Effect of Supply noise on oscilloscope bandwidth? - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 6th August 2009, 11:30 AM   #11
jaycee is offline jaycee  United Kingdom
diyAudio Member
 
Join Date: Apr 2004
Location: Norwich, UK
On the main PCB's i wouldn't try to add parts - simply renew the electrolytic capacitors already there.

what you may want to do is put some capacitance, say 1000uF in the scope where the power supply pcb wouldve been, to minimise the effects of the long cables from the PC PSU's.

Also, how are you getting +15/-15V from PC PSU's ? they typically only supply 12v.
  Reply With Quote
Old 6th August 2009, 04:18 PM   #12
diyAudio Member
 
RocketScientist's Avatar
 
Join Date: Mar 2008
Location: Northwest
Quote:
Originally posted by Craig405
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.
One other thought is grounding. If you've somehow altered the grounding for the power rails that could be creating ground loops or have ground currents flowing where they're not welcome. Try to re-create the original grounding as closely as possible.

Also, always think "loop area" when handling any wiring that carries any sort of significant current around sensitive circuits (including digital circuits). In this case that means keeping the power and ground wiring twisted or bundled together and as short as possible. If the power and ground wiring, or even the two rails, are split apart, and forms any sort of "loop" (open space between the conductors) you just created an antenna that will broadcast all the current spikes on the power supply throughout the scope's circuitry. Those spikes could well glitch sensitive logic causing the re-boot you're seeing.

Given the age of your scope it probably has some really power hungry circuitry in it. It was harder to do 2 Ghz sampling in the mid 90's. So, somewhat revising my earlier opinion, I'd say you may well have a power problem. It could be your hacked supply isn't up to the job, it could be a ground or loop area issue, it could be you need more capacitance close to where the power is needed, or it could be the switching regulators are getting marginal.

If you can clean up the power without taking the scope apart, and that solves the problem, great. Otherwise, you have nothing to lose by taking it apart further and replacing the switching regulator caps (unless it's partly usable as is and you're worried about breaking something).

And, AFAIK, the more popular and lower-end new LeCroys don't run any flavor of Windows. I know Agilent and Tek both largely gave up on that approach for a whole bunch of reasons. I think LeCroy even finally has a sub $1000 scope now to compete with similar models from Agilent and Tek.
  Reply With Quote
Old 6th August 2009, 08:51 PM   #13
diyAudio Member
 
Craig405's Avatar
 
Join Date: Jul 2005
Location: Sussex
Jaycee, I meant to put these parts in the leads to the scope, ill attach some photos to be clearer . I would break out the molex in the first pic and insert a filtering board, or even shorten the wires from the scope and fix it inside in the old PSU bay.
My +/-15V is generated from two laptop power supplies (toshiba) but they are knock offs from Ebay and im suspicious of their power quality.

Grounding could be a an issue, I tried to give the power supplies and scope a good start with a star ground via a large copper bus bar (see next photo) whilst im sure this is much better than splicing wires all over the place there may still be issues with long wires to the scope and from the PC supplies themselves.

I also used every ground wire available from the PSU's 24 way molex to minimise resitances from bus bar back to the SMPS.

I hope there is minimal loop opportunity as I have tightly bundled wires to the scope but i didnt twist
The coloured cable going to the scope has:
10 molex connectors occupied with ground wires
10 wires for +/- 5V
4 wires for the +/-15V

Yep this thing is thirsty! LeCroy told me to provide:
-5.2V @13A
+5.2V @14A
-15.1V @ 4.5A
+15.1V @ 4A

MY two ATX power supplies are brand new antec units that can do 20A on the 5V rail each and the laptop PSU's can do 5A at 15V.

I think ill go ahead and start shortening wires as much as possible and if you think my above suggestion of the LC filtering for the 5V rail is suitable ill implement this as close to the scopes internal power connector as possible (but still inline with the wires)

phewf that was a biggun!

P.s. i have acquired a load of ferrite clip on beads, would placing these on the ends of the wires also be beneficial?
Attached Images
File Type: jpg dsc01090.jpg (94.8 KB, 124 views)
  Reply With Quote
Old 6th August 2009, 08:53 PM   #14
diyAudio Member
 
Craig405's Avatar
 
Join Date: Jul 2005
Location: Sussex
next pic of bus bar.
Attached Images
File Type: jpg dsc01091.jpg (95.4 KB, 120 views)
  Reply With Quote
Old 6th August 2009, 08:55 PM   #15
diyAudio Member
 
Craig405's Avatar
 
Join Date: Jul 2005
Location: Sussex
And one more of the setup, scope front in foreground and rear of my infernal power supplying contraption in the background!

I apologise for these poor photos, indoor lighting bulbs and a phone camera make for shoddy pics.
Attached Images
File Type: jpg dsc01092.jpg (96.0 KB, 118 views)
  Reply With Quote
Old 6th August 2009, 09:09 PM   #16
diyAudio Member
 
RocketScientist's Avatar
 
Join Date: Mar 2008
Location: Northwest
Those are some long wires. You def. should have some serious capacitors as close to the scope as possible (see my earlier comments). And, ideally, the caps should ground to the respective grounds for each rail--again, as close as possible to where the power feeds the scope's boards.

Is the ground bar your addition? A star ground might not be the best approach if that's not what the old supply used. LeCroy may have intentionally isolated some of the ground paths for different supply rails. And if you tied them together, even on a bus bar, you might be creating problems. So I'd trace out how the old grounds terminated and try to duplicate that as closely as possible.

Generally, at some point, I would expect the scope to have completely isolated ground returns for the digital and analog grounds. Where that happens in the scope, I don't know. But if you tie those points together in some arbitrary place you're may degrade the analog performance and possibly create other problems.

Also, some PC power supplies dislike having most of their load on a single voltage. They're not designed for that and they may not function correctly. The feedback loops in the power supply have to balance the needs of the various output voltages against the PWM duty cycle. And some have minimum load requirements on the outputs or they're prone to tripping their overvoltage, falling out of regulation, or other problems.

In modern PC's it's often the 12 volt supply that's used (via the 4 pin ATX connector) to feed the hungry CPU switching regulator. In older PC's it's the 3.3 volt output. So if your PC supplies have no 12 volt or 3.3 volt load, they might fall out of regulation trying to provide 15+ amps at 5 volts because the lack of load on the other outputs is causing the PWM duty cycle to be way too low.

Also, check out http://circuitspecialists.com/level.itml/icOid/8518 for cheap single voltage 5 and 15 volt supplies that have an adjustment trim to give you 5.2 volts. They may also have a dual or triple that would work. You can get full data at Power One.
  Reply With Quote
Old 7th August 2009, 10:17 AM   #17
diyAudio Member
 
Craig405's Avatar
 
Join Date: Jul 2005
Location: Sussex
Ok heres what ill do:

1) shorten all wires as much as possible.
2) find the analogue ground wire(s) and keep them seperate of the 5V GND. I checked the LeCroy PSU and it does indeed appear to have seperate grounds, cheers I never thought to check this
3) build a board inside the scope in the old PSU bay with a load of inductors and caps, anything else worth adding to this board before i build it?
4) ill maintain the star bus for the 5V ground.
5) shall i connect the isolated 15V analogue ground to chasis earth via 10 ohms or so?
5V ground is connected to earth by default due to the internals of the SMPS.

With respect to the loading of rails on the PC PSU's, I actually did this with appropriate resistors and a heatsink but removed it as it worked fine without. Putting this back is no problem.

That link you gave is excellent! I may pick up

http://circuitspecialists.com/prod.itml/icOid/9105

to replace the PC supplies, the specs are better and theyre really cheap, just need to find a UK supplier.

By the way thanks for your continuing help, im learning quite a lot here!
  Reply With Quote
Old 7th August 2009, 04:23 PM   #18
diyAudio Member
 
RocketScientist's Avatar
 
Join Date: Mar 2008
Location: Northwest
Try to use a DMM or visual inspection to see what LeCroy did with the analog ground. Usually it gets tied somehow to either the main star ground, earth, etc. It might be directly (at the right point), through a resistor or resistor/cap combination.

I also don't think you need inductors. At the currents you're talking about they will be somewhat expensive and could even cause problems if they're not sized correctly. I'd use no more than 4700uf for the 5v supplies and 1000uf for the 15 volt supplies with a 0.1uf stacked ceramic or film cap across them. Anything more is probably overkill, and your supplies may not appreciate the start-up surge current of larger caps (it might trip the over current limit).

And before you make a fancy board, you can probably just temporarily solder things up hanging in space to see if it's going to work. You also probably want to experiement with grounding before making the board to find out what works, or doesn't work.

As for loading your PC supplies that's hard to know if you don't have a scope to monitor the rails. I guess I might temporily add the loads back (if you still have them) and if you get the scope working correctly, you could then try removing them and see if it makes any difference then. But for now, the more problem areas you can try to rule out, the better.

And I forgot you're in the UK. Those power supplies really are a bargain. And I've had good luck using them a few of them for various applications.
  Reply With Quote
Old 7th August 2009, 05:52 PM   #19
jaycee is offline jaycee  United Kingdom
diyAudio Member
 
Join Date: Apr 2004
Location: Norwich, UK
If this was me, I would probably look at getting some 18-24V supply that can supply some hefty power, and then build a DC-DC board to go where the old PSU went. Linear or Fairchild will probably have some suitable regulators to turn 18-24V into a symmetric 5V at hefty amperage (with external switching mosfets).

For the 15V supplies i'd use linear regulation with a gyrator to clean up the SMPS crap.
  Reply With Quote
Old 9th August 2009, 08:50 PM   #20
diyAudio Member
 
Craig405's Avatar
 
Join Date: Jul 2005
Location: Sussex
Just got back from a weekend away, got more time now to sort this power supply out now.

Ill imitate as far as i can tell what LeCroy have done for the grounds.

Im also a bit wary of using inductors, ill attach an image of the beasts i had in mind, an uneducated guess would say they are in the order of uH's maybe 50-100uH and i know for sure they are rated at around 30A .
So this eliminates cost and power issues but if they arent appropriate in an LC configuration suppling logic then ill leave them out, I assume there are conditions where they might cause oscillation with the C's?

Jaycee, I did have a look at some DC-DC converters from Traco actually but they were very expensive, however I didnt look at Linear or Fairchild. Also money being a factor and desk space ( i can stack my current solution under the scope neatly) the beasty Krell look-a-like bench supplies are off limits too

But Im all in for Linear 15V supplies and have been looking at Jung regulators with LM1084's (5A LDO variable) in the front end and seeing if the circuit can be modified with good performance to give the currents i need.
Ive read a little about 'Gyrators' as Wavebourne has used them recently.
Attached Images
File Type: jpg dsc01094.jpg (74.8 KB, 76 views)
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
LeCroy 9354AM oscilloscope power supply Craig405 Parts 3 19th December 2011 07:34 AM
Effect of supply bypass positioning. Arx Chip Amps 10 10th June 2007 08:13 PM
Effect of Power Supply on Soundstage Width foxyb Solid State 62 13th December 2005 07:00 PM
X Aleph Bandwidth & Power Supply macka2 Pass Labs 21 30th August 2004 04:08 AM
power supply noise flux01 Pass Labs 12 9th February 2004 12:42 PM


New To Site? Need Help?

All times are GMT. The time now is 07:48 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