• Please review our updated Terms and Rules here

PDP-11/05, won't deposit from front panel

R

RetroHacker_

Veteran Member
Joined
May 21, 2006
Messages
574
Location
Niskayuna, NY
I have a PDP-11/05 which won't deposit into memory from the front panel. I can examine, and I can run code already in memory, but I can't deposit. Toggling the deposit switch will display the correct value on the lights, and will increment the address, but nothing ever actually gets written to memory. If you go back and check, the memory location is unchanged.

Fortunately, I have a friend who recently acquired a working machine. So, the problem has been narrowed down to the Control board (M7261) - one half of the CPU. The memory and the front panel itself is fine. The computer works properly with the good control board from the other machine, and the other machine exhibits the same problem when the bad control board is installed.

I can install the good control board in this machine, toggle in a program, shut it off, switch the bad board back in, fire it back up, and run the program with the bad board. This has proven that the computer *can* write to memory with the bad board installed, from program control. It just can't do it from the front panel. And, from what I can tell, instructions all appear to work with this board.

Attempts to troubleshoot the bad control board haven't been terribly fruitful so far. Lacking card extenders, it's very difficult to probe or test signals. Thanks to some creative cabling and a logic probe, I've been able to verify that the basic circuitry driven by the deposit switch - a couple flip flops and an inverter (E97-99 on the board) all seems to work. Further than that though, I'm not really sure. The actual console switch register is on the data path board - not this one.

Anyone have any suggestions or pointers? I've got a headache from trying to comprehend the nomenclature on the DEC schematics - and from trying to read the many-generation photocopy that was scanned for bitsavers. I'm not really sure where some of these signals go, and I don't fully understand what's supposed to happen when that deposit switch gets flipped. What signal actually triggers copying the console switch register into memory?

And, of course, failing repair, does anyone happen to have an extra M7261?

-Ian
 
I sure would like to figure out a way to help in these situations. It would give me something to do, and maybe make a little cash in addition to helping out the community. Just collecting "Known bad boards" would be a start for anyone who wanted to act as a repair depot.

I happen to have a set of DEC extender cards I used when troubleshooting the 11/70. I also have the tools here to perform chip change outs of thru-hole parts. I know others on the forums too, like me have considerable troubleshooting skills and equipment, if not actual systems that could act as test beds for bad cards. [I used to have an 11/05]

It's been discussed offline, how to share resources, and one of the most inhibiting factors is not wanting to disappoint or become embroiled in any kind of "shoot'n match".


You're about 800 miles from me, so a quick "drive by" is out of the question. I think there are others who are much closer, but still not on your doorstep.

I wonder if we could setup a "friends network" of some sort, with defined rules that would prevent misunderstandings, and allow sharing of capabilities?

Guess I'm just rambling here. Half the fun is fixing it yourself.
 
RSX11M+ said:
I sure would like to figure out a way to help in these situations. It would give me something to do, and maybe make a little cash in addition to helping out the community. Just collecting "Known bad boards" would be a start for anyone who wanted to act as a repair depot.

I happen to have a set of DEC extender cards I used when troubleshooting the 11/70. I also have the tools here to perform chip change outs of thru-hole parts. I know others on the forums too, like me have considerable troubleshooting skills and equipment, if not actual systems that could act as test beds for bad cards. [I used to have an 11/05]

It's been discussed offline, how to share resources, and one of the most inhibiting factors is not wanting to disappoint or become embroiled in any kind of "shoot'n match".


You're about 800 miles from me, so a quick "drive by" is out of the question. I think there are others who are much closer, but still not on your doorstep.

I wonder if we could setup a "friends network" of some sort, with defined rules that would prevent misunderstandings, and allow sharing of capabilities?

Guess I'm just rambling here. Half the fun is fixing it yourself.
Click to expand...

I agree with your last comment. Much of the fun is fixing it oneself. But not knowing how was my problem too.
Or not being sure of how to troubleshoot. I got lots of help from Vince S. during my 8/i experience. But then
it was a much easier machine to work on. Each circuit cards was a small flip chip so not much logic on each.
Having a decent scope and knowing how to use it helped too.

I am not that close to you either. I don't have much pdp stuff left now either. Hopefully you can find someone
closer to you and go from there.
 
I have plenty of troubleshooting experience and the necessary soldering skills and equipment to replace the components. I've done tons of board level repairs on many machines. The only tool I'm lacking is card extenders, which I've been making up for by extending the signals I want to test, rather than the entire card. The problem is that I don't fully understand this schematic - the processor is quite complicated, and I'm just having problems figuring out, from the nomenclature on the schematic, what *else* to check - i.e. - what else gets cycled when the deposit switch is flipped. There's a lot of discrete logic on this card, and lots of custom ROMs that make up much of the logic. This is a rather complex machine, and a bus structure I'm not familiar with. It's not something simple like a TRS-80 or an Apple II - if it were, I'd have fixed it by now and I wouldn't have had to ask.

I, too, would much rather fix it than replace the card. But, it also has several parts that I don't know if I'd be able to replace - those ROMs that perform many of the logic functions - is there a replacement for those? I don't know offhand exactly what type of ROMs these are (bipolar, fuse link, etc), but I know the supply of blanks for this sort of chip is pretty low, and the speed requirements are such that I probably can't kludge in something else, either. Not to mention the fact that the boards must fit close together in the backplane, which rules out any kind of adapter board carrier.

-Ian
 
First - I apologize for my earlier comments. I didn't mean any disrespect of your skills or equipment.


My intent was to make the general case for a Member Maintained Service Network.


Please check your PMs, there may be something of interest.
 
RSX11M+ said:
First - I apologize for my earlier comments. I didn't mean any disrespect of your skills or equipment.


My intent was to make the general case for a Member Maintained Service Network.


Please check your PMs, there may be something of interest.
Click to expand...

I meant no offense to your skills. One doesn't always know what skill a person has who posts here.
I hope you find someone with what you need to get this solved.
 
Douglas Electronics *still* sells dec card extenders. I bought two dual height extenders from them about two years ago. I say, invest in some card extenders and roll up your sleeves!

The roms are likely fuse link roms. If indeed you had a failed rom, someone else out there could read theirs and burn you one. Or, you could make a plug-in eprom replacement (I have done that.) Sometimes the rom listings are in the print sets.

For what it's worth, I would love to accumulate known broken Omnibus (pdp-8a/e/f/m) cards. I do component level repair and will gladly provide a loving home to any dead Omnibus boards.

How do I get membership in the MMSN (member maintained service network)? All this time I have been calling myself the resident (at my home address anyway) customer engineer....

Lou
 
Well, we took a good look at the Control and Data Paths schematics last night. I think we've got it narrowed down to a 74154 that decodes what sort of bus transaction the processor is going to do. It appears that a single stuck output would produce the problem we're seeing; that is, the console switches are unusable as an input to the ALU. RetroHacker_ and I are going to actually test the hardware tonight and will report back.
 
You guys are good, if the group online can't get that M7261 working again don’t think anyone can. Would be interested to see what the problem was once it's resolved. Did you get an extender card or are you still installing wire test points? Also wonder if you're using a logic analyzer to look at these events referenced to clock or write cycle or just using a logic probe or scope, I have an old HP logic analyzer and once I learned how to use it that’s about the only thing I work with now on digital junk.
 
I second that comment on logic analyzers! If a problem at all looks slightly more complicated than I can chase down with the two channel scope, then out comes the logic analyzer!

Lou
 
Folks,

I just wanted to mention that I have confirmed that Douglas still has card extenders and prototyping boards. Here is a photo of what just came that I ordered from them: http://www.vintage-computer.com/vcforum/album.php?albumid=2&attachmentid=9297.

I bought a quad height extender and an omnibus prototyping card. It's really nice to know that this stuff is still obtainable. And, it's cheaper than what it goes for on ebay.

Lou
 
You must log in or register to reply here.
Back
Top