new build with problems
A single board computer using an electromechanical CPU
Brought to you by:
jhallen
Menu
▾
▴
Hello,
I just completed the build of the Relay Computer. The power supply I ordered can be set so that the center pin is either + or -. Which is correct. Since the computer does not look right when powered up, I may have gotten the polarity wrong. Also I unstalled U3 and U4 backwards, so I may have zapped some or all of the other U ic's.
I was meticulous with mounting of the parts and the soldering. So far I have checked all the solder joints and reviewed that parts are installed with the correct orientation. My handling of the IC's may be at fault as I did not have a wrist band static eleminator when I handled the chips. I plan to order all new U ic's. Is it possible to get replacements for the two ic's that came with the board?
Right now the keyboard and digit displays do not work.
I would really appreciate some advice on all of this>
J Durgan
First-off, congratulations on building a Relay Computer.
The 12v power supply center pin should be + (positive).
If you have always had the power supply in that configuration, you should be OK. In looking at the specs of the U29 voltage regulator (12v -> 5v), it would probably tolerate a reversal in voltage and not pass on a bad voltage to the rest of the IC's. So either way you should be reasonably safe from reverse voltage damage. A reverse voltage to the 12v relays shouldn't cause issues.
Most of the modern IC's that are being used in this project are not overly sensitive to static electricity, so unless you were working in a very static electricity environment, the the IC's should work. Static Electricity usually damages these kinds IC's so that they have shorter life spans and not usually out-right deaths. So I don't think your handling damaged anything in the short term at least.
Hopefully you have a meter or something to verify that the +5v is working correctly. (see section 4 - Bring up section in assembly instructions).
Having U3 and U4 installed backwards could have damaged them, but probably not any other nearby IC.
You can get inexpensive IC testers (~$20 US) from places like Ebay to verify that your IC's are still working.
Depending on what test equipment you have, you will probably need to trace signals on the various IC's to see if there are any voltages present that aren't either near 0v (low) or 5v (high).
IC Voltage levels of 2 or 2.5 volts are usually an indicator that something is wrong and should be looked into.
Hopefully these insights help, let us know how it goes...
Hello,
Thanks for the complete and prompt reply.
I discovered that I did not solder all the pins on U23. Once I did that, the digital displays and keyboard all work. I then did the FF (step) to turn on all the leds, followed by the 00 (step) to turn them all off. The results are the two pictures below pic 1 is FF pic 2 is 00.
I did check the u204 and 205 resister arrays, and they appear to be correctly installed and no unsoldered pins on the sockets.
In my previous email I indicated that I had reversed U3 and U4 in the ic socket. Would this have made a difference. What should I look for to find the problem with the led’s.
Thanks
John D
John,
Glad to hear that you found the un-soldered pins on U23 and that you are now at least getting some activity on the displays and keyboard.
As best that I can see in your images, it appears that the PC (Program Counter) bit 0 LED does not come on when you do FF <step>. Please verify that is what you see as well. To help see the PC LEDs turn-on and off, try entering 00 <step>, and then keep slowly pressing <step>, you should see the PC LEDs count-up in binary for each <step>.</step></step></step></step>
If PC bit 0 is indeed not turning on, look at Relays K1, K2 and K3 to verify that they are installed correctly and there aren't any missing soldered pins. If you don't see anything wrong, try replacing each of the K1, K2, K3 relays (one-at-time), be sure to mark each Relay so you can keep track of which one is which. If you don't have extra relays you can trying swapping the K4, K5 and K6 relays with K1, K2, K3. After swapping/replacing the relays try doing 00 <step> and then slowly stepping the PC to see if bit 0 now works and in the case of swapping relays check to see if the other PC bits stop working. It is possible that just removing and re-inserting the Relays, bit 0 starts working.</step>
Depending on what you see/find, the problem is due to either a bad relay or a bad socket connection. I'm assuming you used sockets for the relays...
Let us know how it goes...
Hello,
The 0 PC LED is definitely not lighting. I did as you suggested and swapped out K1 K2 and K3 one at a time. the 0 LED still did not light. at the end of this test, relays 1,2&3 were installed in K4,5,& 6 and 4 , 5 , &6 were in Position K1,2,&3. All PC Led’s lit except 0. This must mean that the relays are good. I desoldered all the joints on socket K1,2,3 and LED in position 0. Still does not light. Can you suggest anything else I might try?
When I input 10 and run. the led display shows 21.90 the program stops after bout 5 clock cycles. Is this right?
The trim pot for clock speed dose not make any change in the clock speed. What could be wrong here.
Thanks so much for your support
J Durgan
If you try FF <step> do you see the PC bit 0 LED blink briefly? If so, it tells you that the 1 is making it to the bit, but the relay is not latching (hold resistor problem). If it does not even blink, then the 1 is not making it there. It could be the capacitor, or the path back to the driver chip.</step>
The speed control knob should work when the speed is set to 0. Do you have it set like this?
Hi,
I have the same problem with the trim poti and after extensive testing we found that the board has faults. There are also other faults in the board, so at least my version is useless.
Thanks
Marcus
Hi,
I have similar problems and it turns out that the board is faulty. Some connections just do not work. So mine can't be used.
Thamks
Marcus
Marcus,
I've built two of these relay computers back in 2018/2019, one (2017 version) from Joe Allen and one variant from another enthusiast who made some changes like adding a FTDI USB interface. The only problems I had was from my own overlooking of some solder joints.
The manual clock adjustment using the trim pot works fine on both of mine.
So what generation board are you working with ?
Francis,
I have to look it up but it is the newest version (I received it in early April) . It sits with my electronics (professional) guy after I have given up and after two months and resoldering the board twice he has given up too. Poti is one issue (so the poti actually works and delivers the required voltages) but the bigger issue is relay #5: #5 shows a somewhat erratic behaviour. Sometimes on, sometimes off. Changing the relay didn't help, neither did checking and resoldering. Resistors are all ok too.
Thanks
Marcus
Hello,
Since we last communicated, I have gone over the solder joints with a high powered glass. All seems ok. I have also replaced all the chips except for the pre-programmed pic chips. I did use a wrist anti static wrist strap for this. The machine still stops on step 5 showing 21.90 on the digital display when I run the test program. I believe it was Joe who suggested that I buy a TTL cable using terminal console software. I’m not terribly familiar with doing this, but I know that a baud rate and com number are required. Can you advise as to what these should be set at. Also what should I be looking for on the terminal screen?
Thanks again
John Durgan
John, I've attached a copy of one of Joe's documents that describes the serial connection and abilities (and more) . See page 36 "Serial Console" in the attached PDF.
Besides getting the USB -> TTL cable you will need to install a Serial Terminal Emulator. Here is a list of a number of ones available for Windows based machines, there are others like MiniCom for Linux:
PuTTy <-- I've used this one for a lot of different applications.
TeraTerm <--- I've used this one for a lot of different applications, it works well and you can create scripts to automate commands/transfers.
HyperTerm <-- From Microsoft, it is a bit hard to find, but does have some useful features.
Look through the document, it contains a lot of useful information.
John,
OK, it sounds like K1, K2, K3 aren't the problem.
Let's look at the driving signals to see if we can see changes.
The PC Bit 0 LED is connected through R57 from R202 Pin 8 through R202 Pin 9 through U21 Pin 7.
Using a voltmeter or DMM (which I hope you have) see if you can see the voltage levels change when you change PC Bit 0 using <step>. Starting first at U21 Pin 7.</step>
Note: You can check R58, R202 Pin 7, R202 Pin 10 and U21 Pin 6 to watch how PC Bit 1 changes.
You should also check to see if R204 Pins 1, 2 have 12v.
Also check the K13 and K26 Relays.
Check U16 Pins 6,7 (5v) and U20 Pins 11,14 (12v) to see how they change when doing <step> changes.</step>
Somewhere there is probably a break in the signal path.
The 10 program should run for quite a few clocks, it is probably stopping early due to the PC not being able to allow the program to sequence through correctly.
It appears that Joe Allen has joined the discussion, he is the designer of this great relay computer, be sure to read his messages.
Hello Again,
The problem with the 0 LED turned out to be that one lead of the resister leading to Led 0 was run through the wrong hole in the board, there being two holes very very close together. So now the 0 LED works. Also the clock speed issue was that I did not set the clock speed to zero, then use the trim pot. The only remaining issue is that the program 10 only runs thru 5 clock cycles, then stops. 21.90 appears in the digital display. Any Suggestions?
John Durgan
Nice progress!
I would check again the PC FF and 00 test, make sure the PC LEDs turn on and off. This the most likely problem.
Next, do you have the serial cable? Try running the program with trace on ("t on" command), we can then compare it with a working system to quickly see what's going on.
Hi,
Where do I get “the serial cable”
JD
See my response to the TTL Serial cables post.
Hello,
I finally got the correct til serial cable, and got connected to putty. When I type the t command I get a “huh?” response. Is that correct?. Attached is a listing of what I think is the built in demo program, When I run it on the machine it completes only about the first five steps then stops. Please advise
If this is not what you need please let me know what I need to do>
Thanks very much
J Durgan
John,
The 't' command takes an argument, either "on" or "off". If you don't include the argument the system doesn't understand the incomplete command.
I've attached a log file showing what you should see when doing "t on" and then "g 10".
Compare the log file to what you are seeing, hopefully something will stand out...
Hello,
I am finally getting back to this project. Attached is a log file of what happens when I run the test program using g 10. only the first several steps work then it quits. the last thing that prints on the screen is some PC code that does not appear on your printout. Any suggestions as to what might be wrong would be greatly appreciated.
Sincerely
John d
John, I don't find your log file attachment. I only see my original log file that I attached.
Hello,
I must have forgotten to attach it. Here it is
J Durgan
In looking at your trace log, the program is exiting (program counter set to 1ah) immediately after storing the initial start values and doing the first comparison. The comparison is checking to see if the value in memory location 01h is equal to 0. The initial value in 01h is supposed to be E9h, but the trace is showing 21h.
Do a dump 'd' after your program stops to see what is actually stored in 01h.
The program exiting early seems to say that the value in 01h is 0h, instead of E9h or even 21h
So we may have one or two problems, the value storing mechanism isn't working and/or the jump if equal comparison is jumping with a value other than 0.
Please attach a new trace log with the additional dump output so we can see what the 01h value is.
Remember the stored data value is only the last two digits of the memory location. If you want you could manually store 0000_0000 into the 00h and 01h memory locations before running the program, this will help you visualize those data locations more easily.
Here is the new trace:
Thanks for the help
J Durgan
It is interesting that the first 'st' command is able to store a 90h into memory 00h, but the 2nd 'st' command doesn't store E9h but 21h into memory 01h.
Try manually storing values into memory. You can do this from the unit's keypad:
00 ADR
00000090 INC
000000E9 INC
Now do a 'd' (dump) and see what the values are in memory 00h and 01h, they should be 00000090h and 000000E9h. Let us know what you see.
Try single stepping through the program and watching the LEDs on the unit to see if you can see anything that doesn't appear to be right. You should see 90h and E9h trying to be stored.
On the terminal console you can do 'j 10' to move the PC to 10h and then do 's' to single step each command. Or do the following from the unit's keypad:
00 STP
STP
STP
STP
...
Hopefully you/we can spot something wrong. At this point I'm not sure what is causing the memory storage problem and why the 'jeq' statement determines that the main program loop should terminate on the first pass of the program.
(I'm bumping this response, since Sourceforge didn't appear to send out a notification about the posting...)
Last edit: Francis Bauer 2021-09-04