Sony-Tektronix 308 Data Analyzer

[jrharley]

Greetings TE junkies,

A Sony 308 followed me home from a swap meet awhile back. Unit is in very clean shape, powers right up, nice crisp screen. It displays what the manual describes as a fatal error- 03,04, which indicates bad RAM. Since I know nothing about this unit, or anything like it, my ignorant question is this:

Will the unit recognize new RAM chips? And if not, what's involved in getting it to do so? Be gentle, I'm really weak with digital equipment.

Thanks in advance for any insight!

JRH

[TurboTom]

If you can find genuine 2114L 1kx4 SRAM-chips, swapping them in shouldn't be a problem and the instrument should perform as expected.

Despite the 308 is a cute and at its time probably a versatile instrument, nowadays it's got little but collector's value. It may be useful if you're into repairing of 1980s era instruments that are microcontroller-driven for signature analysis, though. Got one of these oldtimers myself as well... 

[jrharley]

Thanks TurboTom! That's actually encouraging, appreciate the response.

Yes, I thought it might come in handy for the stuff I work on. Vintage HP gear from that era for example, comes across the bench now and again. This unit has the basic probe complement I believe, so, if I can get it booting up I might be able to use it to troubleshoot at some point.

I'll look into it a little more then, I think I can source appropriate chips based on initial research.

Thanks again, really appreciate it.


JRH

[jrharley]

Sony-Tektronix analyzer update here -

OK, so I tried replacing the RAM chips with what I think are appropriate units. It was tedious to do, but I was able to replace them without deconstructing the board assembly. It's still too soon to draw any conclusions from my limited perspective, but I didn't get a very encouraging result. I've attached a shot of the screen on power up.

Yeah, looks like something is scrambling it up. I'm going to do the deed and take out the board to confirm that I've got the chips in there right, but any feedback on the scrambled display result in the meantime is appreciated.

Later,


JRH

[LazyJack]

Congrats for the instrument. Cute little device, very useful for HP style signature analysis. Good that you have the probes, pretty much useless without it.
Are you sure about the RAM? Might be address decoding, etc. showing up as RAM failure. I would look at the associated circuitry, not only the actual RAM chips.
The service manual has quite extensive signature tests, but of course you would need a signature analyzer to troubleshoot this signature analyzer :-(

[TurboTom]

I assume that you replaced the display RAM (U480 / U482). System RAM is of the same type, so I mentioned this just to be sure. Have you got an oscilloscope available and some experience probing digital circuitry? A completely messed up screen like yours is unlikely the result of defective screen RAM, much more probable - as @LazyJack mentioned - is a defective line driver or an address decoder so the RAM doesn't get initialized properly or just with the wrong data.

Please check the troubleshooting tree in the service manual (around pg. 166 in the PDF document available on the web).  If you have another signature analyzer, this may come in very handy for troubleshooting. Depending on the particular firmware version of your 308, the available signatures may be of little help, though.

If you're not successful, I'ld have a look at U462 and then work my way back through the strobe/addressing circuitry U454c, U488b, U416, U412. This may be an arduous task but fortunately, the computer inside the 308 is very simple and made up of all-standard parts. If you need further assistance, I may be able to take reference measurements on my 308.

[jrharley]

Thanks folks! Really appreciate your comments.

At the outset, I was just following the screen error LazyJack, you could certainly be right that something else is causing the issue other than the chips. I'm personally not one to be too suspicious of chip failure in troubleshooting, based on just visual observation and raw probabilities, so your point is well taken.

I replaced U420 and U422 TurboTom. After reading your thoughts I may have done something boneheaded, not sure at the moment. They are both listed in the parts list as 1024 x 4 Static Ram. I naturally went there, since the manual indicated that if I got the 03 or 04 error, then either U420 or U422 were bad. It can never be that simple though right? I've got my eye on the troubleshooting tree for sure, thanks for the tip!

My replacements were sourced from the Bay of course, and they have the designation SCM 2114, and are allegedly a Sprague part. You never know from the Bay though.

I have worked my way through to understanding how they packed this thing in the meantime and have the board assembly out of the chassis. There's a fair number of things to document and keep track of, but I think I'll get her back together in good 'nick. I'll probably become an expert in the dis-assembly by the time this is over!

Thanks so much for the comments and information! It's already yielding positive things. I'll post up with any new developments, and as always appreciate the feedback.

Later
JRH

[jrharley]

308 update here –

You just never know what’s around the next corner on these journeys. I’ve had an unexpected development emerge in the testing.

I decided to put the original RAM back in, arrange the boards out of the chassis to facilitate component swaps and have a better view of things for testing. What I expected to see upon power up, was the original boot up screen showing the error code “03, 04”. Instead, I got the same garbled screen as I did after replacing the RAM chips with new ones. See photo for more info.

I’m currently pondering this development, trying to resist the urge to briefly short two pins on the MPU board labeled “RESET”, and studying the troubleshooting tree. The tree reminds me to confirm power supply voltages, so, that’s where I’m headed next.

As always, all constructive comments appreciated!

Later

JRH

[Tony_G]

I know absolutely nothing about these machines so after checking the power supply, I would follow the troubleshooting trees.

That said, if you find yourself at an impasse it might be enlightening to put a logic analyzer on the address and data lines to see if you have any stuck bits or lines.

TonyG

[jrharley]

Thanks for weighing in Tony_G!

The power rails seem to be solid on first look, pretty much spot on. Just for grins I looked for any AC and it was what I think is negligible at 2mV. I'm going to proceed as if the rails are good unless I find another reason to dig deeper.

For now, I'm currently chin scratching on how I might rig a connector to the front panel, since some of the LED indicators are part of the troubleshooting sequence. The front panel is a direct 20pin plug to socket through the chassis to the boards, so, I'll need to devise an umbilical to connect 'er up. I think I can do it. Sure wish I had more experience with how these logic based guys work! This is going to be a puzzler for me I'm afraid. Maybe I'll learn some things.

One interesting thing of note is that during the probing of the power rails, I must have crossed something up briefly, because the display changed to a grid sort of pattern and locked on that, until I did a re-start. Then it went back to the hieroglyphics. Interesting stuff ...........onward!

Later,

JRH

[jrharley]

Sony - Tektronix 308 update here -

I think the test setup is ready for following the troubleshooting tree now and I'm working my way into it. It took climbing a little learning curve on connectors to get here, but I think we're good to go for now. The tree is pretty easy to follow but I'm starting to run up on stuff I don't understand.

One possible path is telling me to "fix the kernel". Oh boy, I'm toast I think. What the heck does THAT mean?

Anyway, a couple of attached photos show the setup.

As always, any insights and comments are appreciated.

Later,

JRH

[Tony_G]

Hey JR,

Not sure if this applies here but typically in the software world, the kernel is the core part of the OS.

Maybe they're suggesting either a firmware reflash or replacing any spinning media.

TonyG

[Tony_G]

I grabbed the manual and took a look at it - It appears that the reference to "kernel" refers to U400 - This is the MCU in the unit and is, as far as I can tell an Intel 8085.

You're now in the realm of signature checking (I think you actually looked into this for the 8656A but I can't remember if you ever got a solution) -

Anyway, they're referring to the operation of U400 and the components that feed it information.

Sorry I can't help more.

TonyG

[TurboTom]

I'ld approach the problem differently: Since there still seems to be some structure on the messed-up screen that resembles the original display patterns, I'ld initially try to press the operation mode selection buttons "Parallel Timing", "Parallel State", "Serial State" and "Signature". If the instrument reacts to these inputs and the screen changes accordingly, I'ld expect the CPU, keyboard I/F and probably the analog / input circuitry to be working okay.

In this case, I'ld look at the video ram and the circuitry mentioned in my post some places above. This may simplify fault finding somewhat.

[jrharley]

Thanks Tony_G, appreciate the thoughts!

I’ve since done a bit of research and that does appear to be the case with the “kernel”. Best I can tell so far, the operating system is not setting up the RAM chips to get things started on power up, as it’s supposed to do.

After pecking through the troubleshooting tree I think I’ve come to a, let’s just say “pause point”. It’s telling me to “fix the kernel” via setup 2.1. (See attached pic) I also attached a pic of the screen after removing jumper P410. The whole screen then filled with little question mark symbols, which I took to be the “blanks” described in the tree.

A number of questions to think about at this point, and I don’t have a functioning analyzer to test the over 30 pins that reveal the operation of the “kernel” per the prescribed procedure. I suppose this is why they refer to the error code “3,4” as a fatal error! Replacing the RAM in this guy must involve more than sticking new chips in. There’s no procedure for replacing the RAM outlined in the manual, only which pins to test on U400, so, I’m not sure how I’d “fix” it, even if I had another signature analyzer.

Maybe someone like Ian Johnston or Kerry Wong or several others that we all know and love, could get this cutie up and running. Not this Bodger.

Unless some logic Guru comes out of nowhere and says something like “Yeah, just hold down the #3  button while you power up”, I think I need to carefully re-assemble things and stage it for whatever comes next. I will research Tektronix support options like any forums and see what I can learn moving forward, but yeah, this one’s going to have to park for a while.

As always, I really appreciate everyone’s comments!

Later,

JRH
PS - I see others have weighed in so I'll absorb those comments and re-evaluate.

[jrharley]

Hey thanks TurboTom! Really appreciate you trackin' this.

All good thoughts. I'll give the button push a try. So far, keys have been unresponsive, but I haven't been specific in my approach.

With reference to "original display pattern", I feel like it's important to keep in mind that the original display pattern was a recognized error code - 03-04. So I'm thinking the "kernel" was operating properly at the point of initial power up. After messin' with the RAM chips, I haven't seen that original screen again.

I think I'll go push some buttons.....

Later

JRH

[TurboTom]

Signature analysis instructions start from page 192 of the manual PDF. FYI, I attached the page starting with signature analysis of the ROM system. Topic A specifies where the control lines are to be attached to and what polarity should be selected. B gives you informations on the configuration of the DUT. C is a cross-check to make sure the signature analyzer is configured correctly. In table D, finally, the test points are specified together with the signature that should be read there.

This continues over several pages so it's possible to signature test the whole digital circuitry of the Tek/Sony 308. There's little that's not covered in the manual, but it's a huge amount of information and sometimes it's difficult to find exactly what you're looking for.

[jrharley]

I'm freakin' out here Turbotom!

I just went in to the shop to power the unit on and confirm that the buttons you referenced didn't respond. Upon power up, the guy showed the original error screen again !! I guess you just never know what's next. I powered it up, expecting to see the scrambled screen I have been seeing, but the damn thing booted up with the original error screen again.

The only change since the last power up was re-installing the jumper, P410, back into circuit. Oddly, the jumper is composed of individual, across 2 pin blocks. One Dupont style connector block jumps across 2 pins. There are about 8 of these jumper blocks that make up P410. Connection issues seem to be complicating the analysis here. The "triggered" LED indicator on the keyboard also lit up, which didn't happen earlier.

This certainly changes the troubleshooting tree path. I'll post up when I learn more.

Sheesh!.........

JRH

[Tony_G]

Nice - Might be the ol' dirty contact issue.

Look forward to the next update.

TonyG

[jrharley]

I am so jazzed right now I could stay up all night!
Thanks for the comments guys!

First step tonight was to clean all the terminals on the P410 jumper assembly.

TurboTom, your original statement has proven to be spot on.    I realized you were right about that when I re-visited the tree, based on the recent developments. It led me straight to a box that just said - "replace the RAM device". And so I did once again. This time, I was a little more selective in replacing one of the original chips. U420 always looked a little "off" to me. It didn't have the same markings as it's sista' chip, U422, which matched several other chips in the circuit serving different functions. So I put a new RAM chip in there first.

Sweet, ever lovin' analyzer, if the thing didn't boot up with what appears to be the correct screen!
I've got to go through the start up checks of course, but this guy appears to be working! Tonight, I will dream of electric sheep........

I'll post up shortly with pics and info as I get my mind around this amazing and unexpected development.

Later,

JRH

[TurboTom]

Congratulations -- this is great news. Obviously, the "advanced" diagnostic capabilities introduced with signature analysis had their pitfalls: The neccessisty to be able to disconnect the data bus by a jumper block provides a new point of failure / contact problems which apparently caused the majority of the observed problems. Well done spotting that!

You may consider replacing the individual jumpers (which often only provide a single contact surface) with a female 1/10" 8x2 header with the neighboring contact pairs bent over and soldered short. From my experience, these provide much more reliable continuity.

[jrharley]

 Thanks TurboTom!

I know right? It's weird that they opted to go with such an assembly for that obviously important link. Only thing I can figure, is that if you really know what you're doing in this circuit, you could remove specific jumpers in the stack to aid in isolating a problem. Who knows? At any rate, I think the connector has been restored to at least as good as it was out of the box.

I'll be starting the re-assembly process soon, and I'll post up with a final sound off on this thread when I get 'er back together. Meanwhile, see the attached screen shots.

Later

JRH

[Tony_G]

Outstanding - The great thing about getting this working, apart from the feeling of getting it working, is that you now have a signature analyzer you can use on other gear 

Congrats on seeing it through.

TonyG

[jrharley]

You're right Tony_G,
Much satisfaction, much future utility.

Always appreciate your contributions!

One final post up with this thing hopefully. I feel like I should outline a couple key points to finish up, that should show up on a Google search if some other poor fool is working on one of these.

Later

JRH

[Tony_G]

NP - Always happy to offer whatever, even minimal, help I can.

You should do a post on how to use the 308 to read a signature - Just to flex those signature analysis muscles...

TonyG