Agilent 54825A Repair

[aXit]

Hey guys,

I recently acquired an old broken scope, an Agilent 54825A (4ch 500MHz, independent 2Gs/s sampling) , VIN # 033. It has all the extras like LS-120 drive, sound card, LAN card, GPIB, and has the acquisition board with the hybrids instead of the ceramics. Anyway, it'd be worth the repair even without all the extras, so if anyone could shed any light that'd be great.

Anyway, it came to me in great physical condition, but with what seemed to be a software bug. The waveform display area was solid orange. Still had a UI, menus worked, relays clicked when changing vertical scales, just always with a solid orange in the area where waveforms should be.

Easiest fix was software, so I followed this guide for recovery (using the external CDROM drive method): http://pdub.net/projects/agilent-recovery-cd/ . Which had no effect. The orange seems to come from the default marker colour in software, changing the marker colour in menus will change the colour of the fill on the screen.

After reading the service guide, figured out that the windows UI is overlaid with the waveform display by the VGA mux, I started some hardware debugging. Seems to be that the display board is working fine, as the windows UI gets through, and the card can display both on the internal LCD and an external VGA monitor. The acquisition board also seems to be fine, as there is lots of high speed digital data coming across the W11 ribbon cable (acquisition->interface). However W14 (interface->display) seems to have no data, probing it out shows 2 clocks, and the rest of the lines are either held high or low.

So to me this seems like the interface board is faulty, it's receiving data, but not sending any data to the graphical buffers on the display board. However I'd like your opinion on this, as I'd like to be pretty sure of the diagnosis before dropping $500 on a replacement interface board.

If you had any other ideas or suggestions that'd be also very appreciated. Thanks for your time. 

[free_electron]

The graphics card is most likely dead. They use a special dual ported memory i believe. C&T6555 chipset.
The other board has the altera and battery.

I have those boards. (More than one)
Where are you located

-people please fill in your country ! In your profile.

[aXit]

I'm in Australia.

Interesting that you reckon it's the display card. The dual port memory is on the interface board with the altera FPGA. Do you know if the graphics chipset or the FPGA is clocking out the data from the VRAM? Just quite confusing why that bus (interface->display) has no data across it, even with active waveforms. There's a 25MHz clock that I can see on that interface, that seems reasonable for the 18ns access time on the VRAM.

The display card is much cheaper as a full replacement, can find a few on ebay for around $200. But it wouldn't be hard to replace the VRAM memory, and I've just found a seller in australia selling it at a reasonable price.

Thanks for your help.

[aXit]

Did some more probing, seems like the VRAM isn't connecting directly through to the ribbon cable. The high speed output seems to go into some agilent stickered PLDs, which have lots of connections to some cypress SRAM. These then connect through some 74 series logic, and then 33R series termination before heading out to the ribbon cable. Doesn't help that the board is 6 or 8 layers, and can't follow any traces or data buses.

Out of those though, seems to be that the most likely failure is the old VRAM, like free_electron said. I'll probe deeper when I get a PCI extension, assuming this doesn't throw off all the timings. Or if anyone recommends the direct VRAM replacement, I'll do that straight away.

[aXit]

So, found some VRAM on an Australian ebay auction. Pulled from a working Power Macintosh. Exact same VRAM chips with same timings. Anyway, replaced those two chip on the interface board and.... Absolutely no difference. Damn.

So free_electron, were you saying that there's more dual-ported memory on the graphics card that might be the culprit? I think I've sourced a replacement chipset, the C&T6555. Did you have any more ideas on what could be causing this?

Thanks again for your time.