Alive but not working

[Architect]

I finished the kit and it seems to work. However, the CLEAR button doesn't clear the data LEDs. The INPUT LED lights up though. Some other key combinations result in most of the LEDs to go out, so it is not the K-register. Any data key pressed leads to all data LEDs to light up.

I understand that the keys pressed are mapped onto 377 and the LEDs are mapped on 200. So it could be the memory. But that is just a first impression. I will look into it in more detail tonight.

[Geoff Harrison]

I finished the kit and it seems to work. However, the CLEAR button doesn't clear the data LEDs. The INPUT LED lights up though. Some other key combinations result in most of the LEDs to go out, so it is not the K-register. Any data key pressed leads to all data LEDs to light up.


That's similar to the problem had at first, especially having all the data lights light up for any data key pressed.

It sounds like the K register is shifting when it shouldn't. Try checking the ICs and signals driving pin 6 on IC 104 & 105 (i.e. ICs 107, 116, 120, 121). Make sure pin 6 is not stuck low. In my case I had a bad 7420 @120.

[Grant Stockly]

Do you have a way to test the TTL devices?

I hope you find the problem easily!

[Architect]

Also in my case the culprit was a 7420. It was IC52 (page 24). I found it by using signal IN (pin 5 of IC117, page 04) to trigger my oscilloscope (also vintage) to see what happens the moment 377 is read/written while I push CLEAR or one of the data buttons. It is not that IC52 was functionally faulty, but it negatively influenced IC40 and IC41 to push up the signals WD and WT to 1.6 V which makes them 1 in stead of 0, causing a delay-line full of 1s.

Of course I cannot run programs now because I used IC75 in its place. I'll solve that tomorrow.

[Grant Stockly]

Wow...I must have a marginal batch of 7420s. Geoff had a problem with one of his 7420s too. Geoff, was yours functionally good but having out of spec logic outputs?

You can use any 20, like a HC20, LS20, L20, etc. You might be able to find one in another piece of hardware.

If you can't locate another one immediately you might be able to use a pull down resistor to bring the low voltage into spec?

Grant

[Geoff Harrison]

Geoff, was yours functionally good but having out of spec logic outputs?

I haven't tested it yet, I had to return my IC tester. I'll check the chip when I get it back. Or perhaps I'll jut put it back in the Kenbak and scope it.

[Reactive]

If you can't locate another one immediately you might be able to use a pull down resistor to bring the low voltage into spec?

Grant

I'd use a RAMWall, personally.


Henry

[Architect]

The days you could run to the local store and pick up a 7420 are definitely over. How sad. No Kenbak programming this weekend. I ordered several 74**s online and they will probably arrive on Wednesday.

The LS versions are 0.52 a piece and so I simply had to order another $ 20 of stuff I do not really need to prevent the $8 shipping. But I was good: I didn't end up with 50,000 LEDs.

[Grant Stockly]

But I was good: I didn't end up with 50,000 LEDs.

Too bad. Better luck next time!

:)

[Architect]

To make a long story short: see the attached picture of IC25. The problem had nothing to do with a faulty 7420 as I mistakenly claimed before. It really is this simple. Only some people (Geoff) find this faster than others.

[Geoff Harrison]

Hmmm, that looks familiar :)

Congratulations, I'm glad you got it running. What are we up to now, three running Kenbak-1 Series-2 machines? Four, including Grant. That makes us a more exclusive group than owners of original Kenbak's :D

[Grant Stockly]

To make a long story short: see the attached picture of IC25. The problem had nothing to do with a faulty 7420 as I mistakenly claimed before. It really is this simple. Only some people (Geoff) find this faster than others.

I bet you felt like yelling, laughing, and crying all at the same time when you found that. :) Were you studying the trace layout of the board to debug the circuit and see the lack of a solder joint by accident?

So, have you run any of the examples from the manual yet?

Grant

[Architect]

I suspected the IC52 and I replaced it with the one in IC77. Then the problem of all the LEDs being lit constantly was gone, and I wrongly concluded I found the answer. But it just accidentally worked because I took out IC77. When I put in a new 7420 at IC77 it was back.

Since I swapped IC25 and IC41 before (with the 7400 from IC99 remember), the only thing I could think of is that IC25 was upside down which can easily be detected by checking the voltage at 7 and 14. Only then I saw 1.5 V at pin 14 of IC25 and started disassembling the machine.

Warned by Geoff, I carefully checked the board after soldering and back then I found two similar incidents. But apparently overlooking a third.

[Architect]

I suspected the IC52 and I replaced it with the one in IC77. Then the problem of all the LEDs being lit constantly was gone, and I wrongly concluded I found the answer. But it just accidentally worked because I took out IC77. When I put in a new 7420 at IC77 it was back.

Since I swapped IC25 and IC41 before (with the 7400 from IC99 remember), the only thing I could think of is that IC25 was upside down which can easily be detected by checking the voltage at 7 and 14. Only then I saw 1.5 V at pin 14 of IC25 and started disassembling the machine.

Warned by Geoff, I carefully checked the board after soldering and back then I found two similar incidents. But apparently overlooking a third.

[Architect]

So exactly how do I go about using AS? I want to write a program to calculate E (Euler's constant). The series converges rapidly and you only need to maintain the sum and the term.

[Grant Stockly]

I'm a little confused, is it working yet or are you still having problems? :confused:

What operating system are you using? Linux/Windows/OSX?

The version of AS on the internet does not support the Kenbak. I downloaded his source and added the Kenbak myself. If you want to compile the C source code yourself I can e-mail you that. Otherwise I can compile a binary for you.

[Architect]

Everything works properly and I tested some loops to initialize the term and the series. Tonight I wrote the part that will do divisions on arbitrary numbers, but I didn't put it in yet. It is 62 bytes. I need to do the one that does the adding, but that is pretty trivial I guess. So I must be able to calculate E with a precision of 64 bytes which is 154 digits.

And I prefer Linux since I know that best. So if you can send me the sources, I would appreciate it.

[Grant Stockly]

I e-mailed you the source code and a fixed include file. Please let me know if you didn't get it. Its a tar.gz file so I doubt a virus scanner would reject it. ;)

Also, let me know how it compiles for you. Its going to be exciting to create a library of KENBAK-1 programs!

A project on the back burner is a RS-232 device which could automate button pushing. For example, it could load an intel hex file into the kenbak, or read the kenbak memory image out.

Since the "PC" and registers are regular RAM locations the state of the KENBAK could be preserved without power to the KENBAK.

The project I'm planning on would use a single ATMEGA8-16PC, the RS-232 level converter, and a few capacitors. It will be open source. All of those parts should be easy to get where you live.

Unless you want the pretty PCB then you could just build it yourself. :)