• Please review our updated Terms and Rules here

Honeywell 200 resurrection

I finally got around to setting up the website http://www.honeypi.org.uk/ although there's only a couple of pages on it so far. For the time being if you have any comments about it, especially about my amateur attempt at website authoring, please post them here. When I make a significant update to the website I'll mention it here again.

On the hardware front I've been nursing forty year old PSUs back to life and adapting them to provide all the voltages needed by the core memory logic. Once that's done we can start testing the core memories themselves and find out whether the project is viable. It'll be a lot more viable when the weather gets warmer here.

Somebody told me that many surplus machines from the Honeywell 200/1200/2200 range were bought up by EUR Systems in Mechanicsburg PA and cannibalised for spare parts for their own machines, which they were still using in year 2000. I wonder what happened to them after that, but the fossil hunt will have to wait as we have enough parts to play with for the time being.
 
The link on this page to the Programmer's Reference manual is actually a duplicate link to the logic guide: http://www.honeypi.org.uk/links.htm

Do you have a copy of the PDF for the Programmer's Reference manual?

I was going to write an emulator using a Raspberry Pi. I know the goal is a total reconstruction, but being able to write EasyCoder and get things up and running from a software point of view would be a good parallel task.
 
Mea culpa -- That's Latin of course, allegedly a dead language but not as dead as Easycoder. I've corrected the incorrect link now, so you should find it's okay if you refresh the page in your browser. Thanks for telling me and also the "omg" in the Retro Computing forum. I get that feeling myself quite frequently but it's just a matter of taking things a step at a time and so far we haven't found a reason why we can't succeed -- eventually. So far all that I've produced is the smell of ancient grime burning off of long disused equipment which is experiencing electricity again.

I did explain the reference to Pi that confused you on my website, that the project may be irrational and never end just like Pi, a fatalistic mathematical joke. Although you are contemplating an emulator on the Raspberry Pi our intention is to use an Arduino ATMEGA2560 with an interface to a PC to run tests on the partly completed hardware, but in fact the ATMEGA2560 could easily emulate the H200 and run the control panel all by itself.

I have an incomplete and primitive PC-based H200 cross-assembler, emulator and memory dump programme written in C++ for Borland C++4, itself now virtually vintage software as I bought the Borland package eighteen years ago and still use it. I wrote the software as disposable code just to test my Pi Factory Easycoder programme, but if it's of any interest I could make it available on the website.

I was intending to put the details of the Pi Factory algorithm on the website soon anyway as it is just a mathematical approach to the calculation of Pi designed to work on the simplest form of computer, not even needing multiplication or division capability, which the most basic H200 didn't have. I'll try to add it to the website in the next few days. As it is progressive, outputting digits as soon as they are available, there is no predetermined number of decimal places to which it will run. It just stops when the whole of the available memory is full of partially completed calculations and it can't add any more. That means that it gets slower as it progresses as well, so it is an interesting test of both how fast a particular computer runs it and also how far. It is so undemanding that it could probably be programmed into any of the vintage machines covered in this forum, which might give rise to some interesting comparisons. Of course the original Honeywell 200 only had to deal with one comparison to sell successfully from the outset and the result of that comparison is history.
 
... 1000 logic boards? Wow. How many gates per board? ...

I finally compiled an inventory of the ICs in my stockpile of Honeywell logic boards. There are over 5400 ICs which provide in total approximately 5000 AND-OR gates, 3000 AND-OR-NOT gates and 1000 flip-flops. That should be enough to build a small computer, certainly something more than a PDP8. Perhaps someone has some statistics on the number of gates in early computers. I just wonder how many of the ICs will have to be transplanted onto new PCBs before we've finished. I have a PCB manufacturer with wave soldering equipment interested in the project, so perhaps we'll need his help eventually, but is there such a thing as wave unsoldering equipment?:)
 
Last edited:
As promised my website now contains this page The Pi Factory which describes the demonstration programme. The Easycoder version with its manic optimisations is shown but there is also a saner version in C++ to assist understanding of the mathematical process used. The source code for that is included in case anyone wants to try something similar on another computer.
 
... is there such a thing as wave unsoldering equipment?:)
Yes, although it's usually called a blow torch ;-)

If the pins aren't bent and the boards are small enough you can also try putting them upside down into a toaster oven set on broil; a good tap and everything falls out.
 
Last edited:
I understand that the overheads in a PDP8 were similar to those in an H200, i.e. lots of gold-plated edge connectors and a complex wire-wrapped backplane, if you're planning to build it in the original manner rather than just as an electronic equivalent. Our advantage is that we have over a thousand logic modules with their edge connectors intact but, having checked on the cost of getting new custom PCBs made, about half of that cost appears to go on getting the edge connectors gold plated, especially the way the price of gold is currently. Then there's the silver-plated kynar wire for the backplane and the matter of finding a wirewrap gun, preferably electric for extensive work, with the right size bit for the backplane pins. I've got that side of things covered though. I read somewhere that gold-plated edge connectors were prone to deterioration over time but ours seem okay. Maybe that's because they were removed from the backplanes for long term storage. All in all this project seems to be as much about logistics as electronics.
 
Yes, although it's usually called a blow torch ;-)

If the pins aren't bent and the boards are small enough you can also try putting them upside down into a toaster oven set on broil; a good tap and everything falls out.

It looks like all the ICs have at least two pins bent. Probably they were inserted by hand during manufacture as I have Honeywell's manual assembly instructions for them. The jumper wires on the boards are the usual giveaway as machine-inserted jumpers tend to have dummy components on them while manually inserted ones are plain wire; at least that's how I read it. For the ICs to drop out during heating I'd have to clamp weights onto them, which I have considered doing if I go in for bulk disassembly. Today I managed to wrestle two reticent zener diodes off of one of the boards. At that rate I'll be an old man before I finish. Come to think of it, I already am an old man so it doesn't really matter.
 
I understand that the overheads in a PDP8 were similar to those in an H200, i.e. lots of gold-plated edge connectors and a complex wire-wrapped backplane, if you're planning to build it in the original manner rather than just as an electronic equivalent. Our advantage is that we have over a thousand logic modules with their edge connectors intact but, having checked on the cost of getting new custom PCBs made, about half of that cost appears to go on getting the edge connectors gold plated, especially the way the price of gold is currently. Then there's the silver-plated kynar wire for the backplane and the matter of finding a wirewrap gun, preferably electric for extensive work, with the right size bit for the backplane pins. I've got that side of things covered though. I read somewhere that gold-plated edge connectors were prone to deterioration over time but ours seem okay. Maybe that's because they were removed from the backplanes for long term storage. All in all this project seems to be as much about logistics as electronics.

I am planing to build a close copy of the pdp8 but using modern transistors on copys of the original pcbs in a copy of the pdp8 straight 8 case, I think why gold-plated edge connectors could deteriorate is if there is contaminants under the gold and damp gets in, it would probably be worse if the backplane pins are not also gold plated, and the board left in backplane and damp
 
Last edited:
... I think why gold-plated edge connectors could deteriorate is if there is contaminants under the gold and damp gets in, it would probably be worse if the backplane pins are not also gold plated, and the board left in backplane and damp
Perhaps the important point was that they weren't in the backplane as they spent almost forty years in open crates in a garden shed, probably not the serious conservationist's ideal environment. I was scraping the rust off the cases of some of the power transistors yesterday and the wiring on a few boards has signs of verdigris but at least I've cleaned off all the spiders' webs and removed the dead woodlice. Some boards have small electrolytic capacitors on them which look like they've suffered from frost damage, but that's fixable. So far everything not obviously broken that I've tested has worked amazingly. Nevertheless even if our creation looks and behaves like an H200 it is still going to be more like a Frankensteinian monster under the skin. I think I'm definitely earning my stripes as a resurrectionist.
 
I was working on a badly water damaged commodore 4016, and one power transistor was very badly rusted, I found that socking in a vineger with a bit of washing up liquid, and scrubbing with and old tooth brush every day, the transistor and rusty case screws came up nicely, I then sealed it with wood vanish. you would have to unsolder the power transistors, and 40 year old electrolytic capacitors are likely to unreliable after 40 years of non use and cold, even if reformed, I would love to see some pics of the work in progress on the Honeywell 200 and the old boards before and after.
 
Last edited:
Actually I'm not too worried about the rust; hopefully it's superficial. Some of the logic boards had mould growing over them; it has a particular liking for the plastic insulation on wires for some reason. I've already tried the washing up water and toothbrush treatment on them along with carefully drying with a hairdryer and they seem quite happy with it without any need to disassemble, but I wouldn't attempt that with the power regulator boards. I did consider painting the power transistors but if so would have used heatproof engine block paint I think, but anyway I'd prefer to give the cases the chance to dissipate heat. I know it's only a small proportion of the total but every little helps.

The only electrolytics of any importance are the Farad's worth in the PSUs and they are bearing up okay so far after reforming, but then they were never left in the garden shed. I'll be using the PSUs as a bench supply during the development so that we know how reliable they are when we come to put it all together. As we have no idea what the total power needed will be there's no point in considering the final design for the PSUs now. The main transformers to hand can handle a kilowatt but I've no idea whether that will be enough. There's also the complication that this is American equipment marketed in Britain, so its design uses a mixture of British and American mains voltages and I have to be careful just connecting up the mains wiring. I've discovered that it isn't that easy to buy ordinary American domestic three pin plugs in Britain but, having found a supplier, I can now be more disciplined about what plugs in where.

When I find the time (!) I'll put a gallery of photos on my website, but my workroom is so tiny and full of shelving that taking photographs with anything but a fisheye lens is pretty difficult. Certainly both I and my colleague Marcel have taken photos of the individual components and I'll try to put some of those together in a meaningful fashion. To photograph my entire collection of H200 components in one go I had to convert our whole living room into a photographic studio but my wife insisted that it be restored at the end of the day, so that exercise won't be repeated very often. She does understand me though. When I gave up the project a couple of years ago and started to write a novel instead she thought I was mad, so which is saner, writing a novel or building a mainframe computer in the spare room? I suppose it just depends who you are. As the novel appears to be one of a trilogy I'm still working on that as well whenever I get frustrated with this project. That's another reason why the photos may take a while coming.
 
be careful, some vintage computers like the pdp8 used a special transformer that is self regulating but is locked to supply frequency not voltage, it is easy to spot is it had a winding with only high voltage capacitor or capacitors across it, and no regulator just some power diodes and smoothing capacitors.
and I cant wait to see your work.
 
You mean a constant voltage transformer I assume. Yes, Honeywell loved them and they're all like that but the ones that I have are clearly marked for 50 Hertz use, one of the few concessions that they made to our mains supplies. They came out of equipment used in the office where I worked and I'm following the original engineer's manuals for the wiring, so there shouldn't be any problems. There was no other voltage regulation in the PSUs themselves as you say, but in the equipment from which I removed the backplanes there were shunt voltage regulators acting like high power zener diodes which trimmed the voltage to the right level and kept an even load on the PSUs. I'm keeping the same setup for the bench power supply, but that was only for the 5 Volt supply to the logic boards. Marcel's collection of original second generation boards which drive the core memories need 15 and 28 Volt positive and negative supplies as well, so I'm having to improvise all those. I've converted a couple of dual diskette drive PSUs from Honeywell Level 6 computers to provide 34 Volt supplies which can be regulated down to 28 Volts using the shunt regulators. The diskette drive PSUs, which date from the 1970s, used CVTs as well as voltage regulator ICs to provide a level voltage, but of course I've removed the ICs because they're the wrong voltage apart from being outside the project terms of reference at present.

As you can't wait to see my work here's a picture of the 5 Volt regulator screwed up above my workbench. IMG_0264A.jpg The regulator is an unmodified Honeywell module and the economy resistors above it, made out of electric fire element wire, reduce the current to a level that the regulator can handle when the output load is light. The meter indicates whether the regulator is operating within a suitable range. This is just test bench equipment though, not a part of the final machine. I haven't fitted the cooling fan yet but it will need one.
 
why is a shunt regulator used? it is not good for the electric bill and makes a lot of heat when the load is low? I have only ever seen them used on wind generates, as they need a load at all times to stop over speed. you work it very neat.
the pdp8 used constant voltage transformers for -15v +10v and +36v for the core memory, I will not be able to find constant voltage transformers so I will use conventional transformers with Linear regulators like the 7810 7915, and split the load across the regulators and transformers, I have a horde of 12v transformers that will do the -10v and +36v, and some others that may do the -15v, this will not be as efficient as the original constant voltage transformers used, and will probably be heaver! the 78 and 79 type regulators are the only ICs I will have in the main PDP8.

what will you use to store your programs on the honey pi? and punched cards can still be found, but I like the idea of using used rail tickets, and they are free!
 
Last edited:
Back
Top