Gould Ultima / Nicolet Accura DSO software needed.

[Parus]

Any luck yet?

Recently I have saved one of these with the repair of itś PSU. On the disk I found the original (UK) installation files and utilities. These will install a kernel plug-in in the OS for communication with the custom PCI card and the scope program. I have upgraded the Tyan board with a new PIII/1Ghz processor and more memory. With XP pro installed no problems, after that with W7 no luck. 

A manual and maybe some of the special intelli-probes would be very nice

Oscar

[shakalnokturn]

Getting there slowly, it's mostly a working scope now.
I previously mentioned the CH4 front-end shielding having been removed... It does appear there's some kind of grounding problem on CH4, applying pressure on the front BNC's can add quite a bit if noise to that channel and others to a lesser extent.

Also a problem on 8 bit real time sampling at 2GS/s (4CH interleave) with maximum (1Mpt) sampling memory it just won't display the waveform. I'm not sure if this is a hardware fault or software problem.
Can other users confirm how their Ultima 500 reacts to this setting?

It would be great if you could get an archive of the original install files online somehow for people with dead hard disks. KO4BB's website is one of the most reliable places to go for TE documentation and has an upload option.
I'll try and upload the user manual there soon.

As for the dedicated probes I've never ever seen any around but must admit to never having searched.

[Parus]

Sounds like the huge thermal cycles have done their job, That board really gets hot with 160W on itś legs. Specially on the right hand side (channel 4) with the airflow on the bottom going from left to right. Inspection and reflow of that complete area, but most likely you have done that already.

I will create some zip archives of the different versions of installers, automated measurement scripting and reporting software Proview and put it on KO4BB if i can find it all together. The old HD is quite a mess with several Gb'ś of rubbish and many backups of the same.

For your question about the upper limit behaviour i have no answer yet. After many experiments i am now putting it all back together and a sticky message came up  about missing/corrupt skew and ets calibration data. A auto calibration does not solve this, "see manual for special calibration instructions" but i don't have ...... (little hint: my email is public ). Inserting a old calibration file from one of the many backups might solve the message but this is not the way to go.

Still some work to do.

[shakalnokturn]

Yes I noticed that (front) end got quite a bit hotter than the CH1 side.
I'll do a close inspection of the solders or bulk reflow I only hope it isn't the preamp ASIC that's whacked.

I'm not sure this could affect your ETS/skew error, mentioning it just in case:
After patching up my cracked resistor network (CH2) I was experiencing some intermittent offset problems on CH3, poking around at the acquisition board while running lead me to the resistor network on that channel. Problem was solved by reflowing the solders where the pins connect to the ceramic.

It's worth checking the power connector to the acquisition board and power header solders on acquisition board they run warm.

A couple of other disturbing findings about this scope: A lot of interconnect is through longish FFC's.
Hooking a probe to the probe compensation output without attaching the ground clip you can expect some DC offset, the current draw from the front panel makes the probe comp. output above ground...
Similarly there are PS test points at the rear middle of the frontend board. On my scope the +/-5V supplies only read around +/-4.85V most of the loss is in the FFC's between the acquisition and frontend board.

[Parus]

I have uploaded a zip to KO4BB with the utilities, a OS system driver for the Jungo PCI of the Tyan PC board, 3 full setup versions (V3.05/7/8) and a GUI only V3.10 that runs fine on the 3.08 drivers.  Many thanks to Shakalnokturn (and Kowilli) for the 3.10 GUI and a preview of the manual.  Shakalnokturn has also uploaded the manual to KO4BB, so soon the full package is available to download for those in need with a crashed or missing HD.

Yeh i have noticed some weird and shortcut engineering in it's design. Looks like it was a quick and dirty HP Infinium clone job with the same look and feel to get a piece of that market share for way less $ (but still 15K $). Don't let these things fool you, it may just be so by design. It got me with the repair of it's PSU, chasing rabbits based on some weird measurements. Just wondering how many of these where sold, and how many are still alive today.

My skew/ets problem is not an error, its just wants a calibration of that (with a very sticky boxed message)  but this is not a part of the standard auto calibration and there is nothing in the manual about it. I have been swtching around al ot between different OS's and setup versions and one time it did a not-so-standard calibration, but how i got it there i can't remember anymore. I was busy getting the system to work stable,, and calibration was step 2. It's a bit time consuming.

btw. 4-ch interleaved, full BW, 8 bit is one of the auto calibration tests

[shakalnokturn]

Thanks for the uploads...
Now because of that, once I get CH4 sorted, I feel I ought to install a windows XP on a much faster mainboard / CPU. No doubt that will get me into a whole lot of extra trouble.

What was wrong with your PSU?

btw. 4-ch interleaved, full BW, 8 bit is one of the auto calibration tests

Yes I noticed that as I've been through a few auto-calibrations myself.
Funnily it does not throw an error when calibrating that mode, maybe I'm foolish in assuming it should if something is wrong. 
Searching around the system files I noticed there were several versions of the Jungo drivers:
GNT_PCI_4700_9050_V503.inf
GNT_PCI_2730_V505.inf
GNT_PCI_4700_V522.inf
GNT_PCI_4701_V522.inf
GNT_SYNCHRO3_V522.inf
I was expecting my scope to use the GNT_PCI_4700_9050_V503 driver as the slow acquisition board is PLX9050 based but it doesn't. Maybe I ought to experiment forcing other drivers to see how it behaves on 4 channel interleave.


Does your cal error actually show anything obviously wrong when looking at a waveform in ETS such as a very jagged / jumpy waveform with sample points out of sequence?
I'm not sure where the calibration is stored and if all in a single place, I've attached whatever seemed calibration related on my scope (CalibrationData.dat / CalTestResults.txt / VGen.log) maybe replacing CalibrationData.dat and running an auto cal over that will be enough to satisfy yours...

[Parus]

If it ain't broken, don't fix it. Specialy a mainboard upgrade is a bit useless and might fail. It's very timing sensitive, changing the more advanced settings in the bios other than default often makes the system hang.  My NT4 system had errors that I could not repair with the standard install cd because it was a high security version (that made me smile), no USB support, obsolete network and a terrible user interface. I have upgraded CPU, memory (windows really likes that), networkcard and a extra USB card to create a bit of extra head room for a "newer" OS. With XP running it's just a bit more user friendly but not a better scope. If you are happy with your W2K install, just leave it.
Because this is a little bit difficult to accept for us explorers you can also add a second scratch HD and start experimenting.

For the search enigine archives:
The PSU in the Ultima is a third party Ferrus  FP250-710. Itś a low frequency (36kHz) push-pull smps with 3 separate switching circuits on a common PWM, 7 output rails at 700W max total.
A real pain in the nose to work on, no schematics, no info, compact, components hard to reach, testleads don't fit or hitting other components and because of the layout of the 3 circuits on the baseboard no clear separation between prim. and sec. So most of it is hot and can bite humans and equipment pretty bad what made live testing a bit tricky.

My PSU was very dead, no output, not looping in startup and no visible damage. First thing: check primary, all ok DC present. Next: gate signal at the regulated startup rail, present but low and not a happy PWM, at least not enough to pull the FET open. That signal comes from a driver transformer which is just a small ring core with 3 solid wire windings, not a real failure suspect. So I started chasing a short of some kind that pulled the PWM down in the other circuits.
Driver circuit second switcher, OK. Third switcher OK, feedback OK. PWM circuit: SG3525 fully functional but output weak. This is a common old age thing where the  open collector output drivers are getting leaky, higher voltage drop, lower output voltage and some extra heat that accelerates this proces. Chip will fail in the near future, specialy if it drives the transformer directly.. Simple fix, just replace. Now with the PWM on level again i'ḿ back to starting point, but this time on the primary side of the driver transformer. Desoldering it's leads, check: open ....  shoot.
Easy fix: made with a piece of wire a new primary winding on top of the other, and with this it came back to live.
Next big question is how is this possible? The 100mA running can't even scratch that wire so it must be a combination of thermal and the ultra-sonic coil beating.

PSU fact: it has a small 12V standby DC/DC supply that is powered from the main bulk capacitor. Without powerswitch the complete primary is always on including the divider/bleeder resistors. That explains the 15W standby power usage.

For the calibration i will try a "smart" approach with compairing the files of the old with the new (and yours).

A bit long weekend story, salut

 

[Parus]

After reading the service manual of a Gould 400 DSO i tried the same approach on the Ultima 500 and found the "hidden" section of setup and calibration.
In the GUI go to utils menu with a.o. auto calibration, push the last button of the vertikal screen colunm 4 times and an extra option "engineering" will appear.  In this section all the low-level magic can be monitored and calibrated.

I have cleared my sticky Skew/ETS warning, but meanwhile CH2 got very noisy and the trace is jumping around,  grrrr

[shakalnokturn]

RaymondMack: Thanks for the photos and too bad for the HDD ..
It looks as if the hardware doesn't share much in common with the Ultima 500 after all, the acquisition and front-end PCB's are very different. The common parts seem to be the PCI slow acquisition, front panel, LCD, CCFL inverter. (Maybe PSU?)

Parus: Nice find that secret service menu, I wouldn't have thought of it...
Maybe worth reflowing the resistor network on CH2 as mentioned in a previous post.
If you have removed the coax interconnect make sure the connecors have been mated correctly. If you don't align them perfectly the inner pin can easily plug just next to the two curved spring contacts, giving a poor contact and the symptoms you have.
If this has happened the spring contacts will need realigning slightly. Don't ask me how I know...

[shakalnokturn]

Went back inside and see if the CH4 intermittent noise could be solved. Took a few more shots.
In the end the most likely was that there were also weak solders on the other half of the shielding (the one that wasn't almost floating around) and that was causing the problem.
Took a few pictures and found a couple more details in there that didn't impress me: Lowish clearance between the aluminium plate that holds the BNC's and tracks on each input, what looks like solder paste residue, some found on the tops of a couple of SOT-23 packages, worse on the front sides of two of the FE ASICs, over several leads.
Also a tendency the shielding has to rip one of the tracks to a nearby capacitor... (Pictured on CH2.)

[Parus]

Hi, thanks for sharing these nice detailed pic's, can be usefull. I hope this will clear your noisy problem.

My patient is on a waiting list for it's inspection and reflow operation of ch2. A selftest also reveilled an attenuator problem on ch4. On the lowest range (zero path) everyting is fine, but on the middle or high range the attenuator kick's in with a relais click and the signal (and trace) are gone. So it will not only be a reflow session but also a probing one.

Not a happy scope yet