Tek CSA7404/TDS7000 repair project (Upgraded memory/CPU)

[Xyphro]

Hi again,

will not forget about this, saw it yesterday after writing my email also:
"That is why I said would like to see shunt voltage/prim current and gate driving to see is the current impulse happens during mosfet ON or OFF period."

Maybe I continue this evening, can take a transistor from the digital supply until replacement arrives. Have likely a 12h-16h job for a bringup in the more comfortable dual digit GHz domain today...

Best regards,

Kai

[elektropionir]

No need to hurry, do not risk damage to other parts. Golden rule is if it isn't broken, don't fix it.
Exception would be, if you already have the unit out, you could replace electrolytic caps of small diameter. I am not talking about large 1500uF nichicon placed at outputs (if i remember correctly) but those tiny 5mm elcos 63V some 100u and some other values. Those limit your lifetime and die all the time. Often causing other failures. I really do not like finding those in my PSU.

[elektropionir]

This is a short step by step on how I got Win-XP to work on TDS7104, and presumably other TDS7000.

1) I bought a used SU810 motherboard, that apparently came out of a Gateway PC. These are identical to intel MBs you can sometimes find on ebay. Intel has also deleted all the documents about their legacy products so finding any software for these was not easy.
Motherboard has a bios version P12, that i cannot find anywhere, and I cannot find anything about it.
This motherboard fits perfectly in the tray and ports align perfectly.
I placed 2 pieces of double sided 256MB RAM sticks for total of unbelievable 512MB.

USB Jumpers: Motherboard has 2 usb ports. There are two unpopulated jumpers behind a port. You must populate them and fit jumpers on position 2-3 and 5-6. Upper usb port is internally wired and used to talk to front panel. More info can be found in Intel SU810 PDF documentation.

2) I found a rare Celeron 100MHz FSB that work on 1.1GHz so much better than 500MHz that was originally there. I still cannot confirm if 133MHz fsb CPU would work here. I even found for 5 euro, unused aftermarket cooler for socket 370 and it is very quiet.

2.5) I kept old HDD as a backup but installed MSata to 44pin IDE DeLOCK 62495 adapter. I used it in few other instruments like CMU200 and it always worked wonderfully. And it is much faster.

3) Install standard Win XP. I used small and cheap 20 euro wireless keyboard and mouse and a usb dongle, because I like to use a mouse with my scope. I do not like to touch the screen of my instruments. It worked perfectly during installation. OS recognizes most of the hardware.

4) Most of the problems I had was with drivers for Texas Instruments PCI-1225 CardBus Controller. These were creating conflict on PCI bus and preventing PowerPC board to talk to the PC.
Drivers were luckily provided by user snoopy, but I believe that they were also posted in this thread in a zip file "tektronix.zip".
You must manually force OS to use the driver unsup.inf. and device will be registered as "unsupported" but it will work.

5) Graphic drivers for Chips and Technologies AGP Rage do not work with  WIN XP. The correct driver is in the same zip file under CHIPS. You must again manually force use of this driver in device manager.

6) Front panel driver is also there. As well as driver that windows shows as "other PCI bridge". That is driver for tektronix PowerPC board. It is in folder PCI2PCI.

7) Install IVI driver that can be downloaded from TEK website called IVIDriver3.9_066045315_0.exe. First install part in shared folder then COM.

8 ) Install TekScope driver/Firmware again found on tek website. File is called TDS7104_7054_V2.5.5_066031808.exe.

By this point VXworks.st would load and tekscope app should work.
I had issues where file TekScope.o did not want to load for some reason. There was an error shown in console. But I had a backup of hard drive with Win98 that was previously installed and working, and I copied the same file over and then it loaded without problem. I think this is probably something i did because I was experimenting trying to get it to work. And I installed and uninstalled bunch of these drivers multiple times.

That is all. I just wanted to write a short to the point guide since information in this thread is all over the place and it took me too much time to set this all up.


BONUS:
-HDD Raw copy tool is perfect for making backup images and cloning drives. I used it for all kinds of mediums, from ssd to compact flash.

-The old BL440 motherboard had Tektronix splash screen when PC was booting. This new one has an ugly Gateway logo that is a cube with cow colored pattern. I found this guide how to change splash screen in these motherboards. I will try it out later. Actually if you want tektronix splash screen you can skip the entire first part on how to prepare an image and convert it to .usr file. There is file called LOGODATA.USR in old win98 tektronix/bios folder together with iflash utility. But I will use iflash that comes with a bios for this motherboard. Hope it works.

[elektropionir]

OK, another small update. I managed to change the splash screen. I found a SU810 bios version P10 on rebyte.me
You can find it in the attachment for those who need it together with tektronix logo file. I know this is just visual effect but the cow logo was really bothering me 
There is another version of P10 bios update utility file SUAP10EB.EXE that is supposed to work straight out of windows but it did not work for me. This DOS variant works ok.

1) it is a self extracting file (runs in win xp, not in win 10), place it in new folder and follow the steps in the intel guide.
2) run.bat creates bootable floppy so copy the logo to it, and rename AUTOEXEC.BAT to AUTOEXEC.BAK. Name does not matter really but if you do not change extension the process tries to only flash the bios part, and it does not give you control in iflash utility. Everything is done automatically within few seconds and then you must restart. The user area where logo is stored is unchanged.

The rest is simple as described in the guide, you just update user data area with the logo in .USR format, and reboot.

[Gandalf_Sr]

I find this thread so interesting, it reflects a kind of "failure is not an option" engineering approach to be able to fix anything.  It even made me go look on eBay for a TDS7000 to fix up but the financial aspect doesn't look particularly compelling for me.  On eBay it seems to cost $1,200 for an "as-is" TDS7104 that I then have to spend a load of time and money on and even after I make it all work, I need a $350 TNC-BNC adapter for every channel and what I end up with is an old 1 GHz, 4 channel scope.

If I found a TDS7104 in a dumpster and was offered some TNC-BNC adapters for $100 each, maybe.

I'm not knocking you guys, just reflecting on some real-world financial aspects.

[Xyphro]

Hi Gandalf,

I compare such things always with prople spending time in solving Sudokus or crossword puzzles. Looks like a waste of time, but entertains. Certainly you are right, but solving technical issues is the major fun of engineering. And in worst case you learn a new problem solving technique or get to know a nice technical concept that other engineers designed, to learn from.

Best regards,

Kai

[Gandalf_Sr]

Hi Gandalf,

I compare such things always with prople spending time in solving Sudokus or crossword puzzles. Looks like a waste of time, but entertains. Certainly you are right, but solving technical issues is the major fun of engineering. And in worst case you learn a new problem solving technique or get to know a nice technical concept that other engineers designed, to learn from.

Best regards,

Kai

I agree 

[Lahay_Alexey]

Hi Gandalf,

The TDS7104 does not require a TPA-BNC adapter. It is equipped with conventional BNC. A TPA-BNC adapter is required starting from TDS7154 and higher.

[Xyphro]

Hi!

Short update from the power supply:
I have turned my isolation transformer into an adjustable DC supply with a rectifier+capacitor, so that I can inject higher voltages than 30V into the primary side.

After lot of debugging, checking, if the voltage feedback loop works, ...

It turns out, that my external load resistor on the +5V was not connected properly anymore.

Basically it turned out, that when this resistor is not present, the choke L13 begins to "sing", thus making audible noise. As secondary effect all supply voltages drop in that condition heavily.

After connecting the 15Ohm resistor again to the +5V, everything was all of the sudden working again, with 1 exception: The +-15V only has voltages of about +-12V to +-13V, as function of the load on the +5V.

While a load dependency is not surprising (only +-5V are regulated), the fact, that the +-15V is so much off is a bit surprising to me, because I think the whole intention of the supply is to generate a voltage higher than nominal with the switched mode supply part and then limiting it with a final stage of a discretely build up linear regulator, to get less power supply ripple and better accuracy.

I could not find any further fault. There are clearly not shorts in the secondary path. All diodes are properly working too.
Only the 2N2222 transistor on the RT/CT / ISENSE branch was broken + I replaced the FMMT619 and FMMT720, because I broke them myself due to a measurement fault.

The 2N2222 transistor did not break again. If this missing load on the +5V could have caused this?

Cross fingers, that the acquisition board is not broken due to a fault in the supply.

Best regards,

Kai
PS: I attached a measurement of OUT and the voltage accross the RSENSE resistor. It is clear, that it occurs, when OUT is turned ON.

[Xyphro]

Oh btw. Current ramps are also present on the ISense!

[Xyphro]

Strike, it works!

Loading both +5v and -5v with 15 OHM got btw. also the +-15V to its expected levels.

Have plugged the supply back into the scope:
Very happy for today, it works (see picture showing it still only half assembled) !!! Thanks Electropionir for your remote support in this thread.
Next Step is to get windows XP + ssd running.

Best regards,

Kai

[elektropionir]

Sorry for my late reply. I was occupied yesterday whole day.
That looks good. I am very happy that it works.
I am trying to understand about underlying fault to evaluate if it is likely that this will occur again.
I can understand that transistor failing in case of large current peaks as they were present on the shunt. That is completely possible. One simple preventative measure here would be for example to place a small micromelf signal schottky /40-60V/ in series with emiter of MMBT2222 and 619 ohm. That would take care of negative vbe and I think prevent the transistor from damage.

I checked some photos online, since i never took out PSU out of my unit.
From the primary setup I assumed this was a 2 switch flyback, however your last oscilloscgraph and presence of L13 inductor makes certain that this is actually a 2 switch forward. That is quite popular topology for stabilizing multiple uncontrolled outputs.

However such large currents can only (other than a hard short) be present in case of one of the diodes on the primary side still conducting at the turn on moment.
These diodes would usually have large reverse recovery current, they are old components, and that would be a huge current impulse that could excite such oscillations. No load conditions make that more likely but still...That is quite a reach in logic that would require more evidence.

Contact me if the problem appears again. Good luck with setting up windows. I just went through that few days ago as evident from my previous post.

I really like the different color matched diodes for different channels. Very nice!
Did you replace them yourself? What components did you use?

[Xyphro]

Hi again!

The idea with the diode is good. Maybe this prevents further breakdowns. But I agree, it does not feel like the full final solution yet. Let's use it and monitor further behaviour.

The LEDs I bought from Aliexpress. Could not find similar LEDs anywhere else:
https://de.aliexpress.com/item/32820923105.html

Theese were the closed ones, compared to the original ones. I just see, that they are not available anymore, but if you google for "1,5mm LED diode Licht Weiß Gelb Rot Grün Orange Blau", then you will find other available LEDs of the same type.

I could not identify within this thread, if SF810 supports Windows XP - only SU810 seems to be confirmed here, still will try it out and let know the outcome.

Best regards,

Kai

[awjennin]

I repaired the Power Supply CCA via another thread, but have had a heck of a time getting Tek Firmware to load... I have tried Win2kSP4, and now XP. I've loaded the graphics drivers (Asiliant), disabled the Intel drivers, loaded drivers for front panel, and PCI / PPC. The power-pc status led blinks "-", and even loaded the IVI drivers. Ive read through this thread like 3 times.. and although i'm usually pretty exhausted when i finally get time to work on this scope, i feel like i can't make any progress on this issue.

Every time I try to install Tek 2.5.5 firmware, I get an "UNSUPPORTED HARDWARE DETECTED" error... has anyone seen similar?

Note: My unit came w/o HDD, caddy, or working cd-rom. Currently, I have a WD Scorpio Blue IDE HDD (NTFS) plugged directly in to the 40-pin IDE cable, Secondary slot. Again, i'm not using the Hard drive caddy - so i've plugged it direct - after removing the little connector converter card.
I found a replacement for the CD-rom (now, a dvd-rw) which i plugged in to the Primary 40-pin IDE cable.

Everything installed fine, and SW loads, PPC blinks "-", and after boot i hear relays clicking on the Acq. board, and fans change speed. But each time I try to install that  Firmware software.....

(Thanks in advance!!)

[elektropionir]

Clicking relays and everything you described tells us that the hardware is working and that PC motherboard communicates ok with PowerPC. If you check the console output you will probably see that VXworks loaded.
What I do not understand is how can that happen BEFORE you install the TekScope 2.5.5 firmware??
That firmware is what makes vxboot and other folders to appear on C. You do not copy that manually.

Have you tried following the guide I posted few posts ago?

[Johnny10]

That is an interesting observation.
How does the scope function start without the software installed?

[awjennin]

@elektropionir
I had copied the vxboot folder on to the drive from a previous installation attempt (which was obtained through opening up ghosted images). I loosely followed your guide. However, it did call me to question my Mobo - Radisys SF810. Would it stand to reason that this was original component in TDS7054? Think there's any harm in loading TiN's bios in to it? i've spent 2 months on this, and i'm ready to trash the scope. if a bios bricks it, whats the loss at this point? ha
I can deal with troubleshooting HW,  but SW puts me in a bad mood.

@Johnny10
I had manually copied the vxboot folder on to the C-drive as stated above. When i do that, and the scope clicks as it initializes.. It gives me hope that there's a path to success. So, in the interim, i've renamed the folder to vxboot_old so that it doesn't initialize on boot.
When App SW blocks an installation because it thinks its not a TDS7k scope... its extremely frustrating. starting to worry that i got a custom HW-blend of scope that isn't going to be restored easily / if at all.
If i knew what specific area of HW was flagging the SW and causing the issues, at least i could be rationale in my troubleshooting.

Today i borrowed a Tek HDD caddy from someone and verified that there's no special wiring in the caddy that the installer's checking for. I also moved the HDD and DVD drive back and forth between Primary and Secondary IDE slots...  This did not allow the installer to run either, however, the DVD drive did not show up when in Secondary slot.. odd.

So i'm at a loss.. anyone have similar challenges?

[Johnny10]

Reading through those exceptionally long threads can wear you out.
But they do reveal various issues different users have come across.

I have repaired a couple of these scopes and if I remember correctly.
That is not the correct hard drive.
Also TIN and a few others have explained in detail how to connect to the startup software routine.

I wanted to upgrade operating system and memory but I have that very limited BL motherboard.
Yours is the more upgradeable board.
It is not as simple as upgrading your home computer hardware.
BIOS is Tek specific, those PC100 memory sticks are even layout specific.

I cloned the hard drive from one machine to the next and scope started up without a problem.
Only difficulty was changing serial number to get options to work correctly.
But I had same motherboard across units and ordered same vintage hard drive online, wasn't there a GB limit?

[Xyphro]

I installed yesterday on a fresh ssd windows XP on my SF810 mainboard with 512MB RAM and 900 MHz CPU.
It worked like charm with Electropionirs description (thanks!) and the files you find here from other contributors (also thanks) . Did not need to copy any files from the old HD, also the licenses are still there.

The look and feel is much better, much more responsive. The scope application performance itself did not change as expected :-)

2 hints: an external vga screen is required for setup & the native lcd resolution is 640x480.

Best regards,

Kai
Ps: I used an older 32GB SSD, an adapter with 32GB msata ssd got not detected by the bios at all.

[elektropionir]

@xyphro

Great news! Now we know for sure that SF810 motherboard works also. I expected it to work since they are almost identical to SU810.
I also had some issues initially with msata and adapter with a scope until I found a correct position for the jumper. When bios recognized the scope I had no more problems.

There are posts in this topic discussing importance of tektronix specific BIOS firmware for SF/SU810 motherboards. That is not correct. Multiple members (myself included) have stock intel BIOS. Actually I have some unspecified Gateway Desktop bios, that is a version of official P10 intel bios. The scope works, if operating system and drivers work correctly.
Furthermore, bios present in Tektronix/BIOS folder in the units that came originally with SU810 boards is identical P10 bios that was on intel website (before it was removed) and can now be found on other bios databases like rebyte.me.
I just wanted to write this in clear terms because I wasted time thinking that bios is important.

@awjennin

Disk adapter makes absolutely no difference. I did most of the repair work without hdd adapted, plugging the disk directly into the cable because it was easier to remove and connect to my PC to make a backup.
I would not mix files/drivers from TiNs scope since TDS7404 is significantly different than 7104/7054. I cannot assure you that it would work. Keep HDD and CD drive in the positions marked on backside of NLX riser board.

Try to follow the guide, format the drive, install fresh copy of XP and proceed from there. The only thing you need from other scopes are the drivers I described. And nothing else, and certainly not vxboot.

Also check the console output and you can post results here as others have done. Perhaps there is some obvious error shown there. I assume you already got correct header to DB9 adapter.

[Xyphro]

I tried different jumper positions out , did not work. But I have another adapter as the one you have. The same setup is recognized well on my R&S FSEM, which is I guess older as the tds7104. It also got detected by a usb to PATA converter.

Anyway, was happy to have found the older pata ssd.

[elektropionir]

It is not as simple as upgrading your home computer hardware.
BIOS is Tek specific, those PC100 memory sticks are even layout specific.

Dear Johnny10 this is a misconception. I believe that spreads from few very old topics on tek forum.
My unit works perfectly with non tek bios, and I know for a fact that others who replaced BL440 to SU810 have no issues with bios and the scope will work.
It is quite literally as upgrading home PC.
I have 2 sticks of 256MB 133MHz rated ram, from 2 different manufacturers Hyundai and Infineon, and I got those for 1euro on ebay. I ran memtest for 4hours without issues.
The important thing is that the memory is 2 sided, that it has 16 ICs soldered on both sides. Reason why that is important is that the bios for these old motherboards could not address more than 16MB per IC for some reason. I cannot find links now but I discovered articles from about 20 years ago describing this.

@Xyphro
Great, I was also looking for those old SSDs that are native pata, but they are really rare and often expensive. I found some no brand Chinese ones but I would not trust those. I have no basis for the assumption except that there are no data on reliability, and I can get known brand msata for same money...

[Johnny10]

As I said, I have repaired two of these units that I bought broken.
The BL board is different from the board you have. You are correct the memory needs to be double sided.
I bought 4 different types of PC100 and PC133 and only the one worked.

I did shy away from trying to find a SU810 board because of reading about BIOS problems.
Jwalling gave me the best memory tips at the time. Also that changing memory on Power PC board wouldn't speed the scope noticeably.
Have you found that also?
elektropionir you are quite capable.
I went back and read your posts after writing my first post, thought of deleting mine.
awjennin sounded a bit tired of it all and I wanted to say that there is enough information in this thread to fix the 7104 if you keep at it.
Xyphro, thanks for your posts also.

[awjennin]

Thanks to everyone who posts details on here!! This thread is a treasure trove of information, and i have read through it about 3 times now

@elektropionir
for some reason i missed the dedicated positions for CDROM/HDD on the riser card, which is also indicated in the service manual. My mistake! What I found tonight was odd - If the DVD-RW is installed in the Primary IDE, its recognized just fine. However, if it's plugged in to the correct Secondary slot, its not recognized. In fact, if i manually set it to CD-ROM and save&reboot, it comes up as "none". The HDD shows up and boots just fine in either Primary or Secondary slot... very odd! Also, I noticed that both 40-pin IDE cables that were in my scope are the same part number (174-4231-00). This is called out as the CDROM cable in the service manual. the HDD cable is called out as 174-4321-00 in the manual. However i speculate that this is just different length cable for cable management purposes. a 40-pin ribbon cable is a 40-pin cable, right? (high-speed signal-integrity aside)
When i deemed the original CDROM bad 4 wks ago, i recycled it. Now i'm really kicking myself - as i wish i could test that datapoint. I found another laptop drive (another DVD-RW), and it is not recognized in the Secondary IDE slot either.

@Xyphro - i really appreciate your comments on your install. Its an excellent data point to know that WinXP on an SSD, SF810 mobo, with 512MB is a valid combination. That helps me exclude those variables when troubleshooting. Much thanks!

@Johnny10 - I think i made a mistake early on, starting to read through this forum half way through - and because many people are troubleshooting a scope that is simply not launching Tek app software, i got the idea that the vxboot folder was a prerequisite for installation. However based on your comments and rereading this forum, i understand that the "Tek Firmware 2.5.5" (for my tds7054) includes the scope software, as well as vxboot, and other critical files. Ya, i am getting a little exhausted on this troubleshooting. However, my OCD keeps me coming back each night to continue hammering on it. (that might include a real hammer one of these days - lol)

[awjennin]

In a display of exhaustion and confusion, i was lying face down on the floor, contemplating many things...  then it hit me, i tossed the old CDROM in the recycle - but i hadn't dropped it off at the scrap yard yet... So, I checked and sure enough it was still there. Brushed it off, and installed in to the Secondary IDE location. the bios recognized it!!

Also, I mis-diagnosed the cdrom as a bad drive. It turns out that its just old enough (2002) that it doesn't read CD's that were burned at high speed. The OEM WinXP CD is read just fine, though. These old pc inter-op details are starting to come back to me from my earlier days...

So, I reinstalled XP sp3 again (on 160Gb PATA, NTFS formatted to 146GB), just to get a clean start. Installed the graphics, front panel, pci drivers, and IVI softwares (and also putty, 7zip, and notepad++). PPC board is booting through step L (it attaches to backplane/network, and attaches DIO client... but can't find the vxboot folder (as i haven't installed yet), so it then reboots and tries again. looping.)
Then, I attempted to install firmware from TDS7104_7054_V2.5.5_066031808.EXE, yet I still get a failure....  out of ideas. all I can imagine is that:
1) theres an interface / memory that the installer is polling, but it doesn't ack or the data is wrong, so the installer pukes.
2) The HDD is too big, and i either need to partition it down to 16GB or similar... or simply find an old PATA HDD that's smaller.
I will keep digging..

So my contributions tonight are these:
1) certain DVD drives are recognizable in the Primary IDE,  but NOT the Secondary! This includes:
  - HL (Hitachi-LG) GMA-4082N-Z DVD-RAM
  - Toshiba/Samsung TS-L462 CD-RW/DVD-ROM
2) The original CD-RW that is recognized in Secondary IDE is a TEAC CD-W28E