Author Topic: Tektronix 2445 - bad microprocessor S68A08?  (Read 22682 times)

0 Members and 1 Guest are viewing this topic.

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #25 on: April 28, 2015, 06:12:36 am »
Hi Mskobier

I tested every IC which is connected to data, and address bus excepting the U2134/2034. I soldered the MCU directly to the PCB to make sure the socket is not the problem, but the problem remains. Sometimes the A/B trigger flashes, sometimes few leds lights, and sometimes(rarely) i can see the display, but timebase not running and scope doesn´t respond to any key.. Now i am suspecting the EAROM itself, or the MCU. Next step is to make a tester for the EAROM and try to get some data from it. Good luck with your scope! :)

Edit: The fight is over, MCU is definitely dead. Also in Diag mode, there is no activity on the address bus. In detailed wiew the surface of the pins is oxidated, and i think this oxidation bites into the package and gradually damaged internal pinning. Maybe it was just a matter of time, when this happens. Maybe the bad ESD care was made from my side. And maybe combination of these two things. I´ll try to get MCU programmed for my A5 board version, or the whole A5 board..
BTW thanks to all for help :-+
« Last Edit: April 28, 2015, 05:18:26 pm by stephunk »
 

Offline mskobier

  • Regular Contributor
  • *
  • Posts: 97
  • Country: us
  • Test Equipmentaholic!
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #26 on: April 28, 2015, 06:09:09 pm »
Stephunk,
    Bummer on the mcu being toast. Fortunately, they are readily available and cheap! I thinking I have a bad processor or the two ICs controlling the data lines are bad in mine. According to the troubleshooting tree, that's where it leads me, or the ram, rom, or processor is bad. The tree splits there and leads me down either path. I have pulled the roms from my board and read them in the programmer. I then compared the rom contents to a set of known good roms I downloaded from the net. All data was exactly the same, so that rules out the roms. So that leaves the processor, ram, or the two data control IC's. I'll get back to this scope in the next couple of weeks. Learning a lot and having fun!.

Another poster had said the roms in the 2445 and the 2465 are the same. If so, you are welcome to a copy of the bin files that I read from my roms so you can compare data. Just let me know.

Mitch
 

Offline PA0PBZ

  • Super Contributor
  • ***
  • Posts: 5121
  • Country: nl
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #27 on: April 28, 2015, 06:15:29 pm »
I´ll try to get MCU programmed for my A5 board version, or the whole A5 board..

I don't think the MC6802 has a rom, so not sure what you mean by 'programmed'?
(you say S68A08 but the schematics show MC6802?)
« Last Edit: April 28, 2015, 06:17:48 pm by PA0PBZ »
Keyboard error: Press F1 to continue.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #28 on: April 28, 2015, 06:57:45 pm »
I´ll try to get MCU programmed for my A5 board version, or the whole A5 board..

I don't think the MC6802 has a rom, so not sure what you mean by 'programmed'?
(you say S68A08 but the schematics show MC6802?)

You're right, it does not have a ROM (or flash).

08 parts are identical to 02, but 08 doesn't include the on-chip 128 byte RAM.  Tek has the RE (RAM Enable) input permanently low, so the internal RAM is never enabled anyway, hence allowing use of either part.

A versions are 1.5MHz,  B are 2.0MHz, and no letter is 1.0MHz.  From earlier in this thread we figured out the cycle time is 1.25MHZ so we need at least an A part.

I have no idea why the schematic and parts list says vanilla 6802.  It threw me too.
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #29 on: April 28, 2015, 09:10:07 pm »
I didn´t study the datasheet for 680X family in details, because as it is a microcontroller, i thought it has to be programmed. This is a part of electronics where i´ve got a lack of experience and also this device itself is older as me ;) So, if i will somehow get the MCU, put it into board, and everything other is OK, should my TEK be alive again?
And yes, the schematic says MC6802(Motorola), but on board is S68A08P(AMI semiconductor).
 

Online edavid

  • Super Contributor
  • ***
  • Posts: 3381
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #30 on: April 28, 2015, 09:27:00 pm »
I didn´t study the datasheet for 680X family in details, because as it is a microcontroller, i thought it has to be programmed. This is a part of electronics where i´ve got a lack of experience and also this device itself is older as me ;) So, if i will somehow get the MCU, put it into board, and everything other is OK, should my TEK be alive again?
And yes, the schematic says MC6802(Motorola), but on board is S68A08P(AMI semiconductor).

In the old days we would distinguish between a microprocessor (MPU) and microcontroller (MCU).  The 6802/6808 is an MPU, not an MCU.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #31 on: April 29, 2015, 12:20:24 am »
...
So, if i will somehow get the MCU, put it into board, and everything other is OK, should my TEK be alive again?
...
Anything's possible.  I thought you had it partially working before with the 74HC138 with the EAROM errors.  I think the scope will still run with a bad EAROM while you debug it (although you will have power-on errors).

It sounds like there are now other problems, and you've disturbed a fair number of components.  I think the chances of popping in a new processor and having it work are not as good.

But don't get me wrong - if you find one, you might as well try it.  From prior testing you know for sure it was bad.  I just think you have more troubleshooting ahead of you.
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #32 on: June 01, 2015, 08:24:10 pm »
The new MPU´s arrived few days ago. I ordered MC68B02P, put it into socket and tradaaa, i have the readout and the trace both. I need to check the performance in details now.
Best wishes to all of you sirs :-+ and thank you for your help. I learned a lot with this scope! ;)
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #33 on: June 01, 2015, 09:08:41 pm »
That's great!  With no power-on errors now?

Hopefully your calibration is intact too and it was all a processor issue.
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #34 on: June 01, 2015, 09:52:23 pm »
Failed test 04, error code 11. So there is an issue with earom, bad checksum and/or parity error on read..By pressing A/B trig, it turns into normal operation. Tomorrow i´ll look closer again  :)
 

Offline mskobier

  • Regular Contributor
  • *
  • Posts: 97
  • Country: us
  • Test Equipmentaholic!
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #35 on: June 02, 2015, 04:09:49 am »
Stephunk,
       I have an early version 2465CTS that I have copied the roms to BIN files. For most purposes, they should be the same as is in your 2445. If you are interested, I can send the files to you to try and replace your current corrupt roms BINs. The one I have has four roms. I also have a later (I think) A5 board that only uses two roms. I have not copied them, so I do not know if they are good. After I copied the roms to file, I compared them to what is supposed to be known good versions I downloaded from the net. The files matched perfectly, with no differences in the code for all four roms. Something to keep in mind, is that the roms in that model of scope are UV roms. What that means is that they can only be written to once. If the data needs to be changed, then the entire chip needs to be erased using a UV light source. Then the chips can be reprogrammed via normal programmer. I haven't looked lately, but I think those chips are still available, at least in NOS.

Let me know if you want the BIN files.

Mitch
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #36 on: June 02, 2015, 08:31:04 pm »
Mskobier
I already have some bins from 2465, and it doesn´t match.
I´ve got installed this UVEPROMs: 160-1625-06,160-1626-06,160-1627-06,160-1628-06. I think, that each such numbered part has its own unique content.. So if you have 2 chips, i.e. 160-1625-04 and 160-1625-06, dumps from these two will be different. I proved it right now, using Vbindiff tool. However i am looking forward to your dumps, feel free to send it. I´ll be glad ;)

 

Offline mskobier

  • Regular Contributor
  • *
  • Posts: 97
  • Country: us
  • Test Equipmentaholic!
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #37 on: June 03, 2015, 01:42:21 am »
Stephunk,
     What I meant was that the code from each of the rom chips matched the downloaded rom files, not that they all had the same code. So rom 05 from my scope matched the rom 05 that I downloaded. The same for 06, 07 and 08. The bin files are in another computer that is in another location. I'll get them copied and send them to you. It will be early next week before I can get to that computer.

Mitch
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #38 on: June 03, 2015, 02:49:19 am »
As explained in the "Diagnostic Routines" section in the "Troubleshooting" chapter, the ROMs are tested with a checksum on power-up, before the EAROM.  They're fine since no error regarding them was displayed.

Do you have some other indication to suspect them?

If not, now that the processor is replaced, I would suggest rewinding your testing to post #18:

  https://www.eevblog.com/forum/repair/tektronix-2445-bad-microprocessor-s68a08/msg651562/#msg651562

And try to find the problem with the EAROM.  Your cal data may still be in there, but inaccessible.
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #39 on: June 03, 2015, 05:53:24 pm »
I don´t suspect the UVEPROMs any more, since i powered up the scope with new MPU. Everything is clear. Scope is running with error message related to EAROM and associated circuits. I am just curious, if the UVEPROM dump from Mskobier is same as mine or isn´t.
In case of EAROM related error message one can assume, that there is error in checksum(maybe i lost calibration data), or parity error. Next fact is, that 1V amplitude is displayed as 10V.. i keep going to study  ;)
BTW, I did exercise 02, and read content of some addresses. There are some data, but at this point i don´t know, if these are correct.
« Last Edit: June 03, 2015, 06:05:59 pm by stephunk »
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #40 on: June 03, 2015, 06:46:50 pm »
BTW, I did exercise 02, and read content of some addresses. There are some data, but at this point i don´t know, if these are correct.
There's hope!

If you decode the hex values, do you notice if any bits are always stuck as a 0 or 1?  Only locations 0x00 to 0x63 are used, and they are a 14-bit words (so the top 2 bits will always be 0 anyway).

In particular, cal data is stored at locations 0x01 to 0x4C and are also a 14-bit word, but with 13 bits of data and one bit of odd parity.

Details in the in the "EAROM Exerciser" section.


EDIT: No!!  I got that wrong.  I was thinking this EAROM had a nibble bus but it does not.  It's purely serial.  Sorry about that.

It might be useful to know, however, if there's any pattern to the failed values (like all 1's), especially in the cal data, or if the read back values are not consistently the same each time.

It's possible the contents may just be corrupted.
« Last Edit: June 03, 2015, 08:42:45 pm by MarkL »
 

Offline mskobier

  • Regular Contributor
  • *
  • Posts: 97
  • Country: us
  • Test Equipmentaholic!
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #41 on: June 04, 2015, 03:08:56 am »
Stephunk,
      I got a chance to get to the other computer this evening. Here are the BIN files if read from my UVROMS. Hopefully they are of some use to you.

Mitch
 

Offline stephunkTopic starter

  • Contributor
  • Posts: 24
  • Country: sk
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #42 on: June 04, 2015, 05:58:39 am »
MarkL
Yesterday i ran the exerciser02 and each time I read the EAROM content i can see the same values. Every value has a maximum 14bit width. Every calibration constant has odd parity, but the last one, on address x04C has 0082 value with X(EVEN parity). In the entire range of cal data(x01-x4C), there should be odd parity everywhere i guess, am i right?

Mskobier
Thank you, i´ll check this in the evening.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #43 on: June 04, 2015, 08:26:51 pm »
MarkL
Yesterday i ran the exerciser02 and each time I read the EAROM content i can see the same values. Every value has a maximum 14bit width. Every calibration constant has odd parity, but the last one, on address x04C has 0082 value with X(EVEN parity). In the entire range of cal data(x01-x4C), there should be odd parity everywhere i guess, am i right?
Right, 0x01 to 0x4C should all be odd parity with no "X" displayed.


Looking back at the error, "FAIL 04 11" is a parity error and a checksum error on the cal data.

If the parity for locations 0x01 to 0x4B are ok and it looks like there's reasonable data in there, chances are good that it's complaining about that one location, 0x4C.

A "spiral-add" checksum for all the cal data is stored at location 0x00.  So, if we know the checksum and we know there's only one bad word, it might be possible to reconstruct the bad word.  I don't know exactly what algorithm they mean by "spiral-add", but it could be figured out.  But then the problem is how to write the correct word back to the EAROM at 0x4C.  That could be done too, but it's not trivial.


It would also be useful to know what location 0x4C affects.  Watching the EAROM address for 0x4C while changing settings on the scope might shed some light on it since the scope will access that location when it needs to know the number.


I had an EAROM on a 2465 that was reading out bad constants some of the time.  Providing the EAROM with a slightly lower operating voltage allowed me to dig the constants out, but since you're getting consistent results, that's probably not the case here.


And last but not least, there's always going through the standard calibration procedure to rewrite the value.


If anyone knows of a hidden menu to write arbitrary stuff to the EAROM, let us know!
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #44 on: June 04, 2015, 08:46:32 pm »
In case of EAROM related error message one can assume, that there is error in checksum(maybe i lost calibration data), or parity error. Next fact is, that 1V amplitude is displayed as 10V.. i keep going to study  ;)
It might be related to the EAROM.

Care to elaborate a little on how this test is set up?  A factor of 10 could just be a display factor that's wrong for the probe being used.
 

Offline mskobier

  • Regular Contributor
  • *
  • Posts: 97
  • Country: us
  • Test Equipmentaholic!
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #45 on: June 04, 2015, 10:03:16 pm »
All,
     A quick question. I had the understanding that once this type of rom was written to, you had to use the UV light to erase the chip before you could write to them again, or change any data on the chip. If that is so, how does the scope re-write to the chip when you perform a calibration?

Mitch
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #46 on: June 04, 2015, 10:15:45 pm »
The EAROM (U2008) is an old incarnation of what would be called an EEPROM today.  That's where the cal data is stored.  It also holds the front panel settings.

The scope operating system is stored in separate EPROMs (U2162, U2378, and in earlier models also U2362, U2178), which have the UV erasure window.


EDIT: I meant to say "incarnation", not "implementation".
« Last Edit: June 05, 2015, 06:43:47 pm by MarkL »
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #47 on: June 05, 2015, 06:36:57 pm »
...
It would also be useful to know what location 0x4C affects.  Watching the EAROM address for 0x4C while changing settings on the scope might shed some light on it since the scope will access that location when it needs to know the number.
...
I connected a logic analyzer to the EAROM of a 2465 (which uses the same operating code as the 2445), and I'm unfortunately unable to figure out what settings cause reads from location 0x4C.

I can see that it's being read once on boot, along with a somewhat random group of other locations.  And then it's read again in sequence with the other cal locations.  The second access is no doubt the parity and checksum confidence tests.

Then it's never read again no matter what settings I try.


If anyone else wants to try, I found that Tek swaps the 10's and 1's place of the EAROM address.  So, location 0x4C (decimal 76) is actually stored at EAROM address 67.  It's easy to verify you have everything connected properly by using the EAROM Exerciser.

This 2465 has GPIB and CTT options, so it may have also something to do with that.  Who knows.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #48 on: June 13, 2015, 04:04:19 pm »
MarkL
Yesterday i ran the exerciser02 and each time I read the EAROM content i can see the same values. Every value has a maximum 14bit width. Every calibration constant has odd parity, but the last one, on address x04C has 0082 value with X(EVEN parity). In the entire range of cal data(x01-x4C), there should be odd parity everywhere i guess, am i right?
Right, 0x01 to 0x4C should all be odd parity with no "X" displayed.


Looking back at the error, "FAIL 04 11" is a parity error and a checksum error on the cal data.

If the parity for locations 0x01 to 0x4B are ok and it looks like there's reasonable data in there, chances are good that it's complaining about that one location, 0x4C.

A "spiral-add" checksum for all the cal data is stored at location 0x00.  So, if we know the checksum and we know there's only one bad word, it might be possible to reconstruct the bad word.  I don't know exactly what algorithm they mean by "spiral-add", but it could be figured out.  But then the problem is how to write the correct word back to the EAROM at 0x4C.  That could be done too, but it's not trivial.
...

I was interested in figuring out this "spiral-add" algorithm, so I took a look with a logic analyzer and a disassembled version of the 2465 EPROM.  For posterity, I've attached a C rendition of the algorithm.

What's interesting, as it applies to your parity error, is that the Tek code truncates the checksum at EAROM 0x0000 to 8 bits instead of preserving all the bits.  No idea why.

In doing so, the recovery of a bad 13-bit value at your EAROM 0x004C is likely going to have multiple possible values that map to the same 8-bit checksum.  There's no definitive way to tell which combination is the right one.

Although it's not guaranteed there *must* be multiple solution values, chances are not good there's only one.  It depends on your specific EAROM data.  The only way to know is to use your actual data and compute all the possible solutions.


And I still don't see where EAROM 0x004C is used.  It would still be nice to know if this value even matters.  Now that I have the code disassembled maybe I'll poke at it a little more.

But I'd say it looks like it's time for a re-cal to clear this error.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Tektronix 2445 - bad microprocessor S68A08?
« Reply #49 on: June 15, 2015, 08:53:11 pm »
...
And I still don't see where EAROM 0x004C is used.  It would still be nice to know if this value even matters.  Now that I have the code disassembled maybe I'll poke at it a little more.
...
Ok, I think I've got this piece figured out too.

EAROM location 0x4C, as mentioned before, gets read at boot time.

More digging with the code and the logic analyzer shows that once it gets read, it's stored in RAM as a word at locations 0x010F and 0x0110 (MSB then LSB).

RAM 0x010F/0x0110 is then accessed whenever the delta-V button is pushed and the horizontal cursors appear, no matter what traces are currently on the screen.

This leads me to the conclusion that EAROM 0x4C is the calibration constant that is derived in steps "a" and "zz" in the "CAL 02 VERTICAL" adjustment section.


So, stephunk, I would expect the only thing this bad value is affecting is the delta-V cursors.  Unfortunately, it's only set at the very end of the vertical cal procedure, so I don't think it's possible to do a cal just on that.  You'd have to go through at least the vertical cal, and the manual doesn't generally recommend doing partial calibrations anyway unless you're sure the previous calibration sections are ok.

The only other thing I think you could do is build a circuit to write values to the EAROM, and then try different values that 1) makes the the delta-V cursors work properly, and 2) satisfy the checksum algorithm

But that's a lot of work.  It's easier to buy or find someone with the needed signal generator to do the cal.  Maybe search around for people who have done ad-hoc solutions to calibrate these 24xx scopes.


I think I'm done.  I seem to be the only one left in this conversation.

EDIT: Added pic of 2465 being probed.  Fun!
« Last Edit: June 15, 2015, 09:02:21 pm by MarkL »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf