TEK TPS2000 series scope fault issue

[trobbins]

I have a TEK TPS2012 (2 channel, 100MHz digital scope). 

All functions are normal and the displayed measurement is ok down to 250us/div horizontal scale, but at lower horizontal settings (100us/div and lower) it displays a noisy signal of somewhat lower amplitude (although the amplitude varies with vertical scale, and the timebase varies with horizontal scale). Effectively the same situation with both channels. Using the Probe Comp signal source, there is a slight DC offset with Ch2, but otherwise the same as Ch1.  See attached plots below for 250us, 100us and 1us time-base setting with the Probe Comp signal.

Self calibration fails at Step 7 of 16.  No error shown in error log.

The troubleshooting tree indicates the Acquisition board needs to be replaced, as the Self test passes the first 5 tests.

I guess the fault could be in the input amplifier ASIC (before the isolation), and related to a buffer-gain amplifier section that is used for lower time-base settings.  I can't easily see that this is a calibration setting issue, so am not yet keen to go through the Service Mode calibration procedure just to get stopped in that procedure half-way through.

I can still view higher time resolution detail by capturing a waveform and expanding it, although that is limited to 1us samples for the 250us time-base.

Has anyone changed an acquisition board in this series scope, or tried to faultfind on that board ? 

[tautech]

Looking at the last 2 screenshots I'd be checking PSU caps and any electrolytics associated the mainboard DC-DC converters.

[trobbins]

There just seems to be one processing action that is faulty, associated with the fastest sampling rate ranges, as all other functions and controls and processing are nominal, even the processing of either of those two example screenshots (ie. they can be time-base expanded to show that a signal is being sampled at the demanded rate, as well vertical gain controlled, and trigger control).  It's almost as if a different ADC is used for the fastest timebase settings, and the signal path to that ADC is being corrupted.  The Service Manual has a general description of the Acquisition board.  Prior to the isolation stage, each channel has its own power supply, which indicates to me that the fault is after the isolation stage, and related to the common acquisition ASIC which includes an amplifier before digitising.

I guess I could remove and inspect the Acquisition board to see if there was any sign of discrete part damage, but apart from that there may be nothing obvious except as you say looking for power supply electrolytics (I appreciate that poor electrolytics can sometimes cause quirky circuit behaviour, it's just that this fault seems a bit too discrete).

The only other post on repair of this scope has similar comment that Service Manual is quite deficient: https://www.eevblog.com/forum/repair/tektronix-tps2014-scope/

Service Manual: https://download.tek.com/manual/077030600web_0.pdf

[shakalnokturn]

Faults at higher timebases are typical of problems related to the FISO that isn't used at lower settings.
Check power to FISO ASICs and their temperature regulation.

[trobbins]

The TPS2000 series service manual indicates the ADG421 ASIC acting as FISO/A2D/Peak Detect/Trigger for a pair of inputs, and then an ADG522 for the DSP/Raster/Display ASIC for each pair of inputs. 

The tear down of a TDS2024 ( ) indicates the likely layout of the Acquisition board.  It uses the same ADG522 ASIC for DSP/Display, but uses other ICs for the post-isolator functions.  Given there is no schematic available of the board, or datasheet of the ASICs, it may be somewhat difficult to identify anything but basic powering to the ADG421 ASIC, and the basic signal inputs.  Still I will aim to break open the scope at some stage soon, and check around the ADG421 for any damaged parts.

[trobbins]

A quick exploratory look shows no specific ADG421 part, but an Altera Cyclone FPGA (EP1C4F324C7) and what appears to be associated controller (M68 SEC000PB10 L90M) and a labelled memory.  An LMS1587 LDO regulator is nearby. 

Each input channel has input circuitry that is covered, and then a differential driver AD8131 (Analog), mixer AD8343 DC-2,5GHz (Analog), and custom isolation transformer PT77172 (Datatronics), with a following AD8343 and AD8131 on the isolated side.

No obvious part damage on the top side, so at this stage I think I will just use the scope as is until some more details crop up that could assist fault-finding.

[mmatic]

Did you solve the problem?

I own TPS 2024 and during startup procedure it reports Acquisition CH1 CH2 FAILED. I found out that CH2 has some problems with zero (POSITION - vertical offset) setting. If I set zero to -1V it is -0.8V, if i set to zero to -2 offset is -1,7V,... and opposite on positive side. If I set to +1V offset is 0,8V, if I set to +2V offset is 1.7V,... Channel 1,3,5 are okay.

Does anybody have TPS2024 schematics or any idea what to check?

[trobbins]

The funny part has been that when I needed the scope recently to look at a 10MHz GPSDO output - the scope performed flawlessly!  I can only assume there was a flaky connection/contact that has somehow come good, and can only hope that the issue isn't within a part.  I can't associate it with a specific action, and can only hope that the fault doesn't reappear when I am again in need (hopefully soon if a PIC programmer comes in for the GPSDO). 

[lern01]

A quick exploratory look shows no specific ADG421 part, but an Altera Cyclone FPGA (EP1C4F324C7) and what appears to be associated controller (M68 SEC000PB10 L90M) and a labelled memory.  An LMS1587 LDO regulator is nearby. 

Each input channel has input circuitry that is covered, and then a differential driver AD8131 (Analog), mixer AD8343 DC-2,5GHz (Analog), and custom isolation transformer PT77172 (Datatronics), with a following AD8343 and AD8131 on the isolated side.

No obvious part damage on the top side, so at this stage I think I will just use the scope as is until some more details crop up that could assist fault-finding.

Sorry to bother you. Could you please check the U1701 chip model? Thank you very much!

[trobbins]

lern01, I looked at other photos I took, but can only discern the National Semiconductor logo on U1701 IC, and not the part model.  I'm not in a position to open up the scope to look further, as the original fault may return (to date the scope is still working ok).
Regards, Tim