• Please review our updated Terms and Rules here

Malfunctioning VT100 boards.

MattisLind

Veteran Member
Joined
Sep 30, 2013
Messages
1,158
Location
Stockholm, Sweden
I received two VT100 boards and a VT125 board.

One VT100 board was completely dead. Occasionally it would print some garbage characters on screen and never give the POST beep. This board was still in box and stickers on it that it had been in DEC facility in Netherlands to do repair in 1983.

The other VT100 board did POST self test but showed this output on screen:

sTL00FTl.jpg


0U3ZHEVl.jpg


The repeating SETUP text is weird. Some logic that doesn't work properly? And maybe the font PROM? The other pictures show the screen when entered a few characters in local mode. The vertical row to the right is not something I have typed in but generated "automatically" by the VT100 itself.

The VT125 showed "VT125 IC Error" on screen when powered up.

Before I start dig into this and attaching logic analyser etc I'd just wanted to ask if anybody seen these error symptoms before and could give a hint on where the problem could be.
 
Last edited:
DEC documentation is really amazing. The VT100 technical manual is excellent. I was reading it to understand what was going on in the VT100 and it really tell you all the details.

One of things I recognized when looking at the pictures above was that there were two extra characters in the display. How could that come? The VT100 is 80 characters not 82 characters, right?
Then of course there were jibberish on the screen which I thought had to do with bad character encoder or something similar.

But. There is something called a Terminate character that terminate each character line and start the blanking, the nice tech manual tell me. The special terminate character is stored in the display memory and is the 0x7f. But hey right curly braces is 0x7d - a bit missing? Bad RAM? The VT100 would not pass the POST with a failed RAM chip so it has to be something else.

The missing bit also explains why the tab stops in SET-UP screen says 1014545890 rather than expected 1234567890.
So why is the pattern repeating over the whole screen? Don't know. But if the terminate characters had been working correctly it would not be seen - notice the right curly brace in the first column.

Now I am on hunt for a failing bit in a couple of latches or the line buffer RAM.
 
Mattis,

I am sorry I have not had more time to help you. My day job is all-consuming at the moment. At least it is fun though.

The VT100 technical manual is indeed a masterpiece. Studied carefully you can repair ANY VT100 problem. I have probably fixed at least six VT100 basic video boards over the years. I got three broken ones from Will a few years ago and was so into it that I fixed all three right in a row (all different problems, but the logic was all fresh in my mind.)

At the moment I can only afford to give you moral support, but you are surely on the right track!

Lou
 
A day with rather dull weather and my mother-in-law roaming around in the kitchen (finally ready after the refurbishment) so I escaped in to my garage and found the old VT 100 Basic Video boards that wanted my attention. After some measuring it looked like one of the line memory chips were bad. Replacing E11 with a new 2111A-4 (slower than the original) fixed the problem.

So with the success fresh in mind I continued with the next board. Completely dead. Started to exchange ROM chips. It kept showing various error codes on the keyboard. After borrowing three known god ROM chips it now showed main RAM failure instead... Then I recognized that the bad ROM chips and Main RAM chips had one thing in common: They were all manufactured by Synertek, date codes 80-81. Anyone else has seen this problem with Synertek chips?

So now I need to find a way of creating an adapter so I can use standard 2764 EPROMS. Anyone done this already?

Then there were a few VT100 boards that worked just fine. One had a rather peculiar (DEC made) board connecting with a ribbon cable to both the extra character ROM socket and one of the program ROM sockets. Unsure what it really does.

Finally there were yet one more board with text problems on screen similar to the first board. When touching a pin with scope probe it changed output on the screen. But now I was out of 2111A chips. So just testing with the one used on the other board showed that the 2111A was indeed the culprit.
 
So now I need to find a way of creating an adapter so I can use standard 2764 EPROMS. Anyone done this already?

You may want to just make an adapter that will take a single 2764 to replace all four 2K ROMs. The masked ROMs use two extra address lines
on the OE and CE lines and decode each ROM uniquely on the chip.

Documenting and dumping the firmware, especially for the mod you mentioned would be a good thing. Someone forwarded me ROM dumps so
I'll get those over to bitsavers. The person who did that is also looking for an early buggy version of one of the roms (23-031E2) that was replaced
by 23-061E2 on most VT100s in the field. He is also looking for VT132 firmware.

--

found a 2764 adapter
https://github.com/phooky/VT100-Hax/tree/master/ROM-Carrier

this also has the rom images
 
Last edited:
Yes. One 2764 covering all off them. But isn't the problem with that adapter that the CE is active low but the VT100 Basic Video has CS active high? At least that is how I understand the schematic and the post at day 27 at VT100-hax. If I put the adapter on a AVO board it should work. But not on the Basic Video Board. A small modification to the adapter board with a single gate 7404 and a jumper to select active high or low CS would be nice. Maybe I just should modify the existing design a bit.

This is the strange mod board:

qdiVnFAl.jpg


There are two 2708, a 7400 and a 7402 on the board. The original ROM is the 23-061E2 and it seems that the logic is mapping the upper 1k to the EPROM while the lower is from the ROM. (If I traced the logic correctly). The long cable is to replace one of the program ROM sockets and the short to be inserted in the extra char gen socket. The small black wire is connected to W1 on the board. +12 and -12 is taken from the current loop adapter connection which is looped through on the board. -5V to the EPROMs are generated by a resistor and a zener diode on the board.

Back:

GprB9tTl.jpg


I dumped the 2708: Program ROM. Char Gen ROM.
 
Is there opionons on the speed grade for the 2114 chips? The design makes use of 200 ns parts. But is that really necessary? The two phase clock is around 2.6 MHz for the 8080. I am not very familiar how the 8080 accesses memory. Would it be possible to use the slower 300 ns part? Or does anyone have a pointer to not to expensive 2114 with 200 ns spec?

BTW. Is there a good method to figure out which of the six 2114 that is bad?
 
But isn't the problem with that adapter that the CE is active low but the VT100 Basic Video has CS active high? At least that is how I understand the schematic and the post at day 27 at VT100-hax. If I put the adapter on a AVO board it should work. .

I need to correct myself. According to the schematic there are two jumpers, W2 and W3, that can be used to select what kind of CS sense the ROMs have. On my board W3 is inserted and W2 removed when corresponds to having ROMs with active high CS. So I think that best way forward seems to be to burn four 2716 chips. The original, AM9218CPC is a 350 ns part. But are the 450ns parts I have fast enough? Otherwise I have to go for the adapter route.
 
Is there opionons on the speed grade for the 2114 chips? The design makes use of 200 ns parts. But is that really necessary? The two phase clock is around 2.6 MHz for the 8080. I am not very familiar how the 8080 accesses memory. Would it be possible to use the slower 300 ns part? Or does anyone have a pointer to not to expensive 2114 with 200 ns spec?

BTW. Is there a good method to figure out which of the six 2114 that is bad?

Some people swear to the piggyback method, but I've never tried that myself, and It must lead to some other issues. If it's only six chips, desoldering, installing sockets and swapping might be the best way to go.

Some failure modes in IC's leaves pins strapped high or lo permanently, I've had some success simply testing the leads' resistance to gnd, comparing with the other RAM chips that's in the same environment.
 
I ended up connecting my logic analyzer, using an IC clip-on to connect over each 2114 in turn. A reference board showed a steady stream of write accesses and then following reads, reading out 'A'. On the failing board, when the clip-on was on the last pair it just halted when reading data. So it just had to be one in this pair. At least I was down from 6 to 2 possible suspects. Took a chance on one of them and was lucky! The board started happily.

Related to my previous post on using 2716 on the Basic Video Board: This is not possible since the CS pins are connected to OE (and PGM). So the only solution would be either the 24 pin 8kx8 EPROM or an adapter. Planning to order a few boards.
 
An update to this.

Q90t6QT.jpg


I programmed the Atmel AT28C64 (which turned out difficult since my old programmer didn't really like it - but after seven tries it was verifying OK) with the basic.hex file converted from the vt100hax site binary file.

Then I changed the jumper setting from W3 to W2 and yes it worked just fine.
 
Yes. I ordered Phookys design from OSH park. Phooky describes what he did with on this page. He also mentions that there are very little space left in the PROM space. Since this 8kx8 only replaces the four old 8316 ROMs we have no new space really.

No I never tried the MysteryROM just downloaded the basic.bin and converted to hex.
 
A day with rather dull weather and my mother-in-law roaming around in the kitchen (finally ready after the refurbishment) so I escaped in to my garage and found the old VT 100 Basic Video boards that wanted my attention. After some measuring it looked like one of the line memory chips were bad. Replacing E11 with a new 2111A-4 (slower than the original) fixed the problem.

So with the success fresh in mind I continued with the next board. Completely dead. Started to exchange ROM chips. It kept showing various error codes on the keyboard. After borrowing three known god ROM chips it now showed main RAM failure instead... Then I recognized that the bad ROM chips and Main RAM chips had one thing in common: They were all manufactured by Synertek, date codes 80-81. Anyone else has seen this problem with Synertek chips?

So now I need to find a way of creating an adapter so I can use standard 2764 EPROMS. Anyone done this already?

Then there were a few VT100 boards that worked just fine. One had a rather peculiar (DEC made) board connecting with a ribbon cable to both the extra character ROM socket and one of the program ROM sockets. Unsure what it really does.

Finally there were yet one more board with text problems on screen similar to the first board. When touching a pin with scope probe it changed output on the screen. But now I was out of 2111A chips. So just testing with the one used on the other board showed that the 2111A was indeed the culprit.
I replaced bad Synertec 2114s on two VT100 boards today, both from that time and on both I could feel the legs almost falling off as I cut them. One board had previously had four chips replaced and so I replaced one of the remaining Synertec chips and got it working. The other board was all original and I had to replace four of the six Synertec chips before I got it right (could be only 1 was bad and I was just unlucky). I would have replaced the remaining Synertec chips, but I was out of sockets.
All of the boards I looked at with Hitachi RAM worked fine.
So, it's a fair bet the Synertec RAM from that era can be relied upon to fail.
 
I replaced bad Synertec 2114s on two VT100 boards today, both from that time and on both I could feel the legs almost falling off as I cut them. One board had previously had four chips replaced and so I replaced one of the remaining Synertec chips and got it working. The other board was all original and I had to replace four of the six Synertec chips before I got it right (could be only 1 was bad and I was just unlucky). I would have replaced the remaining Synertec chips, but I was out of sockets.
All of the boards I looked at with Hitachi RAM worked fine.
So, it's a fair bet the Synertec RAM from that era can be relied upon to fail.

I've had about the same luck with Synertec 2114s. I've replaced a bunch of them, probably half of the ones I've run across.

-Dave
 
Back
Top