Keyboard and Case Come Together

I haven’t been posting but I have been doing, sort of.

I wanted to get going and build an externally complete mockup of Don’s TVT prototype, so that I could have it on my desk to inspire and motivate me to keep working at it.

One of the problems with trying to replicate something you have no physical access to is that, the more you research, the more you find out that changes your assumptions.  When I got my Mohawk keyboard, I assumed by the keytops and arrangement of keys alone I had the right beast.  But looking around online I stumbled across a couple of pics where one of the plastic ‘bezel’ blocks Don used around the periphery of his keyboard (to cover up the workings underneath) had gone missing and the others had slid around.  They revealed a metal slotted bracket.  This was not the same as the one on my Mohawk, which led me to wonder — did I have the right keyboard?

I needed to find a pic of the so called ‘surplus’ keyboard outside of the TVT unit.  Hunting around www.tinaja.com, I discovered this article Don had written detailing his ‘improved ASCII encoder’.  Check this out:

keybarticle

Look familiar? This article appeared in 1974 and featured a revised ASCII encoder from the original used in the TVT. For this version, they mounted the encoder beside the keys on the right side.  Now, the width of the prototype is known — the museum gave that to me.  The keyboard pictured as I said is post-TVT and the encoder Don’s attached to the rails on the right side make it too wide to fit.  But the flange looks the same.  So what’s up?

As it turns out, the answer is fairly simple.  Both mine and Don’s keyboard were ultimately made by the same company:  Micro Switch of Freeport, IL.  Micro Switch is pretty prolific.  I’ve actually got an ebay search followed and have seen tons of their former products go onto market.   Micro Switch produced keyboards like ours for Mohawk, Honeywell and others.  From what I’ve been able to dig up, it appears Don’s units may have come from decommissioned Honeywell machines – I’ve seen a few with different coloured keys but the same flanges on the sides.  My keyboard’s switches date to late 1967, so that kind of explains why they were surplus in 1973.

Now, here’s where I get into a battle with my inner historian.  When Don was building his keyboards, these Micro Switch units weren’t rare at all.  They were literally being dumped.  However, here in 2016 these are a lot rarer.  In 7 years of watching around, I’ve only seen mine and one other appear on ebay.  Because the legends on mine are white and I only have a handful of blue keys, to get the black and blue coloration Don had, that means I have to paint.  And paint, as they say, is forever.  Ugh!  What to do!

Well, I guess while I have that debate I can at least take it apart.  That is reversible, and it’ll let me do a mockup of my unit and figure a few things out about my case design.  So.. here we go…

image

The switches are held in place on their rails by two screws up top. Loosening the screws a little is all it takes to be able to move them. I had to cut the mass of wires shorter to work with it, which was in itself painful but — well, it’s not like they had their original connector at the other end anyway, they were cut at about 2 feet or so.  After, I remove several switches to get the number of keys the same as Don’s, then begin to try to rearrange them get the pattern Don had. The basic pattern is already there for the most part. It’s just dropping switches from the right and filling in the blanks with the spacers.

image

With the pattern figured out, I proceed to test fitment. It appears I will have to router into the left side a bit more deeply than right to compensate for the keyboard having more exposed rail on that side.

image

I also work to get the plywood top omin place.

image

image

Ok.. need to trim this down a little

20151004_164630

Good.. now to cut the opening for keyswitches..

20151004_165145

Meh.  My routering skills are a bit unpolished.  But whatever, it’s a mockup.  Alright.. now a bit of glue to see how it goes together:

20150927_171547

Not bad.   Now to router out the sides for the keyboard and try placing it.

20150927_203703

Hey!  This thing’s actually starting to look like a TVT.. here’s mine next to my not quite life-size photo of the original:

20150927_194430

Now the moral crisis.  I’m tempted to ‘leave them keys alone’.  You know, authentic recreation rather than replica.  But of course, if I’m not going to build an actual replica.. what was the point again?

While internally debating that situation, I decided to hunt down some appropriate vinyl.  The vinyl Don used was 1970s faux leather, I think.  Finding the exact colour and pattern would be very difficult, since it was probably long discontinued.  Further complicating matters, photos play tricks on you.  In some photos the vinyl on the prototype looks beige, in others, gold.

I just want to get my mockup built and then do some more photo comparisons.  I found some beige-ish vinyl at a Fabricland store in West Vancouver.  It looked kind of close.  I think Don’s may have been that fake ‘distressed’ leather, although maybe the distress had happened over the course of 40 years.  This vinyl is a bit too polished, but that’s okay.  Again, it’s just temporary to check things out.

20151004_172102

Getting out my ever trusty No More Nails, I began to put the case and rearranged keyboard together.  I ended up having to go deep into the left side to get the keyboard centered correctly, and then had to prop it on the right side with a little shelf.  Here’s the end result (with a printed pic of the switches to stand in for the real ones for now):

20151005_091347

Not bad at all, if I do say so.  Okay.  Time to sleep on the paint thing and decide if I have the stones to go all the way on this.

The Keyboard

Being a ‘TV Typewriter’ and all, it goes without saying that the keyboard is a rather important component of the thing.  No matter how well you design the electronics, they’re useless unless you have a physical means to input data.  And new keyboards in 1973 were expensive.  Unless, of course, you read Radio Electronics for tips on how to get around that problem.

Not that the TVT required an actual keyboard.  In the September 1973 Radio Electronics article, Don Lancaster suggested ‘six switches and a pushbutton’ would suffice.  Entering letters was a matter of setting the switches and pressing the button.  Don claimed with practice decent speed could be achieved, although I think he must have been a pianist to make that arrangement work with any proficiency.  Another option was the ‘low cost keyboard’ project offered in the February 1973 issue of RE.  Part of Don’s ingenuity was finding extremely affordable ways of doing things; his handmade key switches were crude but brilliant, although you still needed to secure keycaps from somewhere.  RE suggested ‘two shot’ moulded plastic keytops made by a company called Mechanical Enterprises of Arlington, Virginia.  I have no idea if they still exist.  The keytops themselves look pretty close to those found on IBM Selectric typewriters, except for extra functions printed over the letters.   You also needed to build your own ASCII encoder which would convert the signals generated into letters and numbers that would appear on your screen.

My first order of business in choosing to model Don’s prototype was to find out what kind of keyboard he used.  To refresh, here’s what the prototype keyboard looks like:

Link: Computer History Museum, Mountain View, CA

When I asked him about it, his response was that the keyboard was an ‘IBM EBCDIC’ surplus unit.  They had taken apart dozens of them and reconfigured into just over 40 complete units that were then offered to TVT builders.   I tried to google search using what he had given me but came up short; EBCDIC, which stands for Extended Binary Coded Decimal Interchange Code, like ASCII, was an encoding standard used my several keyboard manufacturers, not a specific model.   I asked in forums, on mailing lists.   Some claimed familiarity, but could only suggest they were ‘key to tape’ units.  One day I got an old school computer hobbyist who remembered something about the TVT prototype using surplus Mohawk ‘key to tape’ keyboards.  Again I went to Captain Google, and again he came up short.   All I had were photos of the prototype.

I realized I needed a Plan B, because finding anything involving key to tape or punch cards from 1973 today might prove to be impossible.  I thought perhaps I could recreate the keytops on the prototype by finding parts that were close and fashioning something that resembled them.  The underlying switches wouldn’t matter; they wouldn’t be seen.  Since the keytops looked kind of like teletype keys, I did some ebaying around for ‘teletype keys’, and one day came up lucky.  A vendor had 62 ‘Repeat’ replacement keys for military teletypes.  Apart from every one having ‘REPT’ stamped on it, they looked pretty close to what Don’s prototype had.  I figured making them look like the prototype would just be a matter of painting them.  So I ordered all of them.  Here’s what they looked like:

20150923_205221

If you look at the photo of the actual TVT prototype keyboard again, you’ll see these cylinders are set in the lower right corner of black square surrounds that are there to hide the switches underneath.  I assumed (incorrectly) that cylindrical keys slid down through the surrounds when pressed.  Thanks to my years of model railroading, I had learned how to cast plastic parts using two part resin.  I reasoned I could approximate the dimensions of these surrounds, create a mockup of one in styrene, and then make a mould of it and cast it.  Then I could adapt the ‘low cost keyboard’ project from the earlier RE article to it, creating a working keyboard that at least had the same exterior look.

I had that all settled up and began making plans, but then one day I had a stroke of luck and ebay coughed up just what I needed.  While searching for ‘Mohawk keyboard’, sure enough, one came up.  It looked just like this:

From ebay

After close inspection it became very clear this keyboard was what Don had used!  The placement and design of the keys were correct, the only difference being colour.  Woohoo!   This was it: an original Mohawk Data Sciences (MDS) key to tape system keyboard.

My understanding is that key-to-tape was a system that avoided the need to punch data into paper cards.  MDS apparently did very well for a time,  up until about the late 1970s.  I don’t know for sure what happened to them but I imagined it was the same fate as many other proprietary computer brands as the world slowly converged into widely held standards.  Many old hands refer to them as ‘IBM Mohawk’, but I see nothing to indicate IBM’s involvement anywhere at the time these key to tape systems were being produced.  Perhaps someone can clarify.

My elation at finally locating it was was stopped cold by the asking price, though: a cool $600.   I asked around on forums for opinions; even though the price was extremely steep, I reckoned the odds of coming across another unit like this any time soon were slim to none.  I was assured that these keyboards were nothing special — just salvage from old Mohawk key to tape units that had once been plentiful.  Most said they wouldn’t offer more than $50.  I asked the seller if they’d consider offers.  They declined, but eventually relisted for $100.   I opted to sit out, hoping it would go unsold and be marked down again, but it sold for just over $100.   I felt a little bummed — I knew these keyboards didn’t come up every day, and since the sellers themselves didn’t know what they were, searching them out on ebay would require hours of looking up every keyboard in the Vintage Computing section.  Who knows when I might see one again?  But then a year later, I did!  The asking price this time was $100 and I went for it.  Here it is:

Mohawk MDS Key to tape keyboard
Mohawk MDS Key to tape keyboard

Interestingly, it’s not exactly the same as the first one I encountered.  There are fewer switches up top, and the keys are coloured differently.  I’m guessing Mohawk customized these things depending on client and application.  Perhaps someone could shed some light on that.    The keyboard itself isn’t all that big, maybe about 18 inches wide, but man is it heavy.  You know those ‘steelie’ IBM clicky keyboards?  They have nothing on this beast.  Here’s a few shots around it:

20150922_204558

20150922_204645

20150922_204659

20150922_204754

20150922_205400

Way cool.  There are five screws to remove to get into it, and upon removing the case we are presented with the innards:

20150922_205618

The keys themselves sit a good few inches up from the base.  Compared to modern keyboards, the space inside is cavernous.

20150922_205623

20150922_205653

I was surprised to find a speaker in there.  I’m assuming that was there to alert to errors and so on.  The wire bundle looks pretty intimidating given how many there are.  Keeping all that straight looks to be a challenge.  In terms of operation, I was expecting a much more mechanical, clunky feel to them as several forum folks had said these key to tape units were a nightmare to type on (I actually found pressing the keys to be relatively smooth, although a bit awkward with the small size and wide spacing).  The switches themselves I *think* are faily simple contact switches — two wires each, press it down and the circuit is connected.     I’m hoping Don chose these for relative simplicity.  I was also wrong about how they operate — when you press on the cylinders, the entire thing, surround included, moves down.  The switch mechanism itself below the keytops seem massive for what they are.

You can see in this side profile view the solid blocks installed as spacers.  On my keyboard there is a single rectangular ‘LF’ key on the right side.  This key does not exist on the first ebay keyboard I came across; rather it is blocked off.

20150922_205633

According to Don, he cherry picked blue, red and black keytops from various keyboards for colour, and reassembled them to get the patterns seen in the prototype.   Based on what we can see here, it looks like before or after changing/painting keycaps, he removed some keys on the right side and blocked them in, dropped the whole keyboard into his TVT prototype case, and then to center it and hide the metal framing he used to mount it, added a vertical column of blocks on the left side, making it one column wider than it was originally.  Neat!  I don’t have enough ‘blocks’ to fill in the keys I will have to remove, but those should be easy enough to cast in resin and paint.

To be honest, I’m almost a little hesitant to take this apart.  The forum gurus say these units aren’t rare or collectible, but I’ve never seen one myself before and they can’t be that numerous after 40+ years.  It’s just a neat thing to look at and appreciate how much keyboards generally have evolved over the years.  The thought of painting over the original decaling is perhaps just a bit bothersome to me.  I’m sure I’ll get over it, but for now, until we get to the actual construction phase, I’ll enjoy it as is.

Next up, the ICs.

 

Building the Case Pt. 1

One of the challenges with replicating something you’ve never seen in person is getting the details right. In my modelling of the Kettle Valley Railway’s Coquihalla subdivision, I had a problem: the railway didn’t exist anymore. Not the line, scarcely a single building remained. All I had to work with were old photographs and the say so of others. In that situation, you gradually learn to pick out details. How to deduce sizes and distances in photographs using objects with known dimensions. It’s detective work, and quite honestly, I enjoy it.

Now, we’re not quite so far gone with the TVT prototype. The real thing does still exist. I could, if I wanted to, go and see it; however it’s a 2 day drive in another country. And I just don’t have time. Now, fortunately, the prototype’s current owner is the Computer History Museum in Mountain View, California. And they are nothing if not thorough in their catalog of their collection. They had record of the rough dimensions, and had some great frontal and overhead shots. Further, I was fortunate to get in touch with Dag and Chris from the museum, who were very helpful in providing some critical photographs, like this one of the back of the TVT unit, an angle no other photographer bothered with:

IMG_3572

I knew the case was made of wood, but that sliver of silver at the bottom of this picture shows was appears to be metal. Which would make sense, since that’s what my TVT2’s motherboard is mounted to. I’m not sure what the screws at the back are holding up; because the case is screwed shut and the museum understandably doesn’t want to risk damage taking it apart, that’ll probably remain a mystery for the forseeable future.

Here’s a kind of side shot Dag was kind enough to send:

IMG_3579

It’s hard to say for sure with just pictures, but I’m thinking the wood is probably hardwood — maybe fir or oak. Could be pine too but it looks too dense. I don’t see any screws attaching that wood support in the front to the sides. It appears Don routed into the side pieces wood and then secured the front ‘beam’ with either screws that are hidden inside or maybe glue. Probably I’ll go with No More Nails or something like that with maybe a couple of inside screws. Is the front a 2×4?

I asked Dag if the measurements on the museum’s website were fairly accurate, and he went above and beyond and got another museum staffer to measure the unit again, particularly the wood sides. It turned out that the dimensions were pretty correct, the side wood being 4.25″ high by .75″ thick, the length of the side pieces being 13.125″ at the bottom and 12″ at the top (Don cut the ends on an angle) the entire unit being about 11.75″ wide. Perfect! Now I was ready to cut wood. I got my hands on some 1×5″ fir, which actually measures 1×4.5″. I’ll trim off .25″ inches and then cut to length, and then cut my angles.

20150926_141451

20150926_151717

Nice! Now I have something tangible!

Now, I know the width is 11.75″. So factoring in for routing on each side, I figured the pieces on either side of the keyboard needed to be about 11 inches. I felt pretty confident they were 2x4s, and I had some 2×4 cedar around. Now I had this:

20150926_153026

Cool! But… it doesn’t look quite right. The 2x4s just seem a bit too big. Looking at how much space is left in the back, there doesn’t seem like there’s enough room to fit the TVT boards in there. Drat. Something’s wrong. I need to figure out how big those front and middle pieces are.

Aha! An idea. A trick I learned from using air photos and a large format printer to very precisely model a segment of railway.

First, I took the above shot the museum had. It appeared to be more or less directly over top. I then print screened and copied into Photoshop, cropped it, and then set the image size to the known length and width dimensions. From there, I printed it at that exact size to my Designjet 450c. Voila!

20150926_164810

A (roughly) life size printout of the TVT prototype.

I then took out my MDS keyboard and checked the measurements. The keytops are .75″x.75″. Sure enough, my large photo also measured the same. Great. I can trust it. Now I measured those two wood pieces. They come out at just over 1″. So I’m guessing these are 1″x4″, not 2″x4″. That’s going to necessitate a trip back to Home Depot, but at least I know now before I go and router some rather expensive wood ($28.00 for 8 feet).

20150926_170005

More progress tomorrow!