PDA

View Full Version : Reproduction SIM4-01 set?



candrews
November 5th, 2019, 12:05 AM
Has anyone heard of anyone that knows anyone who is offering reproduction SIM4 01 set of boards? I4004
SIM4-01, MP7-02 and MCB-410
I know a few people here have originals but I probably won’t live long enough to find one for sale so I am looking at the reproduction route.
I run a reproduction MIL MOD8 and, by golly, I don’t see how a real one could be any more fun than this, so I am fine with reproduction boards.

Thanks

Craig

daver2
November 5th, 2019, 12:58 AM
Didn't even know that project existed. Thanks for the pointer though.

If I had known a while ago - I would have yet another project on my rather large pile!

I still might obtain a copy of the PCBs and look to hunt out the 'unobtanium' parts over time though.

Incidentally, I developed my own JavaScript emulator for this beast (see over at Forum->Technical Support->Vintage Computer Programming->8008 pure JavaScript Emulator). I have shaken most of the bugs out - and I find MONITOR-8 a really nice monitor - but development work has stopped at the moment due to other projects and lack of posts.

Interested to know what software you run on it to keep yourself amused!

Dave

Dwight Elvey
November 5th, 2019, 07:01 AM
I'm one with the setup.
I'm still looking to get the assembler ROM code for the SIM4-01. I have all but one ROM that look to have good code. The fourth one is dead. I need a A0743 EPROM!
I have the code to program 1702A but I'm looking for the code to do non-A. It wouldn't be hard to write as the only difference is the duty cycle and the port used to trigger the program pulse on the MP7 board. The non-A EPROM is A0542, to go with A0540 and A0541 that I have.
Since the SIM4-01 has so little RAM, I wrote code to do stand alone copy of 1702As. It does it in two phases. I use the LEDs on the MCB-410 to sequence when to put the master and when to put the blank. If one has the SIM4-02 it has enough RAM and one can use the code in the manual.
The schematics are in the manual. If you want I can take photo's of the top and bottom of boards. It shouldn't be too hard to make replicas.
I've not seen anyone else that has the complete set of boards, that post to the MB.
My MP7 is a -03 but as I recall it has a minor difference in the regulator circuit, compared to the -02. The MP7-01 only has circuits to do non-A 1702s.
I've an assembler and simulator for the 4004 but they run in Forth. It uses win32forth but could be adopted for other Forths by modifications to the file I/O.
I've made a board to run the Maneuver Board program but not finished stuffing the board. To many family projects! It uses too much RAM and ROM to run on the SIM4-01 but I believe it could run on a SIM4-02 with a little tacked on hardware for display and key switches. There is enough ROM space in 4K to write a serial I/O, instead.
Dwight

glitch
November 5th, 2019, 07:23 AM
Huh, yeah, I also didn't know about the MIL MOD8 reproduction! Another project to add to the list :)

Dwight Elvey
November 5th, 2019, 07:53 AM
I've seen National 4004s on ebay for around $150. There have been some 4002-2s for under $20 each lately as well. 4002-1 are easy to come by. If anyone wants some 1702As I recently got a box of National ones cheap. I'll send a set of the three A054x coded to anyone that wants them for about $7.50 ( price of a mailer, need to check current price ) ( only for 4004/4040 projects though, not for chip collectors ). The A054x proms run the MP7.
I have stuff on github at: https://github.com/prog4004/I4004-projects/
I don't have everything there but I'll update it with other stuff later. In the Maneuver directory, you can see how my simulator is used.
Dwight

Dwight Elvey
November 5th, 2019, 08:14 AM
Oops! I should say, for the cost of shipping. I forget, Dave is in the UK. It might be cheaper in a smaller box.
Dwight

candrews
November 5th, 2019, 08:33 AM
Regarding the straight 1702 programming. I believe the mil mod mon8 has a lot in common with the sim8 and might further presume that the SIM8 had some portions of the programming lifted from the SIM4. So this may not apply....

That being said, the mil mod monitor 8 doesn’t use a different hardware programming pulse generator for the 1702 to 1702A, it uses a software loop to determine the programming duty cycle.

Regards

Craig

candrews
November 5th, 2019, 08:40 AM
Dave

I had many components for the mil mod8 hanging around but had to order most. They were not hard to find and I have extras for many of them in case someone else gets in a bind and can’t something. The board set comes with boards for both the 8008 and 8080. I suggest getting an extra ROM board so you don’t have to swap EPROMs when swapping processors.

For those that want the boards or looking for the mother lode of MIL MOD information, here is the link to his website. https://va3ngc.weebly.com/repo-boards.html

For anyone wanting to follow my experience with them here is my YouTube
https://www.youtube.com/c/helloworld0000

As for what software I run, I have a pretty low amusement threshold. So a lot with tweaking the resident monitor, writing confidence tests, maybe a few games, but then I go right into interfacing to to things that couldn’t have been even conceived of back then.

Regards

Craig

candrews
November 5th, 2019, 08:59 AM
Dwight

I need to dig deeper, but I was thinking I would go for the last revision of the boards. That may be misguided but I lived through the constraints of writing programs with 256 bytes of ram and 2 kbit of rom and, frankly, I prefer having more of each. So I was thinking the additional capabilities and flexibility of later revisions would be my preference. Sure, if this was just going to hang on my wall, the first version and earliest date codes would be important.

What would be your thoughts for the most enjoyable system?

Craig

Dwight Elvey
November 5th, 2019, 02:42 PM
Dwight

I need to dig deeper, but I was thinking I would go for the last revision of the boards. That may be misguided but I lived through the constraints of writing programs with 256 bytes of ram and 2 kbit of rom and, frankly, I prefer having more of each. So I was thinking the additional capabilities and flexibility of later revisions would be my preference. Sure, if this was just going to hang on my wall, the first version and earliest date codes would be important.

What would be your thoughts for the most enjoyable system?

Craig


The 4040 with a 4289 is more what your looking for. Both are Harvard architecture but the 4040, used with the 4289 can be used with some added instructions that allow the 4040 to write to program memory where there might be RAM as well as ROM. The 4008/4009 combination doesn't allow one to do that. The 4040 has the ability to expand both RAM and ROM beyond what the 4004 can handle.
You do know that the RAM ( 4002-1 and 4002-2 ) for the 4004/4040 has special access that was intended for things like floating point exponent and such. It is difficult to emulated these RAMs with standard RAMs.
One of the disadvantages of using RAM in the program space is that one can only read a value from code on the same page. That means that to read some location A to location B, you'll need to use up some RAM space on each page to read the data. Of course if you only intended to use that RAM space for experimental program space, you can always execute anywhere in program space. It is also possible to setup that RAM space for data only but that is better to just be in normal RAM space without the exponent status nibbles. Both can be done with the 4289 by clever decoding.
The MOD4/40 has much of these features built in, that the SIM4-01 didn't. Still, you have to think in terms of program space and data space as separate, like a 8051 or 8048. Not that these couldn't have such mixing for development purposes with some clever decoding.

The Maneuver Board code, that I've been working on ( not my original code ), runs a display, keyboard input as well as calculates range, distance, time of closest approach, absolute speed and direction for up to 10 targets. It uses a cordic algorithm to calculate sine and cosine, given angle information. It does all that in less than 4K of code ( not really all that well optimized ). 4K is quite a bit but the 4004/4040 is not really all that well an optimized processor. One has to do a lot of LIM xx dd, SRC xx for just about any operation not in the internal registers.
Dwight

Dwight Elvey
November 5th, 2019, 03:04 PM
The SIM8 uses the same MP7 board as the SIM4. It would have the same issue. There is no feedback from the MP7 board to say when it is done. The delay loop for the SIM4 would be the wait for the board to finish so the code can change the address and data. It is an open ended control so it only needs to deliver a constant signal while the MP7 sets the right pulse width and dwell between pulses. The 1702 needs more dwell time between shorter pulses and more pulses than the 1702A. It would either have to use a different port pin like the SIM4 or have a switch to select the different dwell signal if using the MP7. I don't know how the MOD8 or MOD4/40 deal with programming as I've not studies either of them that closely. They may both completely control the dwell between pulse and the pulse rate. The SIM4 does not.
Dwight

candrews
November 5th, 2019, 05:21 PM
The SIM8 uses the same MP7 board as the SIM4. It would have the same issue. There is no feedback from the MP7 board to say when it is done. The delay loop for the SIM4 would be the wait for the board to finish so the code can change the address and data. It is an open ended control so it only needs to deliver a constant signal while the MP7 sets the right pulse width and dwell between pulses. The 1702 needs more dwell time between shorter pulses and more pulses than the 1702A. It would either have to use a different port pin like the SIM4 or have a switch to select the different dwell signal if using the MP7. I don't know how the MOD8 or MOD4/40 deal with programming as I've not studies either of them that closely. They may both completely control the dwell between pulse and the pulse rate. The SIM4 does not.
Dwight
I haven’t found the SIM programmer code online to look through it. My suggestion was that the MIL MOD8 uses the same dual monostable hardware design as the MP7. However, the MIL does not use both pulse generators so it would not surprise me if the SIM didn’t either. The MIL only triggers the shorter one and uses a software loop to determine the duty cycle. They also speed up the overall programming by doing a program-check, program-check loop until the eprom matches and then it does the overprogram. there are many common aspects between the programming hardware between the MIL and the SIM systems, soo.... I would not be at all surprised if the monitor for the sim8 and SIM4 also ignores the dual multivibrator hardware setup, does duty cycle in software, and uses the program-check + overprogram algorithm since that is fastest.

Slob
November 5th, 2019, 05:30 PM
i had no idea that the MIL mod-8 board set had been reproduced. The board set is probably not cheap, but where else are you going to go for one?

As I recall, those boards used first-gen TTL and some of those chips might be getting a bit scarce/pricey particularly in non-LS versions (the -L's might be tricky, for example) if you wanted something period-correct.
And then, there is the memory-but I've got some gold lid white ceramic 2102's looking for work...this would be very tempting. I've even got a MIL8008 for such a project!

I really like the MIL monitor. I use it in an emulator (Dunfield's transcription) and it is amazing what they packed into it. I have read somewhere that there is a bug in that transcription; I've never followed up on it. But my ESP8266 emulator hangs up if I dump a certain location and I've run a lot of other real 8008 software including SCELBAL in my emulator OK. Believe me, I've made more than my share of mistakes "digitizing" worn-out ASR-33 teletype output!

candrews
November 5th, 2019, 05:53 PM
Slob..

I wasn’t going for period perfect so i don’t have any idea how much extra that would take in time and cost. There are a dozen boards in the set, I forget how much they cost. I think Charles has some boards available still.

You are right about the mil monitor, it is very powerful and the symbolic input and symbolic output in the MON80 can’t be beat.

Craig

Dwight Elvey
November 5th, 2019, 07:57 PM
I haven’t found the SIM programmer code online to look through it. My suggestion was that the MIL MOD8 uses the same dual monostable hardware design as the MP7. However, the MIL does not use both pulse generators so it would not surprise me if the SIM didn’t either. The MIL only triggers the shorter one and uses a software loop to determine the duty cycle. They also speed up the overall programming by doing a program-check, program-check loop until the eprom matches and then it does the overprogram. there are many common aspects between the programming hardware between the MIL and the SIM systems, soo.... I would not be at all surprised if the monitor for the sim8 and SIM4 also ignores the dual multivibrator hardware setup, does duty cycle in software, and uses the program-check + overprogram algorithm since that is fastest.

OK, you made me look. For the 1702A, the SIM4 only sends one 0.5 second pulse to the MP7 per address. The MP7 generates approximately 100 ea 3 ms pulses with a dwell time of about .25 ms. It does this only once per address totaling 2 minutes. There is no significant delay between addresses.
For the 1702, the SIM4 only does a 0.05 second pulse for a about 15 ea 3 ms pulses. It does this for each address and then repeats the address/data sequence for 10 passes. The MP7 actually creates the dwell time between the 3 ms pulses but the SIM4 creates no specific delay for each address. Each address gets 10 shorter passes, only for the 1702, non-A.
Dwight

candrews
November 5th, 2019, 08:09 PM
Dwight

Good information to know, thanks

This is pretty much what the MIL documentation SAYS its resident monitor does, but it does not. I think MIL lifted some design aspects from the MP7 and then maybe did one better with the software to speed up the programming. I have had 1702As that only required one 3ms hit in the MIL to get a matching data byte. It can do a 1702A lickety-split.

Craig

candrews
November 5th, 2019, 08:35 PM
Dwight

were the SIM4-01 and SIM4-02 contemporaries that were differentiated by price and capability or are they revisions that came out sequentially and the intent of the -02 was to replace the -01?

Thanks

Craig

Dwight Elvey
November 5th, 2019, 09:17 PM
Dwight

were the SIM4-01 and SIM4-02 contemporaries that were differentiated by price and capability or are they revisions that came out sequentially and the intent of the -02 was to replace the -01?

Thanks

Craig

The -02 had 16 RAM sockets and 16 ROM sockets. They both ran the same software but because the SIM4-02 had enough RAM to hold the entire 1702 code, it had a separate copy program, A0544. I wanted to be able to make copies easier on my SIM4-01 but with only 4 RAMs, I could only do 1/2 at a time so I wrote a program that did it in sequence, first read 1/2, program 1/2, read the second half, then program the second half. I used 4 of the lights on the MCB4-10 to sequence things for the person making the copies. The lights would all turn on any time there was an error in either the copy from master to RAM or after programming, the RAM to to the programmed EPROM. This way the operator new there was a problem. Depending on where it errored, one could also tell if one had bad RAM or programming failure. The MCB4-10 has a switch to turn the programming voltage. I always turned it off when the master EPROM was in the socket, just in case it didn't skipped a step. It used the switch for the TEST input to sequence the steps.
Dwight