HP Infinium 54845A spiraling death

[faroutliving]

Friend of mine in Florida dug out an old HP Infinium 54845A from storage that he needs to use for some RF work he is doing. When he first turned it on, it indicated the HDD was dead. Since he does not have much experience with the software (or digital) side of life, he asked me for some help. After looking around I found the answers here on EEVblog (what a great resource!) and suggested to him that it might be easier for him to ship it up to me and let me replace the HDD and install the software. I have most of the old equipment needed and scrounged the rest, and ordered a new HDD for the Infinium.

This is where things took a bad bend. The scope arrived with the issues above, but only about 1 out of 4 power-ups caused it to actually light up the screen and (attempt) to boot. I assumed he just glanced over that issue, and I proceeded to test the Windows 95 boot disk I created. That is when the power-up went from 1 out of 4 to never. The power supply fans come up, and the keyboard lights momentarily flash, but nothing else. Seeing that the motherboard has no speaker, I dug one out and plugged it in. According to the AMI BIOS docs I could find, the 8 beeps I now get indicates a missing video. My friend indicates that the display always came on when he pressed the power button, so it seems like some kind of progressing time bomb :-(

I've tried re-seating everything including the HP custom "display" board, and instead it has gotten worse where it won't even beep if the display card is installed. Of course, without the card it still complains about no video card :-)

I have a PDF of the Agilent Model 54835A/45A/46A Oscilloscope Service Guide, and I see where they use a ISA POST card and also details on checking the power supply voltages on the acquisition board. I'm also going to track down an old ISA video card to try. Has to be one lying around in those boxes...

!! Now the question !! I will pursue the provided diagnostic steps, but to what end? I see various Infinium ISA cards listed on eBay. How model specific are these cards? Can I use a display card from a different model if testing indicates a display card problem? Just a quick scan of the cards tell me that it is unlikely I will be doing a low-level parts swap on the card.

Thanks for any advice!

Deron

[mcovington]

I'm not familiar with this piece of equipment, but it sounds like a power supply problem, and you will probably find that you can replace discrete components on the power supply board even if you can't anywhere else, because they're big.  Start with the biggest electrolytic capacitors.

[shakalnokturn]

Same thinking here: Would start checking power and NVRAM backup battery even if it seems unrelated.
If it is based on a standard AT/ATX solution then strip the main board down to a strict minimum with a single RAM slot populated, keyboard (helps check if the BIOS is actually reacting) video card, no HDD.

[charlyd]

start with the basics and start with the voltages check... try to get your mobo back to live again. and make you problem smaller.

[Ice-Tea]

Not sure about this model specifically, but I've seen a number of 54825's with erratic behaviour, mostly due to the elcos on the rather pedestrian windows motherboard shitting themselves... Might be worth checking that out.

[charlyd]

for sure on the #vin033 mobo. leaking elcos

[faroutliving]

Thanks for all the suggestions. Good solid debugging.

What I was asking was a preemptive question. If I determine the video/ I/O card is the problem, would only cards from an 54845 work? Or is that card the same between all the HP 548xx? In other words, is the only difference in the capture board underneath the motherboard?

Thanks again,
Deron

[Ice-Tea]

Beats me. Show me the board in question and/or PN and I'll be able to tell

[Jwalling]

Thanks for all the suggestions. Good solid debugging.

What I was asking was a preemptive question. If I determine the video/ I/O card is the problem, would only cards from an 54845 work? Or is that card the same between all the HP 548xx? In other words, is the only difference in the capture board underneath the motherboard?

Thanks again,
Deron

That video card (Part # 54810-66525) is in: 54810A, 54815A, 54820A, 54825A, 54835A, 54845A, 54846A,  E4991A, E5070B, E5071B, 4287A
You can still buy one from Keysight for $473, which is quite surprising.

[charlyd]

is a memory upgrade worthwhile on these units?
if yes what is the best choice? 3x 256gb?

[faroutliving]

I am still struggling with this fragile machine, but this machine here requires memory to be installed in pairs. It has 4 slots so two pairs. Do you have the manual for your motherboard? That was one of the first things I did was find the manufacture documentation on the motherboard. It should outline what kinds of memory that it supports. By my recollections, they used a couple different motherboards so determining which motherboard you have would be the first order of business.

Deron

[faroutliving]

Making progress, but now another strange problem.

First, the power up/video card/booting problem appeared to be a cabling issues. A chewed cable underneath that connected the capture board to the front panel (guessing it was snagged when someone was closing the unit back up?). A couple ribbon cables were not seated square. And the HDD ribbon cable was connected one row offset to the video/ i/o card. I've corrected those issues, and the power supply test points on the capture board and the probe connectors as illustrated in the HP manual all spec out. No apparent capacitor leakage/bulging or other visible damage that I can see. Replaced the clock/battery chip and set bios parameters per HP documentation. I have not removed the motherboard or capture board to examine the back side of either.

It now boots to the Infinium splash screen but hangs there. External monitor and internal flat screen both work and display the same info.

The HP troubleshooting steps only appear to deal with display/power supply issues or capture issues. This problem lies right in between the two! It basically says if you don't see the default oscilloscope screen go to the No Display Trouble Isolation Flowchart. Following it, the Front Panel shows the power-on test complete pattern and so on. Infinium display appears (step 6), and that is the last step it completes. Next step #7 is "Attenuators click as solenoids are preset to known states" and "Front panel LEDs change to a normal operating pattern. Both screens go blank momentarily." Sadly, there is no Troubleshooting steps for this one.

I can pull up the Windows menu and go to the various apps. Some appear and then immediately disappear, but some like the calculator stay up. I'm guessing that the Inifinum Loader app is forcing the splash screen to top over normal windows but palettes/dialogs stay above. I can stop the Infinium Loader task (Windows reports it does not respond etc). I ran ScanDisk on the C drive, and "Thorough" test reports no errors.

An odd symptom is that the keyboard caps lock LED does not toggle when the key is pressed but the function does. Same for the num lock (I use it to move the mouse by keyboard since I don't have a PS/2 mouse I can find). Stranger still is that the screens (internal and external) invert momentarily when I press the num lock. The LEDs will resync when I press control-alt-del and bring up the task manager. This strangeness does not occur in the BIOS setup.

So it could be some simple software error because of corrupt HHD (seems unlikely), or some kind of video/capture card error (seems likely), caused by the poor reassembly at some point in the past? I feel stupid for not catching the error initially but since I didn't mess with them I didn't expect them to be wrong...

Any way to find out what is causing the Inifinium Loader to hang?

Thanks again. Even if you don't hit it on the nose, just getting me to think in a different direction can help!

Deron

[faroutliving]

Alright, looks like it is software? If I stop the Infinium Loader task in the , and then run the Infinium Scope app from the Programs submenu in the Windows menu, it will load up and all seems well. Passed the Self Test and it properly displays the calibration output square wave.

Any idea why the Loader would lockup and fail?

Curiouser and Curiouser

Deron