Did you try running signal path compensation.
It usually takes care of signal offsets.
Go to Utility-Cal, read the screen and run SPC. If everything is fine,
you'll have no DC offset any more.
--Artem
Juerg wrote:
OK, here's an update on my progress:
I cleaned the acquisition board using all methods mentioned in the
thread (toothbrush with simple green and isopropyl alcohol followed by
a dishwasher cycle). I reassembled the scope and - oh wonder - the
sawtooth/triangle is gone but all the channels still have different
offsets (even when the input is grounded). Could that still be an after
effect of leaky caps?
I'll start looking at the bias voltage generation/convertion to figure
where the error gets introduced.
NOW I'm having fun :-)
...juerg
mike wrote:
Juerg wrote:
That's what I figured from searching the web...
Anyways, I took a closer look today and found the D/A converter
(AD667)
for the bias voltages and a bunch (5) of analog 8:1 multiplexers
(don't
remember the part number)... And yes, as Mike suspected, the
frequency
of the sawtooth is the same as the refresh rate of the D/A. The
sawtooth is also visible on the output of a couple of the
multiplexers
as well as on the input and output of a couple of the OpAmps
(TL074C).
But I haven't found any holding caps yet, that also show the
sawtooth
across them.
I'll keep hunting around. Maybe I just replace all the electrolytic
caps in that area...
Thanks for your help, guys
....juerg
Look for quad op-amps. TL074 in the 540.
They're configured as simple voltage followers.
The holding caps are typically 0.1uf SMT on the input.
Sometimes there's a series resistor.
The caps probably aren't bad, unless they got cracked.
There's a lot of stuff on the backside of the board.
Make sure to check for VCC on the op-amps.
Replacing the electrolytics is good, but won't fix it.
If they leaked, the electrolyte is already on the board.
Do the math. It takes very little board leakage to turn your
reference voltages into triangles.
You have to CLEAN the board. First one I fixed, symptoms got better
every time I cleaned it. About the fourth time, it started working.
Clean the whole board, both sides. I found electrolyte droplets far
from any caps. If you try to spot-clean you'll just move the gunk
around.
Use something like Simple Green and a stiff toothbrush. Clean the
CRAP
out of it. You have to get the invisible coating out from under the
chips. I used an Xacto knife to scrape gunk from between IC pins.
Sprayed cleaner HARD in attempt to force it under chips.
Cleaned it again with 99% Isopropyl Alcohol.
Dried with a hair dryer then let it sit overnight on the heater vent.
There's a reason for the two cleanings. Some stuff dissolves in the
cleaner that won't in alcohol and vice versa.
Took FOUR cycles of this before I got it all working. At the start,
I
was very timid about scrubbing the board. By the end, my scrubbing
was
brutal.
Tape up any switches on the board and try to keep the cleaners out of
'em. Alcohol probably won't hurt the switches, but all the gunk
dissolved in it might.
Are we having fun yet?
mike
OH, about replacing caps...I have never seen your scope, this relates
to
the TDS540. YMMV. If one is leaking, they probably all are.
Change them all. There's over a hundred on a 540. Don't forget the
front panel board. You don't want that gunk getting into the
encoders.
If they're the round silver caps like mine, the connections will be
corroded. Go over them one connection at a time to burn off the gunk
and get down to clean solder. Then use TWO soldering irons. Apply
the
heat and twist slightly with the irons to rotate the caps. When it
gets
hot enough, they spin right off. This technique applies shear force
to
the pad and is much less likely to lift the pad.
Clean it while the caps are off.
Mine had two values of caps. I did some voltage checks and decided
that
47uF 25V caps would work everywhere. Again, this was a 540 YMMV.
Good news was that I had a bunch of 'em. First board, I put a .1uF
1206
smt then put the 47uF leaded cap on top of that. By the fourth
board,
I left off the .1s. Didn't seem to matter.
If you lift a pad, the .1 gives you something stable to solder to.
Start on caps where you can get at the trace on both ends. If you
lift
a pad, you'll be able to fix it. Save the blind ones till you've
done a
few dozen.
Now, we're having fun...
--
Return address is VALID but some sites block emails
with links. Delete this sig when replying.
.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
MAKE THE OBVIOUS CHANGES TO THE LINK
htremovethistp://www.geocities.com/SiliconValley/Monitor/4710/
|