...and at the ready with fingers poised on "STOP+INITIALIZE" and "RUN"! Great story and pic, Rob.
| VCF Southeast | Aug 01 - 02, 2026, | Atlanta, GA |
| VCF West | Aug 01 - 02, 2026, | Mountain View, CA |
| VCF Midwest | Sep 12 - 13, 2026, | Schaumburg Convention Center, IL |
| VCF Montreal V2.0 | Nov 07 - 08, 2026, | Saint-Lambert, Montreal, Canada |
| VCF SoCal | TBD, 2027, | Southern CA |
| VCF East | TBD, 2027, | InfoAge, Wall, NJ |
| VCF Latam | TBD, 2027, | Bahía Blanca, Argentina |
| VCF Pac. NW | TBD , 2027, | Tukwila, WA |
| VCF Southwest | TBD, 2027, | Westin Dallas Fort Worth Airport |
-
Please review our updated Terms and Rules here
Honeywell 200 resurrection
- Thread starter RobS
- Start date
RobS
Experienced Member
...and at the ready with fingers poised on "STOP+INITIALIZE" and "RUN"! Great story and pic, Rob.
Well spotted Bill. That was of course the whole point of the photo, which he wanted to send to other former engineer friends of his. He said that when testing the machines with their test software they just kept hitting those same buttons over and over again. He remarked that he didn't remember the buttons being so hard to press so I reminded him that at our age, both of us being over eighty, our muscles aren't as strong as they used to be.
RobS
Experienced Member
I thought that the next time that I was working on a board I would check the batch dates of the ICs against the date stamped on the edge connector to see which was the earlier date. Yesterday I tried that but came to no conclusion. There was just one IC on that board, the rest of it containing transistors and indeed the date on the board was a little earlier than that on the IC just as in your case. However there was also a second date later than that on the IC stamped on the board below the previous one. Just my luck. I'll keep looking.
I recognized that position immediately! He should hang that on a wall.
Those pushbuttons are pretty sturdy, and I can still hear the sound they made in my head (or if I press a few on my H2015 control panel, but I don't need a reminder!
). I can even tell the difference between the momentary type and the on/off type (like the ones used for the SENSE switches)...they had a different sound and feel to them. Mine are still working after all these years.That's something I had a discussion about a few weeks ago with my dad. I was asking him about the FORCE CYCLE knob on my H2015 control panel (and in others). It has positions for 10us, 100us...500ms on it (I decided a pic is worth more...see below) and I was wondering what it was used for. He said that you would punch in a program and use that feature to basically repeatedly reset the CPU after every 10us, 100us...etc. Basically a STOP+INITIALIZE followed by RUN after letting the program run for a period of time. I suspect there were cases where just using a branch instruction wouldn't work. For systems without that feature, like the H200 control panel, there was a board you could insert into the CPU to essentially do the same thing. Many of the systems with consoles had it built into the control panel.
RobS
Experienced Member
Bill: That control panel appears to be solely for running diagnostics rather than operating the machine normally, but the larger machines were operated using consoles instead of the control panel used on the basic H200. To rerun tests initialising ensured that each test started off from the same machine state regardless of how the previous test left it. Nothing could be assumed if a fault was suspected I suppose.
I have now finished building my SRAM memory unit shown below. The unit itself occupies just one of the five connector boards which replace the five jumper boards that normally link the core memory unit to the CPU via the backplane wiring.
This second picture shows the unit installed directly on top of the core memory unit with its ribbon cables running between the boards. The left hand connector board brings in from the CPU the write control signals, the middle two bring in the lower and upper address lines, the right hand one with the SRAM unit on it brings in the data to write and the board a few rows down from it sends back to the CPU the data read out.
In principle with some extra logic and wiring both the SRAM and core memory could be used simultaneously but I have no reason to set this configuration up at the moment and the SRAM unit cannot be regarded as a regular part of the replica machine as such as it uses technology not compliant with the restrictions that I apply to all components used in the replica machine.
I can't actually use either memory yet as, apart from more wiring still needing to be added to the backplane, there are no data registers at all in the CPU yet because I have focussed all my efforts on setting up all the basic address handling components needed. At minimum I need to create the memory local register, which is the data interface between the CPU and memory during both reading and writing data. It is also the register that the control panel accesses, so it will enable me to enter and display the data contents using the panel. Fortunately data registers have far fewer bits than address registers, so their construction is a lot easier.
I have now finished building my SRAM memory unit shown below. The unit itself occupies just one of the five connector boards which replace the five jumper boards that normally link the core memory unit to the CPU via the backplane wiring.
This second picture shows the unit installed directly on top of the core memory unit with its ribbon cables running between the boards. The left hand connector board brings in from the CPU the write control signals, the middle two bring in the lower and upper address lines, the right hand one with the SRAM unit on it brings in the data to write and the board a few rows down from it sends back to the CPU the data read out.
In principle with some extra logic and wiring both the SRAM and core memory could be used simultaneously but I have no reason to set this configuration up at the moment and the SRAM unit cannot be regarded as a regular part of the replica machine as such as it uses technology not compliant with the restrictions that I apply to all components used in the replica machine.
I can't actually use either memory yet as, apart from more wiring still needing to be added to the backplane, there are no data registers at all in the CPU yet because I have focussed all my efforts on setting up all the basic address handling components needed. At minimum I need to create the memory local register, which is the data interface between the CPU and memory during both reading and writing data. It is also the register that the control panel accesses, so it will enable me to enter and display the data contents using the panel. Fortunately data registers have far fewer bits than address registers, so their construction is a lot easier.
That's correct, it was mostly used for maintenance purposes, although you could use it to clear memory as well in conjunction with the SYSTEM CLEAR button. I learned a little about some of these features initially on an H2050A and later an H3200.
Even small systems like an H200 had the console as an option but I think it still required a control panel like yours as it didn't have the reduced version of the control panel like mine (see pic below, I don't recall what the two buttons above the keyboard were) Not all of those smaller control panels that were part of the console option included a diagnostic section like mine does (I want to say the H1200 didn't)...they just had the lower section which basically included the bottom row of buttons on your panel.
Nice job on the SRAM memory unit...looks good! That's gotta be handy for bringing the computer up one step at a time. Seeing the drawer loaded with cards...it's looking like an H200 again!
I guess since much of the logic for those machines was TTL that you didn't need any level shifters on the SRAM board?Having the local memory register working will be useful. How much of the CPU do you have built, or have you started on it yet?
durgadas311
Veteran Member
There was a version of the console, at least on later models, called 220-3 that provided an interactive version of the control panel via the keyboard and eliminated the full control panel, leaving only a few basic buttons.
RobS
Experienced Member
I don't know much about the console but assume that there had to be at least an interrupt button to get the CPU's attention. Equally there may have been a button to disconnect the console when it wasn't being used, but I'm just guessing.
Bill: Regarding that picture, apart from it being too blurred to see much detail you may not understand how little I had to do with the computer back in the 1960's. I was a programmer working in an office across the corridor from the computer room and I seldom had any reason to go there, especially because my programmes almost invariably worked first time, so unlike the other programmers I wasn't involved in debugging them on the computer itself. That might sound like bragging but at my age there is now nobody left to back up what I say about those times, so what else can I do? To a great extent my current project is intended to give me the hands on experiences that I never encountered back then. It was only by striking up a friendship with our resident field engineer Brian (who is amazingly now helping me with my project), whose little workroom adjoined the computer room, that I got to play with the machine with him on rare occasions to find out what actually happened in situations where the programmer's manual just stated that "results are unspecified". Just because the results were unspecified it didn't mean that they were unpredictable, not so long as the machine wasn't modified or replaced by a newer one and in those days the objective was to get the maximum possible performance out of what we had at the time, so I did just that working beyond all standard programming conventions to get the job done. Maybe the only testimony to my past reputation as a programmer is the very fact that so many years later Brian is still so willing to help me with my ludicrously ambitious project knowing that I enjoy tackling apparently insurmountable problems and often as not overcome them.
In those early days not only was the computer just across the corridor from the programmers' room but there was a large window onto the corridor the length of the computer room so that staff passing by could see it working. Over the years as we got bigger and bigger computers the computer room moved further and further away from our room and eventually we communicated with the machine using terminals so never had to go into the computer room, which was by then protected by a security door. Subsequently the computer was installed in a custom built security building a mile away from our office and the only time I even got into that building was when I was on 24 hour call to go there to advise the operators and fix any problems that occurred during the high volume batch runs overnight. Even then I only got into the office adjoining the computer room where the operators' desks and terminals were.
The drawer in the picture appears to contain a memory and if those memory modules are 4k ones then it would be a fully populated 16k drawer as used in the H200. However it is shown with the logic boards in the vertical working position whereas in the H200 lowboy cabinets the drawers didn't slide out but rotated out so that the boards were horizontal during maintenance as my pictures always show my backplanes. The entire layout of the backplanes is designed to make working on them easy that way up whereas with the drawer in the position in your picture everything is on its side and one would go mad trying to work on that for any length of time. Certainly I couldn't wire up an entire backplane with it that way up. Was that drawer possibly in a display case rather than in a computer? I ask because I don't know what the covers on them looked like and I need to make a cover for mine so that I can install the cooling fans and run everything for long periods without it overheating. Obviously on the odd occasions that I saw a drawer open in our H200 Brian was working on it with the cover removed, so I never noticed what they looked like.
Regarding TTL compatibility of the H200 signals, the ICs that I am using are certainly TTL compatible but the earlier transistorised boards in the memory unit actually use plus and minus fifteen volt supplies although the outputs are clamped to the five volt supply I think, so they do seem to be compatible with TTL. I had some trouble with the control panel interface because I think the panel was optimised for use with the transistorised boards and its primitive push button circuits seem happy to generate output signals outside of the normal TTL range and equally the peculiar SCR circuits operating the lights have their own opinions about what voltage ranges they work reliably between. I think the problem is that those SCR circuits require current limited inputs rather than voltage limited ones although I think I now have a reliable solution. I still haven't found any schematics for the original control panel interface in the H200 to help me with that.
As to how much of the CPU I've built, I wouldn't like to think about that. Certainly the control memory loop, which is now in place, is the heart of the machine as so many instructions process fields of indeterminate length primarily under the control of the punctuation bits in memory. In fact in a two address machine like the H200 the programmer has little say in what happens within the CPU as there are no directly programmer controlled data registers in it and almost everything that the programmer needs to think about is in the main memory, where all the results of operations end up. Even the index registers used in address modification are just predefined main memory locations that can equally be used as ordinary memory when no indexing of addresses is in use.
I don't think about how much work I still have to do as this project is a pastime just as playing games is about enjoying the games themselves and not necessarily about acquiring a cabinet full of trophies at the end. All l can say on that subject is what I originally wrote on the opening page of my website, which hasn't changed at all since I first wrote it years ago. To quote the opening paragraph of that page, "It is conventional for a computer to have both a name and a number and the replica won't be entirely Honeywell so its name is only "Honey" and as for Pi, well just like that number the project may be irrational and never end." I hope to have the machine doing something meaningful by the end of this year, but I have other issues to tackle in my personal life that may get in the way.
Bill: Regarding that picture, apart from it being too blurred to see much detail you may not understand how little I had to do with the computer back in the 1960's. I was a programmer working in an office across the corridor from the computer room and I seldom had any reason to go there, especially because my programmes almost invariably worked first time, so unlike the other programmers I wasn't involved in debugging them on the computer itself. That might sound like bragging but at my age there is now nobody left to back up what I say about those times, so what else can I do? To a great extent my current project is intended to give me the hands on experiences that I never encountered back then. It was only by striking up a friendship with our resident field engineer Brian (who is amazingly now helping me with my project), whose little workroom adjoined the computer room, that I got to play with the machine with him on rare occasions to find out what actually happened in situations where the programmer's manual just stated that "results are unspecified". Just because the results were unspecified it didn't mean that they were unpredictable, not so long as the machine wasn't modified or replaced by a newer one and in those days the objective was to get the maximum possible performance out of what we had at the time, so I did just that working beyond all standard programming conventions to get the job done. Maybe the only testimony to my past reputation as a programmer is the very fact that so many years later Brian is still so willing to help me with my ludicrously ambitious project knowing that I enjoy tackling apparently insurmountable problems and often as not overcome them.
In those early days not only was the computer just across the corridor from the programmers' room but there was a large window onto the corridor the length of the computer room so that staff passing by could see it working. Over the years as we got bigger and bigger computers the computer room moved further and further away from our room and eventually we communicated with the machine using terminals so never had to go into the computer room, which was by then protected by a security door. Subsequently the computer was installed in a custom built security building a mile away from our office and the only time I even got into that building was when I was on 24 hour call to go there to advise the operators and fix any problems that occurred during the high volume batch runs overnight. Even then I only got into the office adjoining the computer room where the operators' desks and terminals were.
The drawer in the picture appears to contain a memory and if those memory modules are 4k ones then it would be a fully populated 16k drawer as used in the H200. However it is shown with the logic boards in the vertical working position whereas in the H200 lowboy cabinets the drawers didn't slide out but rotated out so that the boards were horizontal during maintenance as my pictures always show my backplanes. The entire layout of the backplanes is designed to make working on them easy that way up whereas with the drawer in the position in your picture everything is on its side and one would go mad trying to work on that for any length of time. Certainly I couldn't wire up an entire backplane with it that way up. Was that drawer possibly in a display case rather than in a computer? I ask because I don't know what the covers on them looked like and I need to make a cover for mine so that I can install the cooling fans and run everything for long periods without it overheating. Obviously on the odd occasions that I saw a drawer open in our H200 Brian was working on it with the cover removed, so I never noticed what they looked like.
Regarding TTL compatibility of the H200 signals, the ICs that I am using are certainly TTL compatible but the earlier transistorised boards in the memory unit actually use plus and minus fifteen volt supplies although the outputs are clamped to the five volt supply I think, so they do seem to be compatible with TTL. I had some trouble with the control panel interface because I think the panel was optimised for use with the transistorised boards and its primitive push button circuits seem happy to generate output signals outside of the normal TTL range and equally the peculiar SCR circuits operating the lights have their own opinions about what voltage ranges they work reliably between. I think the problem is that those SCR circuits require current limited inputs rather than voltage limited ones although I think I now have a reliable solution. I still haven't found any schematics for the original control panel interface in the H200 to help me with that.
As to how much of the CPU I've built, I wouldn't like to think about that. Certainly the control memory loop, which is now in place, is the heart of the machine as so many instructions process fields of indeterminate length primarily under the control of the punctuation bits in memory. In fact in a two address machine like the H200 the programmer has little say in what happens within the CPU as there are no directly programmer controlled data registers in it and almost everything that the programmer needs to think about is in the main memory, where all the results of operations end up. Even the index registers used in address modification are just predefined main memory locations that can equally be used as ordinary memory when no indexing of addresses is in use.
I don't think about how much work I still have to do as this project is a pastime just as playing games is about enjoying the games themselves and not necessarily about acquiring a cabinet full of trophies at the end. All l can say on that subject is what I originally wrote on the opening page of my website, which hasn't changed at all since I first wrote it years ago. To quote the opening paragraph of that page, "It is conventional for a computer to have both a name and a number and the replica won't be entirely Honeywell so its name is only "Honey" and as for Pi, well just like that number the project may be irrational and never end." I hope to have the machine doing something meaningful by the end of this year, but I have other issues to tackle in my personal life that may get in the way.
durgadas311
Veteran Member
The is the "control panel" that came with the 220-3 console. This diagram has space for an additional button for 4-char addr mode, and 4 more sense switches.
Section 8 (VIII) of this equipment operators manual gives some description of the 220-3 console: https://www.bitsavers.org/pdf/honey.../278_Model_120_Equipment_Operators_Manual.pdf. This includes the commands allowed on the keyboard, to replace the other functions of the "full control panel". It would have been interesting to see just how this keyboard connected to the CPU and performed those functions. My modern brain thinks of some microcontroller and firmware, but I'm guessing that was not how it was done.
Section 8 (VIII) of this equipment operators manual gives some description of the 220-3 console: https://www.bitsavers.org/pdf/honey.../278_Model_120_Equipment_Operators_Manual.pdf. This includes the commands allowed on the keyboard, to replace the other functions of the "full control panel". It would have been interesting to see just how this keyboard connected to the CPU and performed those functions. My modern brain thinks of some microcontroller and firmware, but I'm guessing that was not how it was done.
The photo of the memory drawer is in the "working" position. Sorry for the fuzzy photo but it's around 50 years old. Here is the entire photo for better context:
As far as the covers go. some had a thin (maybe 100mils), clear but frosted cover on a pair of hinges with a magnetic latch that covered the PCB side.
That's a simple, elegant solution to clamp the memory signals to 5V...I like that.
Keep enjoying your work on the machine. Good to hear Brian is there to help out. Hopefully your personal issues are nothing serious.
Exactly, and some had the maintenance panel under a hinged door at the top as you can see in my H2015 control panel. Not all had the maintenance panel and some had different configurations when they did. The coolest, IMHO, was the H2050A which also had a row of small incandescent bulbs for the address bits, much the same as the big control panel does.
Not a great pic of it, but here is a H2050A system my dad had at one of his warehouses. If you look closely, the maintenance panel on the control panel is noticeably larger than for the H2015 to accommodate the row of incandescent address lights which are the same ones used on the H4200/H8200 control panels (snippet below). I still have the H4200 control panel.
RobS
Experienced Member
Ah, now I understand. It was a picture of a lowboy cabinet with the side panel removed so that the memory drawer at the right hand end was visible. That explains the peculiar frame visible in the close-up.
There's a nice view of a 223 high speed card reader in the foreground of one of the other pictures. I always liked that device because of its clever geometry with the angled feed chute and triangular cabinet. Despite the high operating speed the cards went in at an angle but came out the normal way up on the top output slide. One of our chief operator's tricks when he needed to remove just one card to inspect it was to pull back the pressure plate on the top slide so that as the card was ejected it flew up into the air so that he could catch it in mid air like a conjuror about to say "Is this your card?" He had previously attended the same boarding school as me but before my time there and was probably a keen cricketer, unlike me.
As our first machine only had punched card devices a lot of time was spent just feeding boxes and boxes of cards into that reader whereas sites with other storage for bulk data wouldn't have used the device so much. Prior to installation of that first computer the company had for many years used punched card equipment with a tabulator for printing, so all the data was already on punched cards that were updated by hand in the filing boxes. The H200 was purchased primarily to replace the tabulator although we soon used it for much more before a larger Honeywell machine with tape drives was installed about three or four years later. However, that honeymoon period with that small punched card machine was the most memorable part of my career.
There's a nice view of a 223 high speed card reader in the foreground of one of the other pictures. I always liked that device because of its clever geometry with the angled feed chute and triangular cabinet. Despite the high operating speed the cards went in at an angle but came out the normal way up on the top output slide. One of our chief operator's tricks when he needed to remove just one card to inspect it was to pull back the pressure plate on the top slide so that as the card was ejected it flew up into the air so that he could catch it in mid air like a conjuror about to say "Is this your card?" He had previously attended the same boarding school as me but before my time there and was probably a keen cricketer, unlike me.
As our first machine only had punched card devices a lot of time was spent just feeding boxes and boxes of cards into that reader whereas sites with other storage for bulk data wouldn't have used the device so much. Prior to installation of that first computer the company had for many years used punched card equipment with a tabulator for printing, so all the data was already on punched cards that were updated by hand in the filing boxes. The H200 was purchased primarily to replace the tabulator although we soon used it for much more before a larger Honeywell machine with tape drives was installed about three or four years later. However, that honeymoon period with that small punched card machine was the most memorable part of my career.