Tek TDS 360 Digitizer Problem

[cheapskate]

Hello everyone,
I recently purchased a broken Tektronix TDS 360 that has issues digitizing its input. At low sampling rates (<50ksps), the output is mostly correct except for in the middle and end of the sampling interval. At the very end, the signal appears very distorted or doesn't look like the input signal at all. In the middle, there seems to be some kind of phase shift error which results in the wrong part of the signal being digitized.


The phase error is consistent in its location though doesn't always occur.


At higher sampling rates (>100ksps), the signal becomes increasingly corrupted until it disappears completely at 1gsps.
Interestingly, the degree of corruption depends on the frequency of the input signal. A low frequency signal paired with
a low sampling rate and a higher frequency signal and fast sampling rate will result in similar levels of corruption, but a low
frequency signal with a fast sampling rate will be very corrupted.


The oscilloscope passes the power on self test but running the diagnostics utility results in the following errors.

diagAcq_fisoCell
diagAcq_acqMemAddr

I have been looking through the maintenance manual but it seems like repair = replace. I have checked the power supply
and all of the voltages seem correct. The main CPU/Acquisition board only has 3 electrolytics on it and seems to be in
excellent shape. I have been tracing out the signal path using the service manual and have concluded that the issues begins
somewhere between the analog front end and the A/D converter. I have been able to monitor the waveform at the
output of the sample driver and it seems to be correct. The waveform at the input to the A/D is this strange digitally modulated
version of the input. While I can't see the corruption on my analog scope, because of the general blurryness and instability
of this signal, I suspect that the errors are introduced before the A/D. I have also tried disconnecting the A/D from the signal
path, which causes the corruption to disappear. At this point I believe that the errors originate in the FISO sample memory,
which is this strange CCD analog memory device that is made of unobtanium.

Image I stole from this thread: https://www.eevblog.com/forum/repair/pls-help-me-repair-tektronix-tds-380/

FISO driver output bottom, FISO output top. FISO signal appears to be a series of discrete analog voltage levels.

At high frequencies the FISO output doesn't make much sense.

Do you think there is any hope for this scope? Could it be something other than the FISO memory or something that is negatively affecting its operation? I don't want to replace the main board because the scope really isn't worth it, but I do need a digitizing scope for some of my upcoming projects. Do any of you have a lead on this unobtanium part? 

There is one other issue that I haven't gotten to diagnosing yet because it seems to be separate from the sampling errors.
Sometimes, signals exhibit a stair step effect that causes them to shift up and down periodically. This happens even when
the inputs are coupled to ground and the steps appear to be consistent in height. The period is very slow, approximately 1
second. The stepping can't be detected at the output of the FISO driver or the FISO itself, so I'm not sure where it is coming
from.

[james_s]

The fault and the error message both suggest a problem with the acquisition memory. What you have circled as "memory" looks like the ROM that holds the firmware to me. Is that a SRAM chip there in the acquisition section? You might try replacing that. It has been my experience that SRAM is one of the more fragile ICs out there, I've replaced quite a few RAM chips fixing old arcade game boards.

Unfortunately there is very little information out there on fixing the TDS300 series, I fixed the power supply in one but I never have been able to successfully calibrate one and have not found much other info.

[shakalnokturn]

Schematics here:
http://w140.com/tektronix-tds380-tech-ref-schematics.pdf

It would be worth checking local power and bias voltages to the acquisition section around FISO (U590), it could avoid running in circles looking for a problem that was only supply related.

Then considering there seem to be problems related to both analog (FISO) and digital (ADC / RAM) maybe reflow U401 the ASIC that's common to both.

May be worth checking the 60.6MHz acquisition clock for stability and amplitude too.

[james_s]

Wow, I didn't realize there was a manual that had schematics, I looked all over for one a while back.

Now if only I could find a schematic for the TDS400 series.

[shakalnokturn]

Now if only I could find a schematic for the TDS400 series.

Same here...
I'm sure if all the people who've drawn fragments of schematics of the TDS400 during repair attempts put them together we'd have the full picture by now.

[Labrat101]

Hi ,
I have the THD340 which has the same layout . the only differences are there is the BW caps.
I got my one about a year ago and it also had a lot of errors similar . Not as bad as yours .
I would first just run your finger over the top of the Dallas real time chip and see if it has a slight bulge on the top.
If the battery volts drops below 1v this will cause memory errors .and the errors are not always true to the faults.
 Looking at your photos I think you may have 2 problems 1 the Battery in the real time . also I found that some of
the some SMD caps were bad 0.1uf ceramics which on my one I replaced about 7  & the battery which I dremeled out.
and fitted a external battery in a holder 3v lithium.  And then had the fun of recalibration which takes about an hour.

May I ask what was your test frequency? 
What happens if you use the AutoSet ?
I guess you did . try setting it to default settings. shrink to fit screen is OFF .
Also try turning the Auto trigger OFF  (in readout options Display @ trigger point )
Also Make sure that in the horizontal menu  that the Trigger positions are correctly showing . Norm 25% this is adjustable .

It was just a passing thought That some of the setting don't always correct when set to default if the battery is low these don't  update to memory.
 If you have done these . ignore .
All the calibration is held in the Dallas chip . so if this does go bad some of the error become false. 


@ james_s  If the PDF is good for you enjoy . Santa delivered . Just say Thanks 
     

   Now if only I could find a schematic for the TDS400 series. 

[cheapskate]

Well I have some encouraging news. After messing with the FISO some more, I desoldered the op amp in between it and the ADC and the error did not go away. I guess I falsely ruled out the A/D and ACQ RAM after all. I then connected the input of the ADC to its zero point which resulted in error similar to but not identical to those I was experiencing earlier. It seems most probable that the ACQ RAM is bad, though I can't rule out the ADC either. I don't have a logic analyzer to peek at the ADC data, I guess I will just try to see if there is any activity since the input is all zeros. Now I just need to source the AS7C164–12JCTR SRAM from somewhere and try to replace it.

Still no progress on the stair stepping problem. It has disappeared once I disconnected the ADC from the FISO, maybe related to the gain control DAC?

[Labrat101]

Well I have some encouraging news. After messing with the FISO some more, I desoldered the op amp in between it and the ADC and the error did not go away. I guess I falsely ruled out the A/D and ACQ RAM after all. I then connected the input of the ADC to its zero point which resulted in error similar to but not identical to those I was experiencing earlier. It seems most probable that the ACQ RAM is bad, though I can't rule out the ADC either. I don't have a logic analyzer to peek at the ADC data, I guess I will just try to see if there is any activity since the input is all zeros. Now I just need to source the AS7C164–12JCTR
 SRAM from somewhere and try to replace it.

Still no progress on the stair stepping problem. It has disappeared once I disconnected the ADC from the FISO, maybe related to the gain control DAC?

Did you not read what I suggested . I think your looking at the wrong area .

[cheapskate]

Hello Labrat101,
I know the Dallas RAM is a time bomb, but it seems to be working for now and the RTC is updating as expected. I can't detect any bulging. I hope to address this issue in the near future but I don't know much about repairing or replacing them. How did you figure out that the SMD capacitors were bad? How did you check for it? I can only see brown ceramic capacitors which I though didn't have degradation issues.

I don't know much about using the unit, so I haven't really played around with the settings. AutoSet works more or less as expected, it centers the screen and adjust the time base. Triggering works, I think it is unaffected by the errors because it is done directly from the analog signal.

[Labrat101]

Hello Labrat101,
I know the Dallas RAM is a time bomb, but it seems to be working for now and the RTC is updating as expected. I can't detect any bulging. I hope to address this issue in the near future but I don't know much about repairing or replacing them. How did you figure out that the SMD capacitors were bad? How did you check for it? I can only see brown ceramic capacitors which I though didn't have degradation issues.

I don't know much about using the unit, so I haven't really played around with the settings. AutoSet works more or less as expected, it centers the screen and adjust the time base. Triggering works, I think it is unaffected by the errors because it is done directly from the analog signal.

I used an ESR meter and I also have on board cap tester . which save time when checking SMD caps .
 This scope is 20 years old +_ a few yrs. So the Dallas Battery will be for sure low. There is a 1k ram in that chip
when the battery gets low the RAM becomes corrupted and throws out the wrong errors . Which becomes
confusing . Look up the pin out for the dallas there is a place to check it . I don't remember of hand it the ground pin and
one of the pins hard up against the white plug need a needle probe to get to it .
But I suggest stop removing chips as you will kill it for sure .

I also suggest you learn were all the menus are as there are some of the time base setting don't change
when setting default . In save/recall there are factory setting see if one of those makes any difference .
Also if it lets you . go into Utility . on the bottom left button near the power on.
press several times to scroll to Cal . It will run the self cal make sure there is nothing connected to the inputs .
let it run go and have a coffee it takes about 8 minutes to complete . so don't touch any controls while its working.
when it finishes see if all the icons to the right show passed .

[shakalnokturn]

tektronix_tds410-tds420-tds460-oscilloscope-service_manual.pdf

Not bad, the best around, still lacks a component level schematic which is what we all need.

[Labrat101]

tektronix_tds410-tds420-tds460-oscilloscope-service_manual.pdf

Not bad, the best around, still lacks a component level schematic which is what we all need.

Which scope the exact number 410 ,420 , or 460 .  miracles take a little longer 

[shakalnokturn]

The scope I'm currently working on is a TDS 420, however if you are going to do a miracle I expect the TDS 460 schematic would be the best compromise for all those yearning.
Ho, ho, ho !? 

[cheapskate]

Does the TDS 400 series really have sampling rates lower than its analog bandwidth? What's the point of that? Doesn't that result in aliasing errors?

Well I tried running the calibration and it only made things worse. It completely messed up the DC offset and spit out a bunch more errors. I'm not sure if those are related to something I screwed up or if they were triggered by failing calibration. 



Cal error, sampler gainrange
Cal error, slow sampler mode
Cal error, sampler gain/offset

diagAcq_acqMemData
diagAcq_acqMemPat
diagAcq_acqMemAddr
I believe these are related to the acquisition SRAM

diagAcq_fastFisoAcq
diagAcq_fisoShortPipe
diagAcq_fisoCell
diagAcq_peakDetectAcq
These are probably related to the analog front end

I think I am going to try replacing the acquisition SRAM, I can't get the original AS7C164–12JCTR 12ns part from a trusted distributor, but I can get the similar IS61C64AL-10JLI-TR which is rated for 10ns. Alternatively I could get a 256k version which has some extra address pins which aren't connected in the original design.



The stairstepping issue is till there, its even present when AC and ground coupling are used which means it is introduced somewhere after the input.

[cheapskate]

I have also probed the power supply for the FISO sampler chip and +5v seems fine but there are a few pins that I don't understand. VDDL (pin 8 ) measures 3.70v and VSUB (pins 7, 22, 39) measure -1.36v does this make sense? VSUB is supposed to be related to -5v but I'm not sure what that circuit is supposed to accomplish.

[james_s]

tektronix_tds410-tds420-tds460-oscilloscope-service_manual.pdf

Not bad, the best around, still lacks a component level schematic which is what we all need.

Which scope the exact number 410 ,420 , or 460 .  miracles take a little longer 

The one I have with issues is a TDS460. I think the other one is a TDS410A.

[Labrat101]

Can u do a photo of the version Firmware number screen. .  In utility menu .
.
Also it might  be worth checking the undersides of the panel.  Front section.  For tin whiskers.
This is a very over looked fault . Some of the whisker maybe only a few microns thick but can grow to quite along length..
Clean the entire bottom of the board with a soft toothbrush..
As you don't know were this scope has sat or what environmental conditions its been in over the last 2 decades. .
I don't think you have a chip problem. 
 If you change that chip for non original you won't be able to re cal. .

Update . Also I would recommend checking the front control panel . Ie clean all the rotatory & switches with a switch cleaner as this is where Human wear and tear appears . also check PCB for hair line cracks on the most used controls.
These TDS 360 scopes are built like tanks . The errors you have are more likely to be something simple .
 I have seen weird stuff over the years . Even Spider poop across a track will give the weirdest errors .
  worst case its a failed SMD cap with a hairline crack this is not uncommon .
As the cal made things worse there is a bad connection some were . dry joint .
 
 
.

[madao]

TDS300 and TDS400 share only  similar case.

Eletronic is extremely difference.   TDS300 Serie use only  one PCB with ADC, CPU, Firmware, etc. And their case is really empty.

TDS400 is just a  TDS520/540 in other format and slower DAC and less memory. -> Full package of eletronics in case.


Issuses with  waveform look me after bad memory /clock

[Labrat101]

TDS300 and TDS400 share only  similar case.

Eletronic is extremely difference.   TDS300 Serie use only  one PCB with ADC, CPU, Firmware, etc. And their case is really empty.

TDS400 is just a  TDS520/540 in other format and slower DAC and less memory. -> Full package of eletronics in case.


Issuses with  waveform look me after bad memory /clock

He has the TDS 360   . 
The other ref to the 400 were asking for the service manual . we know that the 300 & the 400 series are different.

[cheapskate]

The TDS 360 is disappointingly empty inside, I didn't expect it to be so light either.

The first thing I did was disassemble the front panel control panel as it was filthy. Excellent design, very serviceable and robust. I can't see how this can be a front panel issue since the controls work when used and user interface itself isn't jittery, only the data shown is corrupted. I haven't cleaned the main board yet and have mainly ignored the underside. Thank you for the reminder, I will do that soon.

The firmware version is 1.02 from Nov 6, 1995. The technical reference and programmer's manual I have are both for version 1.05 so a newer firmware does exist.

Why do you say that I won't be able to recalibrate after replacing the SRAM? The original has 1 extra CS line over my proposed replacement (IS61C64AL has an NC where AS7C164 has CS2), but seeing as the diagram shows them tied to GND and +5v I don't think that would be an issue. The replacement also has slightly better performance than the original so it shouldn't be too slow either. Do you think that faster timing would cause issues?

[Labrat101]

I have never used a substitute replacement . But I have a horrible feeling it will mess up the calibration and
you may have to compensate on some of the smd components .
On the other hand its your Baby . try it it you may be lucky .
But the real thing I noticed on your error screen are the 3 CPU error marked at the top . they are displayed on
another page . You will have to remove the  cal jumper to see all the errors.
Also the error list has  number errors .
That why I suggested you check for other faults the number codes on the list are default
The CPU error has to be solved first  . This could be due to several reasons . The timing is getting distorted
on one of the data inputs . Hardware . could be a dozen reasons that's why I suggested looking for bad solder connections .
 I found sometimes helps use a daylight led light on a zoomable camera .
Hairline cracks show up . I use a 21mp HDMI with a long focal lens 250x on a swinging arm that I use to scan over large boards bit by bit .
This scope may have had a hard life and may have been used by heavy handed people .
So I would start by cleaning the bottom of the board . and go over it section by section . and resolder
any thing that does not look good . I have done this many times and 75% of the time fixes most of the
problems . One dry joint or a dead bug . can cause more problems . SMD caps can also fail .
Look for discoloured resistors as well . And not least to check are Plugs and sockets on the ribbon cables
the pins go black . as they are silvered . 
This is a nice holiday project . with many hours of careful work though .
update there are 12 cal errors 3 cpu errors .

[james_s]

TDS300 and TDS400 share only  similar case.

Eletronic is extremely difference.   TDS300 Serie use only  one PCB with ADC, CPU, Firmware, etc. And their case is really empty.

TDS400 is just a  TDS520/540 in other format and slower DAC and less memory. -> Full package of eletronics in case.


Issuses with  waveform look me after bad memory /clock

The chassis is identical, the TDS300 chassis has all the holes for where the TDS400 guts would mount but yes other than sharing the same case and video display they are completely different designs. The TDS300 is cost reduced by integrating the entire scope onto one small PCB, it's much like a Mac Classic, it's mostly empty space. It would have been nice if they included a storage compartment in the thing for probes and accessories, there's enough room to store an entire modern compact DSO inside it.

[Labrat101]

I agree the empty space is a bit of a waste . Made side panel opening . Coffee flask just fits in a treat.
( I'm going to change the resistor jumpers on my TDS340 and reset it to a 360 . Just for Fun )

[shakalnokturn]

I have also probed the power supply for the FISO sampler chip and +5v seems fine but there are a few pins that I don't understand. VDDL (pin 8 ) measures 3.70v and VSUB (pins 7, 22, 39) measure -1.36v does this make sense? VSUB is supposed to be related to -5v but I'm not sure what that circuit is supposed to accomplish.

I imagine VSUB is a bias voltage for the CCD, "-5VA"+2Vf(CR543) so measured voltage seems to make sense.
No clue about use of VDDL but same idea measured voltage makes sense "+5VD"-2Vf(CR542).

I really doubt substituting a 12ns acquisition RAM for a 10ns one will cause any problems.

[Labrat101]

This a TDS  that I repaired recently . Had about 3 pages of errors .
Your all trying to make a dogs dinner out of a very simple scope repair . No Rocket science required.
Firstly the main problem is that the Real time block has 256k of RAM . when the battery becomes old the RAM suffers
from a stack overflow that is then loaded into the main Memory with the stack errors . on each reboot the error list grows .
This 256K of RAM holds all the Calibration setting and the last front panel setting . .
The software v1.02 does not check for this stack over flow as its a BUG with the Dallas 
So Changing the Memory chip will clear some of the error . So will a hard reset done vie the 2 TP pins .
The voltages you measured on the VDDL are correct  .

EXAMPLE::  all you are doing . is Measuring the Back Tyre  pressure when the Nail is in the Front Tyre .

The Dallas Chip has to be reset so does the main memory . But only after you have fixed the issue .
The Scope will then Need (require a full recalibration)

(Image Blue Caps are the new replacements 0.1uf & 0.22uf)
Dallas R/time with external 3v cell

Try removing C230 and C231  these two caps if faulty or damaged
 will cause these effects . You can run the scope while they are removed.
 if the scope works correctly without them they can be left out.
 if there is only miner improvement or it makes it worse replace both with NEW.