TDS784C-FAS Stuck on "MEMORY WRITE PROTECT SWITCH" after aborted Volt. Ref. Cal.

[micaticu]

Hello everyone,

First I will mention in short the present situation of my problem for those of you who may have already encounter this problem and may have the answer right away.

My TDS784C seems to be fully operational at present but I can not recalibrate it. I will explain later why I whish to calibrate it.
The Field Adjustment Software (started with ADJ700C.BAT) is communicating with the instrument and functioning normally up to the point where it prompts "MEMORY WRITE PROTECT SWITCH" asking to put the switch in unprotect mode. Of course I tried to change the switch position but to no avail. The software is not passing over this message, prompting it again and again no matter in what position the switch is.
I have been able to read/ write and verify the DS1650Y and DS1486 nvrams without problems trough the GPIB, using the Unified Tektool. I have also been able to try a clean version of firmware v5.3e and also v5.2e, using the same software (the scope came with v5.3e installed and functioning).
And to conclude the short presentation of the problem I have to mention that when I check on the scope the status of the current calibration, under UTILITY, I have the following results:

-SIGNAL PATH - Pass (and it passes every time I run it)
-VOLTAGE REFERENCE - Running
-FREQUENCY RESPONCE - Initialized
-PULSE TRIGGER - Initialized
-COMM FILTERS - Not present


And now, the longer explanation of how and why I am in this situation, and more details on my setup and what I have tried so far.

I have purchased this scope about six or seven years ago, fully calibrated and running. Reading extensively on this and other forums about the danger of losing the content of the nvrams, I have decided to acquire some tools to save their content and to be able to recalibrate the unit if necessary.

My setup:
-PC running DOS 6.22 plus a GPYB-PCII/IIA card and FAS 700C and 700D installed and running.
-Data Precision 8200 voltage/current source with GPIB
-HP8664A 3GHz signal generator with GPIB (but used in manual mode for calibration)

Upon reception and verification of the scope, I noticed a slight dip in frequency response around 600-700MHz and I decided to recalibrate.
I found and installed the Tektronix TDS700C Field Adjustment Software and started the calibration. I have done this a few years ago and I remember that it did not go very smoothly; I think because of a communication problem with the Data Precision 8200, so after many tries, I decided to try the version ADJ700D (instead of ADJ700C) of the FAS and that went better. I have even succeeded to automate the first part of the calibration (VOLTAGE REFERENCE) up to the very end where the FAS stopped short of saving the calibration. I have tried also to run the FREQUENCY RESPONCE calibration only, but unable to finish it. I have aborted and restarted the FAS from the very beginning a few times until, after having reached the same point as previously mentioned, system got stack, unresponsive, and I had to power off the PC.
From that point on, restarting the FAS, I have never past the "MEMORY WRITE PROTECT SWITCH" prompt even if the scope continues to run normally otherwise, but with the VOLTAGE REFFERENCE, FREQUENCY RESPONCE and PULSE TRIGGER stuck in the states mentioned above.
All this happened a few years ago when I received the scope, but I continued to work with the scope anyway, as it otherwise seamed to work ok, having kept the original calibration intact.

A few weeks ago I decided to revisit the situation, and I have looked what is new on the forums related to my problem. Some very creative and talented people have developed tools (new to me), like Tektool or Tekfwtool to communicate with the scope trough the GPIB port or to use the floppy as a backup device. I have used Unified Tektool and I successfully saved and rewritten the NVRAMs and firmware contents.
But I am still unable to make the FAS advance over the "MEMORY WRITE PROTECT SWITCH" prompt.

Trough many hours of research trough the forums I have found a single mention of my situation being possibly due to some corrupt eprom. In this YouTobe, -->  <--, at 7:27min, "tekhobbycap" mentions that you can get "stuck on 'RUNNING' " as I am with my VOLTAGE REFERENCE calibration but without developing on the subject.

I apologize for the length of this post, but I tried to give as much details as I could regarding my problem.

Any suggestions or advise would be greatly appreciated
Thanks

[m k]

"MEMORY WRITE PROTECT SWITCH" is obviously something more also.
Can you check how it operates?

There can be several traces and maybe one is bad.

[micaticu]

Hello,
I am note sure I understand your suggestion. Should I check traces on the CPU board of the scope?

Also you ask: "Can you check how it operates?"
Are you referring to the FAS (Field adjustment software) or the scope?

The scope is functioning without problems. I can use it every day.
Even the switch for the memory protection seems to do its job. If it is turned to 'UNPROTECTED' mode before powering up, the scope will start in standby mode with a black screen and waiting for communication through the GPIB. In this mode I successfully saved and rewritten  the NVRAMS and the firmware trough the GPIB a few times.

Only the FAS has a problem. It starts communicating with the scope through the GPIB, identifies the scope, and then, when I choose to run any test, it asks to turn the protection switch to unprotect.
If I do turn the switch in unprotect mod, it repeats the request  again and again every time after I press ENTER.

Please let me know if I understood your suggestions.

Thanks

[m k]

There are many possibilities but one is clear, memory can only be write protected or not.

The actual write signal is hardly going through that switch so there are more routes than one.
Routes can also be pull ups or downs, or something more active.
It's also possible that the signal from the switch is inverted somewhere in the middle.

Your situation is clearly indicating that the machine has at least 2 different feedback for that switch.

You have also used 2 different software.
If you know that both should work then it's clearly hardware that is not.

It's also possible that the signal is much more indirect and complex.
Something like one chip is seeing the switch and another is asking the first one how it is.
Then you clearly need a correct software asking correct questions.

[picburner]

In order to calibrate the scope it must be as shown in the attached photo. Each test should be "Pass".
Your situation may be due to a corrupt NVRAM or some hardware failure or both.

[micaticu]

m k

If I understand clearly, what you are suggesting is that my problem might be hardware related, and I should do some hardware tracing on the circuitry of the processor board of the scope, and I will do so again, even if I did it in the past without finding anything wrong. I might have missed something.

On the other hand I think that the timing of the appearance of my problem may suggest that there is a significant possibility that this is software related.
The problem occurred WHILE I was trying to recalibrate the scope using the Tek's Field Adjustment Software when I had to power off the PC and the scope at the end of the first calibration procedure (DCV).
This, together with the fact that the processor board looks to be working in perfect order and I couldn't find any tracing problems so far, leads me to bilevel that the coincidence that something in hardware went wrong exactly at the moment I was doing an action that the FAS probably did not expected (powering off the device) may be possible but unlikely.

I rather think that at the moment of powering off, the FAS was accessing some parts of the memory that are now corrupt.

I might add that I am trying with the help of Strick from the Tek forum to see if maybe a communication problem developed between the PC and scope. He just recently encounter the same problem I have (prompted repeatedly by FAS: "MEMORY WRITE PROTECT SWITCH") by substituting in one of his PCs a new motherboard with a 133MHz bus speed (instead of 100MHz).
Putting back the old mother board in the PC, allowed him to finish the calibration without problems.
The conclusion was that the FAS may not work properly with computers that run to fast.  Probably a GPIB timing issue. Go figure...

Anyway, I will try again to see if somehow  I have missed some traces while I verified the processor board.

Thanks

[micaticu]

picburner

Thanks for your input. I know that a calibrated scope should look like in the picture you show.

My problem is that my scope is in the state shown in the attached picture, and FAS does not want anymore to pass over the "MEMORY WRITE PROTECT SWITCH" prompt, even if I did not find any problem so far with tracing and signaling on the processor board and I installed fresh firmware (tried ver.5.2e and 5.3e) and new NVRAM content downloaded from the web.

In my previous post (addressed to m k) I explained a little bit more about this same situation encountered by Strick from the Tek forum.

Any more specific ideas?

Thanks

[m k]

GPIB timing is a good guess, even better if FAS is known to have issues.

Old motherboards had all kind of settings, like memory wait states and even jumpers for slowing things down.
But then the GPIB adapter must also be an addon board, if it is a dongle then those settings have no meaning.

Write protect shouldn't have anything to do with the memory it self.
Though FAS may have an error where certain value somewhere is mistreated.

If you try to avoid factory reset you can always compare memory dumps but that can be even longer route.

Your gpib-pcii/iia card has some settings.
Have you tried it without DMA and/or IRQ?

[micaticu]

m k

Indeed I have read in many posts and seen in many YouTubes about FAS being very finicky.

I am sure that I tried settings for my PCIIA  card with and without using interrupts (I did this recently) but I don't remember if I tried it with or without DMM. Maybe I did it in the past ( a few years ago), when the problem first appeared, but I am not sure.
I will try again playing with those settings this afternoon and I'll keep you up to date with the results.

Regarding the memory dumps I have read recently on the forums, but I am not sure of what to look for, as a clue for my problem.
Inspired by some posts talking about memory dumps and/or other interactions that can be made trough the serial port, I am waiting for a little adaptor I ordered from eBay ( https://www.ebay.com/itm/325165284729?hash=item4bb55c7d79:g:NXwAAOSwcfdbADO6), which would maybe enable me to get some more information trough the RS-232 port.
I will try though to be very restrained with pocking willy-nilly  trough the serial port, as I read that doing so might definitely brick the scope.

Do you have any ideas regarding this, or things and/or ways to look for in the NVRAM, firmware or some other memory places?

micaticudan

[micaticu]

m k

I have just tested all four combinations (On/OFF) for DMA1/IRQ7 on my PCIIA card (changing of course the corresponding setup in the ADJ700C.BAT file of the FAS software) but with no success.
The behavior of the FAS is the same; it starts communicating with the DUT (scope) and reaching the point where I select to run a test, the FAS is still prompting: "MEMORY WRITE PROTECT SWITCH...".

micaticu

[m k]

Back in the day delaying programs was an unknown concept but eventually it became very relevant, old games were the obvious target.
You can check a game called Zaxxon, its frame picture/control update rate is a direct result of machine speed.

So programs delaying programs emerged but I have no idea how they are today.
Maybe one is still available and you can delay FAS to crawling speed.
But it's not very obvious direction since FAS is pretty new.

For memory dumps you need a before and after dumps of what ever from the same machine.
Then you can start checking what is changed and why.

A space between low and high NVRAM addresses is 500k, so no help from there.
Possible relative start addresses are 131072 and 262144.
Incomplete addressing is also possible, there same data is duplicated.

CAL_START (
hwAInstrSerialNumberDigit(UNIT "" VALUE 327685)
DIPSWITCH_ADDR("8388608")
DIP_SWITCH(UNIT "" VALUE 0); model C & D
DIP_SWITCH(UNIT "" VALUE 1028); model A

Have you tried that A version of FAS?

INTERLEAVE (
PH_CAL_STORE("262144")

One possibility is also that your old interrupted calibration left behind a file or setting that is the cause of your situation.
PC side is easy, just check what are changed and when, no idea how it is done with the scope.

Pricking a scope is a real threat only when commands are too cryptic, long and have generally unknown parts.

[micaticu]

m k

Regarding the PC side, now I am waiting for a mother board and a hard drive image with a FAS installation that Strick, from Tek forum, fully tested and generously offered to send to me, to completely replace my PC hardware/FAS installation. This way I will be able to eliminate any suspicions regarding the PC side (speed, configuration, GPIB communication etc.).
If the problem persists, then it must certainly originate from the scope.

Regarding the scope side, I will install the little adapter I mentioned previously and try to, first, familiarize myself with the way the debugging port is operating.

m k, You mentioned in your email:
"CAL_START (
hwAInstrSerialNumberDigit(UNIT "" VALUE 327685)
DIPSWITCH_ADDR("8388608")
DIP_SWITCH(UNIT "" VALUE 0); model C & D
DIP_SWITCH(UNIT "" VALUE 1028); model A"

and:
"INTERLEAVE (
PH_CAL_STORE("262144")"

Are this commands I can send through the GPIB or the debugging port?

If for now this is a little bit to new or advanced for me, I will ask you guys, to point me to some posts or other reading material I can study in order to catch up with some of your knowledge regarding the TDS memory, software and such. I have time and perseverance. The question is if you will have the time and the patience to "hold me a little bit by the hand" in my endeavor 

m k, you also mentioned the "A version of FAS". I am not aware of what that version is. Are you referring to a TDS700A.ADJ version?
I will try to look on the web if I can find it. Do you have a link where I could download it?
I have tried so far the TDS700C.ADJ and TDS700D.ADJ. I have head in the past (before the "MEMORY WRITE PROTECT SWITCH" problem occurred) more luck with the "D" version, successfully starting the DCV calibration (but not finishing it completely).

I much appreciate the helping suggestions you offered so far.

Thanks

[m k]

No rush, we'll wait and see what your new hardware will tell us.
Maybe all will become clear.

The magic mambo jambo is a snippet from TDS784x.CON of TDS700x FAS.
https://www.eevblog.com/forum/testgear/tek-tds-3054b-vs-tds-744a/msg539818/#msg539818

VAR = hwAInstrSerialNumberDigit(UNIT "" VALUE 327685)
Is an internal function with a return value stored to VAR.
A bit later is
VAR = hwAOption1M(UNIT "" VALUE 327686)
VALUE difference is only 1 so serial number can't be that but it can be something like a validity check.

The whole thing then is input parameters for
TEST = CAL_START (

Later
TEST = INTERLEAVE (
tex = DUT_SET(
has actual console commands.
Others are more or less for internal use and the actual console command using them, if any, is something else.
The whole file is also an input for something else.

If you turn those NVRAM addresses to HEX you'll see that they are not starting from 0 but from 0x10.
The chip addressing is starting from 0 but first there are watchdog, clock and calendar stuff.

[m k]

I just saw TDS754D schematics for write protect switch.

Your machine is very likely the same so if you can change the firmware your hardware is pretty ok.
TDS754D A10 board has 74HC86 U1721, it's the actual write enabler.
One possibility still is that flasher software doesn't check the feedback from pin 3.

Write protect switch is also generating 12V.
No idea what and where that is doing.

[micaticu]

m k
Yesterday afternoon I have found the source to my FAS stuck on "MEMORY WRITE PROTECT SWITCH..." problem.

It turns out that the NVRAM content was corrupt in such a way that it precluded FAS to pass the point where it verified the position of the memory protection switch, but it did not gave any error when self testing at boot up. I don't know the exact inner workings of how that happened, even if I know that it was most probably because I powered off the PC and/or the scope at an inappropriate time, but I am glad I could fix it. Maybe you can give a more detailed explanation.

How did I solve that?
Using Unified Tektool program, I have written through the GPIB into the NVRAMs two new, clean images I have downloaded from the KO4BB site.
The FAS has started to function normally right after, even if I am missing now some of my previous installed options. But I can fix that later.
Now I can attempt again to recalibrate the scope.

Thanks for your input and your interest in trying to solve my problem. I might ask you other questions in the near future regarding the calibration itself and possibly an attempt to semi-automate it.

Best regards

[m k]

Good news for a change, not bad at all.

My guess for the situation is a wrong error message.
where the right one would have been something like this or that parameter is invalid.

For device specific operational stuff I'm the totally wrong person.
I can't tell you anything from it.

When the time comes put up a new thread to Metrology section, they will know it inside out.
Your device is also quite a bit liked one so the forum can have a very detailed instructions already.

[micaticu]

m k
I tried to find references to the tek scopes or the FAS in the Metrology section of the forum but I didn't find.
Maybe I didn't understand very well your suggestion for where to post my questions regarding the FAS configuration and usage.

Meanwhile I will try to post it in repair, as I did with my previous post.

[charlyd]

hi i was reading the topic i had a similar problem few years ago.
it was hardware related in the end my gigatronics 6062A
was not stable that give strange FAS problems. but you could try all steps manual and maybe with a little Lucky finish the hours taking calibration.

is your debug log ok?? i assume yes. and log in the unit is and stays empty?

[micaticu]

charlyd

is your debug log ok?? i assume yes. and log in the unit is and stays empty?

If you are referring to the "Error Log" on the scope, the answer is yes and no. I am in the process of testing and trying various things because my purpose is not only to calibrate the scope but also to semi-automate the process. Not because I intend to calibrate other scopes but as a challenge, and to have the satisfaction of the hobbyist.

So far I have succeed once to finish manually the "Voltage Refence" part of the calibration and to have it "Pass" and at that time my "Error Log" I believe was empty because I did not have any errors on the power on self test.

But immediately after that I have restarted attempts to redo the same "Voltage Refence" calibration in semi-automatic way and I am having problems with the FAS prompting me that it can not initialize the DP8200 through the GPIB, so I had to interrupt the calibration, and as a result... no, the "Error Log" is not empty now and the "Voltage Refence" on the scope shows "Running" or "Initialized"(I am not sure because I am not near the scope).

So this is still work in progress. I am trying to figure out what is the problem with the DP8200 GPIB communication.


If you are referring to the debug log that can be done through the RS-232 port when connected to the 20pin connector on the processor board of the scope, I have not tried that yet. I have just received yesterday the little adaptor for accessing the 20pin connector, so I will have to learn how to use that first, and how to communicate with the scope.
Do you have any experience you would like to share with how to catch data trough the serial port?