really pleased when I saw this one come up on the subscription! It is turning in to a gripping serial!
No, I've easily tracked down and found three faults so far with no circuit diagram. The circuit diagram will make it easier in the long run for sure, but there is still a lot I can do without it that isn't "shooting in the dark". My main issue right now is (lack of) PCB access, not the schematic.
I was rooting for you on this one... unfortunately, you can't fix them all on the first try - maybe next time :)
But it's always fun trying - especially when you finally get it working - it makes it all the more worthwhile and satisfying!
Great video.
i would assume that an expensive instrument would have surge protection on the rails to prevent stuff like that on power supply failure.
I'm getting one shortly! Need an extension cable or card to get the board out of the rack though.
The Actel FPGA is only powered using 5V so it should be ok if nothing shorted the -15v to one of the digital lines.
Usually, Tantalums are my first suspects that I measure and lift legs on.. compared to normals elcaps they have one annoying specialty: they tend to fail to a short circuit which not only messes up your further measuring, but can also easily take out your recently fixed power supply again...
I think the main problem is an input overvoltage (measurement input).
All the things you found is related to analog circuitry.
Maybe the user exceeded analog input voltages blowing up the input stage, then, this stage fail short to the power rails, and then blow up the regulators.
Good luck.
Cracking fault finding... Extremely useful . Cheers Dave.
Love these vids. Please keep the the process, successes, and fails coming. Very nice to see the whole process. Make me feel better for all my fails ;) Cheers Sir
Here is what I do in this situation. Get a large schematic or board layout of the system printed. put it on a dart board and throw 5 darts, check the parts or circuits that the darts land on.rinse and repeat.
I used that model a lot in school, brings back some nice memories. Hope you can get it back to live.:-)
Safe to say the rail that failed and fried things from the PSU was the -18V rail? Notice the CH 2 Half LED flash every time you tapped the BNC? :o
Don't give up, Dave!!!
Agilent would have debug extension cards to make this easy. Not sure it's worth doing that or hacking some sort of DIN cable
But the power supply does not give out more (or less, how do you say that for a negative supply) then about -19 Volts. The maximum voltage across the opamps will have been -19 to +15 Volts, or 34 Volt in total. That could be stretching it for the opamps, but many have an absolute maximum rating of +-18 to +-22 Volts with a nominal of +-15, so depending on the part, they could have survived. Also, I don't think the opamp really cares if it is symmetrical, so I'm guessing they're all fine.
Love these repair videos Dave, keep em coming!
The negative rail going "high" probably wouldn't have mattered if it was single rail, but being split rail you have to take into account the positive rail and the total rail to rail voltage which was over 34 volts and many opamps (at least the 324) can only handle 32 rail to rail
In a production lab we pull boards and swap till it solves the problem. The original suspected board then goes in for repairs. Allowing the unit to go back in service during the repair.
Sure would be nice to have a schematic. Great video and looking forward to the next on getting this thing fixed.
When you say "ADC Gate Array" I keep hearing "ADC Gatorade". Nice sport drink!
Anyways, I really hope you'll be able fix that Murphy-possessed bugger! Waiting for Part 3...
Can't wait for Part 3!
As you had mentioned structural corrosion in part 1 of this series, I've got this impression that the original power supply went mustang, causing power surges and spikes, hitting just about every sensitive component in this unit...
I think you are going to be chasing after more faults and ghosts in this hunk of steel.
I run my own business, I can work any 15 hours a day I like...
love you repair videos and your soldering tutorials had lots of great info and showed me that though i havent been doing it long i have been doing it wrong, keep up all the great videos and great score with the plotter found a cheap one online for myself! cheers!
Never seen that myself, there were 3 others that looked perfectly fine.
All of them, it seems that there will be at least -35V on them according the specs as the specs tells you that that is the max voltage this regulator can handle. Total voltage over the OP-AMPS power lines was at least 50V in total. Most OP-AMPS can't deal with that amount of power. Hope to see a follow up on this.
We don't know what's happen during first failure, what kind of power supply failure was and what was repaired by HP...
Cool video. I admire your skills and thanks for the teachings. Looking forward to finding out if u can fix this bugger or if it will ultimately be too much of a pain in the rump.
In the first video you pointed out that a rubber foot or cushion had melted and spread over the top board that would take a lot of localised heat. Have you managed to track down the heat source yet ?
No, since he´s in the right track, but it at least depends on the ASIC healthy....
I´ve repaired some synths successfully without having any kind of documentation for it. But when you know how it works and its main electronical structure as well then it´s possible. But if it´s someones first repair ever than it would be quite hard...
I know the feeling. Spoiled by the easy fixes. With some experience in most cases reparing stuff comes down to really simple problems. But maybe 1 out of 10 "patients" drag you deeper with every step you make. Hardest thing is declaring something beyond repair after pouring hours into it.
looking forward to part 3.
I feel for ya Dave. I hate it when you think your sure you've tracked something down then get that sinking feeling when power-on doesn't go as planned :(
There's always next time though.
Great video.
How close are the fires to where you live and where you work?
not necessarily. the clipping LEDs being hard on gives some hope that it might just be an opamp (or many opamps...) that have failed and are railing their outputs.
Also you could try to adjust the -19v rail to -18 correctly but it couldn't be too much difference.
Perhaps, but which ones? There are probably 30+ opamps in this thing!
Used them in the past and like 'em, but the Hakko 808 retails for nearly $500 in Au.
Try the good old law of the opamp: "The input's shall have equal voltages." (if in feedback config). That should narrow it down a bit...
If it gets too expensive I think a little aluminum oxide and iron oxide mixed in the proper ratio would be fun!
Don't know if you noticed, but those 'half' LEDs came on when you tapped the connectors at 20:15. Thought that was interesting.
you know.. the input power from the lab supply seems awfully high. ~ 90Watts.
Im guessing the opamps are fried( maybe even the ASIC). The next step should be to test the op amps out to see if they're shorted out.
The benefit of this is that he can keep getting videos out of the repairs, so there's less incentive to write it off as uneconomical to repair. Proprietary/custom parts being broken will stop things, but otherwise it's a potential video.
Dave, the half rail led came on cause your finger brushed the BNC connector.
Have you tried turning it off and on again?
Maybe look for a component that heats up. Mostly faulty semiconductors heats more than normal when powered, as it is known to fail into bad short-circuit.
Love these repair and debugging videos. Would really like to see this sucker running :)
Thank you Dave for another good video
can't wait for the next video :)
If I remember correctly it was just the 240v ac side that died, but it could still be powered by a DC input.
How did you clean the melted rubber of the board?
these repair videos are alway interesting, keep on with this one Part 3?
Hi Dave, didn't you forget to replace regulators on the second board?
A good test a this point would be to check if any of the out of scale rails connect to that ASIC, cause if its failed its probably beyond economical repair.
On a other note, the op-amps could probably tolerate that -19V, a lm741 can take +-22V but an tl081 cant take over +-18V. If its +15V -19V the voltage across the op-amp is less than 36V and probably wont kill it. But don't take my word for it, maybe the spec sheet of the op-amps say it cant take over 30V total or something.
I'm wonder how high was voltage on these rails during power supply failure. It's not so easy to kill -15 V regulator... But shorted regulator means -19V on ASICs so ASICs probably died too... Maybe they're protected. It should be in such expensive instrument, but who knows ?
I did notice that those half LEDs turned on when you were touching the channel connectors. Not sure if that means anything.
The other observation is that you moved the LO varicap. Has that changed the freq of the I2C? Was it 400kHz or 1MHz? Was it that time critical that can't be measured by 1 significant digit of a DSO?
I probably got too much entertainment out of the ribbon cable clunk at 4:03.
I usually pull the ASiCs and reseat them after cleaning the socket.
MORE!! this is gold
I think, the ADC has been blowed away. Since if it´s rated +/-15V it may tolerate 10%or so and -19 are much to low...As i know that many synthesizer hardware (with curtis chips) dies because these chips rated +/-15V and do NOT tolerate +/-16V. Likewise some of the OpAmps witch tolerate +/-18V are dying when one rail goes beyond 19V.
if you want thermite you need iron oxide and aluminium. not aluminium oxide.. the aluminium takes the oxide from the iron which releases loads of energy, so you end up with aluminium oxide (white solid stuff) and liquid iron :P
Have you checked the remain board the analog processor board coul also use the -18 rail that as now seem the dead power line....?
Hi, Could it have damaged the ADC? I'm fairly new to electronics but Dave's videos really help learn. I know he always says that the USB scopes are pants, but does that include the Hantek and Rigol ones. I'm on a real budget and have most parts but a scope! Would be neat to have it all PC based (I'm going to build a USB controlled signal generator), any thoughts?
Beautiful...just....beautiful.
I'm betting on opamp or other analog failures. The common 32V PS differential limit would have been busted with a +15 and a -19.x. Look for an opamp with a railed output. I'm betting that will lead you to the input issues.
It is a tad worrying, as we will never know how bad the PSU went to screw up the pessimistic regs yet knowing the diode let what ever it was through, Grrr. Would the gate arrays rely on the I2C? Hazarding a guess, I am no genus. I won't bug you about the the time circuit anymore, whilst your in repair mode, however, if you get it working we could just pop back in time, just before this unit failed and hook up some test leads to see what happened ;p. All the best
I think this as well, but this had been hapened when the first (the original) powerSupply went wild....it took the whole machine with it, too.....
Great vid, I look forward to the next part =)
Dave, toss the solder wick and get a Hakko 808 solder SUCKER... It really SUX the big one.
I use mine for pcb work and even point to point stuff. Best $200 spent in a long time.. Check youtube videos for Hakko 808... No, I do not work for Hakko, just love this product
How can the unit still be working when regulators just let the current pass trought ?
Btw , nice video , good work as always !
Is it worth checking the data sheet for the op-amps in there to see if they can survive the over voltage?
powered from the DC power jack which still works.
Awesome video dave!!
Great vid Dave!
But now we have... sorry Dave has working diagnostic procedures. and overload lights continuously... I think that's the point something is wrong on the inputs, and that's now good point to continue
I hope you make more videos like this. :)
Well, it makes it that much more interesting, Dave! ;-) I like these "find the culprit" vids! :-D
I'd start checking every op-amp on the -15v rail. They're probably mostly fried too.
Try freeze spray and build a layer of frost. Poor man's FLIR and knock a hole in the ozone.
Great Video! You'll get it working again I'm sure.
Damn it I need to pick up some snacks before the next video is ready.
also apart the fpga the intrument seem to not use hybrids!
If there an input overload my gess is to check the input section, probably they transistors that can be shorted or amplifier (strange because the -19 is on adc board and not on the channel boards). The instrument uses dac that could fail if too much voltaghe gon in it. The adc gate array wish is not gone away. I think that the gate array error is linked to the i2c error. What ther error logs said thi time? Some error go away?
i know you can fix it dont give up untill the end
Now onto HP35670A - Part 3
Of Course, you´re right here, for sure! :)
Half was blinking when you touched the BNC ... ... broken ground? not sure if thats normal
But maybe make a rig and test all the opamps?
he said "lets connect all the cables" just before the test, so i think he did that
Or maybe something is shorted to -15V on inputs side opamps perhaps?
Still can't help but wonder what melted that rubber standoff...
LOL! "Another 5 MINUTE video." Since when have you EVER completed a video in anywhere NEAR 5 MINUTES!! :-)