diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   Effect of Supply noise on oscilloscope bandwidth? (http://www.diyaudio.com/forums/solid-state/148567-effect-supply-noise-oscilloscope-bandwidth.html)

Craig405 4th August 2009 08:20 PM

Effect of Supply noise on oscilloscope bandwidth?
 
Hello,

I have created a power supply for a LeCroy 9354 Digital Oscilloscope using pc supplies as they are powerful and cheap.

The scope boots up and works seemingly well until I turn the timebase to 50uS or less then a relay switches inside for the new time range followed shortly by the reboot of the scope.

I have heard that giving the scope noisy supply rails will limit the available bandwidth of a scope such as this (500MHz, 2Gs/s).
Id like to obtain close to the original spec from this puppy if possible.
Before going too far down the line of repair id like to know if this is the cause and what the most appropriate solution is?

Edit: One more thing can somebody please provide me with a link or PDF of Jungs' articles on "Regulators for High-Performance Audio" they appear to have been removed from the web?
Many thanks.

error401 4th August 2009 11:28 PM

They seem to still be up at Walt's site: http://waltjung.org/Library.html

Click Regulators & References and look halfway down.

Can't help you with your poor hacked scope though :(

Craig405 5th August 2009 12:25 AM

Thanks, I feel stupid now, kept getting page error for this site so I assumed it was taken down. Its fine now!
:xeye:

RocketScientist 5th August 2009 12:39 AM

Re: Effect of Supply noise on oscilloscope bandwidth?
 
Quote:

Originally posted by Craig405

I have heard that giving the scope noisy supply rails will limit the available bandwidth of a scope such as this (500MHz, 2Gs/s).
Id like to obtain close to the original spec from this puppy if possible.
Before going too far down the line of repair id like to know if this is the cause and what the most appropriate solution is?

Hmmm... The ADC's in the scope should have little to do with the microcontroller that runs the user interface, etc. So noise in the converters, and re-booting are clearly two different issues (as you may realize). It's true high performance ADCs need a good power supply. But the power conditioning is usually close to the ADCs themselves not on a different board or module. So, with some luck, the quality of the actual power supply is less of an issue.

Given your scope has had a hard life from your other thread, I'd suspect some hardware damage (sadly) from the previous power supply problems. Or perhaps an entirely different problem. And when you switch time ranges some sort of invalid (i.e. unforeseen by the firmware developers) set of conditions occurs that causes the firmware to crash. Most instruments have a watchdog timer that will reboot them if the firmware crashes.

It's possible the actual noise from the relay coil is causing the problem, but that doesn't seem likely. I would think Lecroy would have properly suppressed the relay inductance, etc. It's much more likely something serious is wrong somewhere.

Does it pass its self test(s) and calibration? Sadly, scopes these days are so highly integrated with lots of custom FPGAs, ASICs, and other programmed devices, in BGA (ball grid array) packages, component level repair of much of the scope is not feasible by anyone other than the factory and they might not even attempt it given the high cost to remove and replace even a single BGA. So repair is generally limited to replacing, typically very expensive, entire circuit boards. But you probably already know that :bawling:

Craig405 5th August 2009 08:36 PM

Hi RocketScientist, Thanks for the response. Do you think its likely to be a damaged piece of hardware then? Would it be worth me giving it a run from some bench supplies in the future to be sure?

Bummer if so, it works perfectly in every other aspect.

RocketScientist 5th August 2009 08:41 PM

Quote:

Originally posted by Craig405
Hi RocketScientist, Thanks for the response. Do you think its likely to be a damaged piece of hardware then? Would it be worth me giving it a run from some bench supplies in the future to be sure?

I guess that depends on how easy that is to do? It couldn't hurt. Does it fail its self test and self calibration? Have you tried contacting Lecroy? Sometimes you can get a sympathetic person who might be willing to offer advice. The good news is you can get some almost decent digital scopes these days for vastly less than your LeCroy cost. In fact, several are probably cheaper than having your LeCroy repaired.

Roar Malmin 5th August 2009 09:24 PM

If everything seems futile, and the scope will not work, I can sell you a perfectly ok LeCroy 9350L for 500 + p&p to UK.

Regards
Roar

Craig405 5th August 2009 10:19 PM

Hi, I forgot to answer you on those points, it doesnt seem to do a self test / self calibration (that it will display for me anyway). I think these are supposed to be run from a floppy drive but i dont have the diagnostic software.

I might ask lecroy for it as its surely free.

I was playing around with the scope tonight and discovered a few strange behaviours:

1) Reducing the number of points stored from waveform acquisition enables me to move the time base down to 10ns (previously couldnt get past 50uS) or so before i get a crash.

2) Using the scope to measure its own rails I get nasty noisy periods of 1-2V pk-pk every so often from the 15V SMPS's im using to power this thing. The last noisy period i observed ended with a scope crash, can i draw a conclusion about supplies here or does the fact I am using the scope to measure its own supply make the measurement pointless lol.

In fact casting my mind back to when i had a lab to test this thing in a year ago it ran perfectly well from some expensive 50A bench supplies (four of the things!) without crashing on me once and I was fiddling with the timebase at 2nS too. Ps this was after I had finished exploding it :)

Roar, thanks again for the kind offer if I had any money I might just have been interested in that (im an unemployed student haha!).

RocketScientist 5th August 2009 10:56 PM

If you have access to lab supplies, it's worth a try. Using the scope to measure its own supplies might be suspect--it's hard to say.

Have you tried simply adding more capacitance to the supply rails as close to the scope's boards as possible? If there's ample capacitance, and you're certain the scope isn't triggering the current limiting or otherwise causing your supply(ies) to drop out of regulation, I would still suspect some other problem--especially if it's old enough to have a floppy drive ;)

I would guess the scope has a switching regulator, or a few of them, to power the high speed digital circuitry which typically needs from 1.2 - 3.3 volts. The electrolytic capacitors in those regulators have a hard life and tend to get weaker with age. It's a common problem on old PC motherboards for the switching regulator that powers the CPU, etc. to start to get marginal causing instability as the caps age (or even sometimes outright fail).

It's possible replacing the regulator caps might solve your problem. Especially if you're feeding the regulators with less than clean power, it be exposing any bad or weak filter caps.
I'm sure LeCroy has a service manual for your scope, if you don't have it already, but most are just for board level repair. They usually don't give you schematics or details of the individual boards. Still, you should be able to identify the caps as they'll be near some power MOSFETs (probably surface mounted to the PCB) and one or more inductors.

The regulators are usually located as close as possible to the most power hungry chips they're feeding (i.e. the ones that run hottest) So they can be anywhere in the scope and there are likely more than one. Look at a PC motherboard around the CPU socket for an idea of what the circuits typically look like if you're not familiar with them.

Replacing those caps, assuming you can find and get to them, should be a relatively cheap repair. But it may or may not fix your problem.

Craig405 6th August 2009 10:04 AM

I dont have access to lab supplies yet, i have to wait a few months to use a university lab. Would be so convenient right now though!

Its a 1996 scope i Believe, its old yes but its pretty fast and most of the bells and whistles of a modern high priced scope, and doesnt run XP like the newer Lecroys do ;)

I may have trouble getting to the mainboard in this thing its all very tightly shielded and id have to do some serious disassembley.

I think ill try as you say and put some localised capacitors as close as possible.
how about 50uH inductor -> 1500uF // 10uF tantalum // 100nF film on each 5V rail, should that do a pretty good job?

For the analogue supplies 2200uF // 10uF tantalum // 100nF film.
Ideally if i can solve this problem by cleaning the supplies up then id use a linear supply on the 15V's and try to modify some good regulators like the jung super.


All times are GMT. The time now is 10:11 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2