Tag Archives: don lancaster

RF modulator build and test

After successfully testing the power supply on my TVT’s mainboard, I now need to move on to actually making the TVT do something.  The next instructions from the guide have me building the RF modulator, which is the component that gives the TV Typewriter its name.  As mentioned before, I have had advice not to bother with this and just go with straight composite video.  However, I felt doing this, while easier, would sidestep the main achievement of the device, which was that it used something everyone had in their home (a TV) already.

Anyway, building the RF modulator is straightforward enough – the trickiest bit is winding the 14ga tinned bus wire into a coil that helps generate a signal for the TV to pick up.  You can see the result behind the trimmer capacitor.  The wire is very thick and does not want to twist into a coil easily.  It requires the use of a 3/8″ drill bit and very strong hands.  There is a point on the second coil from the right where a tiny piece of wire ‘taps’ (is soldered) into it to draw the signal.

After that, I install all the required pieces, including my 300ohm twinlead with ‘gimmick attenuator’.  If you’re of a certain age, you’ll remember that twinlead is basically a flat cable with two wires embedded in either side.  In the days before TVs had coax inputs, you had screw terminals on the back for VHF and UHF reception.  These twinlead cables went between there and the antenna attached to your house.   The ‘gimmick’ attenuator is nothing more than a short piece of twinlead that is cut and electrical taped a few inches up from where the twinlead is soldered into the mainframe board.  This provides a kind of capacitor that further hones the signal going to the TV, so as not to overpower it.

With everything installed, we are ready to do a test!  The goal of this part of construction is to find a channel between 2 and 5 we can tune the device to.  At this stage, we are trying to tune it so that the screen goes completely blank, and any static noise is minimized or eliminated.   Ultimately I found the best channel to tune to was 5.  Here is a video of my dialing it in (please ignore my messy shop!):

You can see how the screen goes to a solid blue.  Offscreen, I am using a small screwdriver to achieve that screen by adjusting the 33pf trimmer capacitor next to the coil.  Anyway, there you have it!  Another exciting first step towards a working TVT!

Assembling the case

The final pieces are on order and now I’m moving to assembling the TVT’s case.  For trial purposes, I built the first case out of fir and plywood.  I had initially thought plywood was correct – in pictures of the original it appeared to be plywood peeking out from the left side where the vinyl covering had peeled away.  I was wrong.  It was metal.  And indeed, if I’d read the manual all the way though I would have known that all along – the article strongly recommended using a metal case to prevent RF interference from the device, which is a no-no!

For the side pieces I went with red oak.  I kept looking at photos of the original and that seemed to be, in terms of grain, as close as I could get.   I even got a bit cheeky and made sure the right side had a knothole, similar to the prototype!

I then routered out the insides.  This wasn’t quite as hard the second time around, although I did accidentally cut a notch through the front of the right side.  Argh!  I’ll have to figure out a way to fix that.

The case is assembled with a piece of 5/4″ wood (yes, that’s what it’s called) for the front ‘palmrest’, and then the back is a piece of sheet metal from Canadian Tire.  One thing I learned for bending metal – you need a brake!  I thought it’d be easy enough to fashion the metal into a box shape with nice crisp corners.  Nope!  Turns out metal really resists corner shapes!  No matter how much force I put on it, I couldn’t do it.  I ended up using a ballpeen hammer and an edge to bang it more or less into shape.  Thank goodness it’ll be covered with vinyl!

 

Looking alright.  To hold it all together, I ended up using No More Nails.  That stuff is quite strong and was able to hold the metal in place.  To bond the vinyl to the metal top, I initially used the same but as you can see in the pictures, it developed serious wrinkles.  So, I ripped it off (I had lots of vinyl; a square yard can cover a lot of TVTs!), sanded down, and did it with contact cement.  Much better, no wrinkles!

The really tricky part of assembly was getting the keyboard to fit – when I first put it together, the return and space keys were binding on the case.  I had to router things out a bit more (hence the accidental notch).  I realized later my case dimensions were out by about 1/4″ of an inch.  Not noticeable visually, but enough to cause these sorts of assembly headaches.

Anyway, after the glue has set, we have a pretty good facsimile of the original.  Check it out – mine is up top (obviously) and the original is pictured at the museum below:

Not bad if I do say so myself!  So yeah, basically I need to find a way to fix that accidental notch, because I am not messing with the router again.  And then I need to get the legends and nameplate made up.  Probably Front Panel Express will be where I go to do that.  But for now, on to getting the main unit actually working!

The RF modulator

Probably the most important circuit on this whole device is its RF modulator.  The RF modulator takes is essentially a ‘converter’ that takes a pure video signal and emits it at a frequency a television receiver can pick up.   As has been explained to me by folks more knowledgeable than I, the RF modulator is essentially a  Hartley LC circuit.

In the TV Typewriter, a coil of solid wire is fed current to produce oscillations at a frequency your television set can intercept and interpret.  There is a 33pf trimmer capacitor installed alongside that allows you to precisely trim the signal.  For example, if I wanted the system to use channel 2, I’d need to adjust until I was putting out signal at 55.25mhz.

Setting out on the project, I had been advised by both contemporary and historical sources not to bother with the modulator and instead just pipe video direct to composite monitor to make it easy on myself.

However to me, the whole point of a ‘TV Typewriter’ is its interface with a TV.  Sure, you could have bought a monitor back in the day, but monitors were expensive.  The genius behind the TV Typewriter was the realization that most people already had a perfectly good ‘monitor’ in their home – their TV!  To build the TVT without a TV interface just seemed pointless.  After all, I already have a TVT-2 that does direct composite output.

Probably I will come to regret this decision later.  I did run into some snags trying to make my coil.  The article advised I needed to take a 4″ piece of solid 14 gauge wire, and twist it 6 turns on a 3/8″ mandrill, spacing the turns so that the whole coil is about 1″ long.  Well, here’s how many twists I was able to get with 4″:

Hmm.  I decided to write Don directly about this.  This wouldn’t be the first error encountered in the directions.  I asked him if I could go longer and what the implications would be.  He said no problem (‘shouldn’t be critical’ were his words) and gave me some pointers on what I was shooting for frequency wise.

I ended up taking a 7″ long piece of wire and got my 6 turns out of that.  With that all shaped up properly, I then turned my attention to how to fix it to the mainframe.  Obviously I can’t direct solder the ends as they have to pass through to the copper traces on the opposite side of where the coil is mounted.  I couldn’t see clearly enough in any pictures of originals I had if they had used special mounts, so for the ends I just soldered in two solid wire ‘posts’ and then soldered the coil ends to that.

That seemed straightforward enough, but I was at a loss as to what to do about a third connection implied by the drill pattern on the PCB:

The instructions, being written for people who actually knew what they were doing, did not explain.  Was that another mount point?  If so, why?

This is where it helps to read schematics.  The schematic is specific that you must ‘tap’ (attach) a third point to the coil on the first turn:

Why?  Not being an electronics expert (yet), I can only hazard a guess that current is passed from one end of the coil to the other, and the third ‘tap’ is placed at a point where sufficient signal/oscillation is generated to be usable.

The trace the third tap is on leads directly to the 300ohm twinlead, which is a flat ribbon-like cable those of us over a certain age might call ‘antenna cable’.  You can see it on Roy’s unit here – it’s long and flat and has two little ‘Y’ connectors that screw into the TV’s VHF antenna terminals at the back:

 

This is more or less the TVT’s antenna, carrying signal from the unit to the TV to intercept at your chosen channel.  As I said before, a lot of TVT builders ran aground on this particular rock.  Bob Rethemeyer ran into this particular problem when he built his – the output was apparently terrible.  Apparently it requires a lot of fidgeting to get working properly, and that’s why so many chose to just direct the feed to composite monitors or gave up entirely.

Anyway, here’s the circuit more or less completely installed:

Personally I think RF output will be the least of my problems getting this thing to actually run.  But, we shall see!

Mainframe and Cursor Board Progress

Progress continues on my TV Typewriter Redux project.  After encountering disaster in the form of weak traces, I cleaned up and readied my second TVT mainframe board (there were three, thankfully I got one right!).  It looks great – the dye isn’t quite as blue as on the last one and it looks about as vintage as I think one can make modern PCB stock look.

I don’t yet have the transformer, fuse clips, or self test posts.  I’m also missing two momentary return DPDT switches.  But I’ve installed pretty much everything else.  The mainframe is easy to work on – everything is spaced out nicely.

Moving up the unit I started work on the Cursor board.  I had run out of old 14 pin sockets from my parts bin, so I hunted ebay and found these blue Cambion wire wrap sockets.  I don’t know what year they’re from but to me they look a lot like the ones my new friend Roy used on his TVT:

You can see Roy installed them only for certain ICs.  I suspect this was a decision based on what was likely to fail, and what components were too expensive to risk when doing self-tests and/or diagnostics.  But where he used sockets, they are those nice, tall blue ones.  I don’t know if they’re wirewrap — doubt it — but I can just install my wirewraps and trim the legs and get the same look.   Here they are installed with ICs in place.

I didn’t exist yet in 1973 but I think that looks pretty authentic if I do say so myself!

You’ll note I also did a test fitment, plugging the Cursor into the Timing board and then the Mainframe underneath.  It had occurred to me that if the Mainframe scan was skewed there was every likelihood the others were as well.  But I had hoped because I made sure to compare the three smaller boards before etching them that they would work.

I can’t say they fit perfectly, but they can be made to fit, and continuity is perfect between boards.  I can easily trace connections from a pin up top to where it ends up on the mainframe.  Excellent!  One small beef – although the first board plugs nicely and snug into the mainboard, the others sit up about a quarter inch from the molex connectors beneath.  I figured out this was because I had soldered the pins in such a way that they were actually pressing against the bottom of the molex connectors, restricting how far the pins from above could penetrate.  I’m not sure if I’ll attempt to remedy this at all, or just trim the pins on each board to compensate.  Hmm…

Oh no!

So I’ve got my TV Typewriter ‘mainframe’ parts install underway.  I’ve got molex connectors, caps, diodes, resistors — the whole nine yards.  I decide to put the board up against a light to check for any trace ‘bleeding’ or accidental solder bridges and then:

Not good.   Not good at all!

This is something any homebrew PCB maker should check before proceeding.  I had been fooled — from topside, the copper obscures the tiny pinholes visible here.  Since these pinholes look like the pattern of pixels in a magazine photograph, I’m guessing not enough toner was deposited to fully protect them from the etchant.

Interestingly, the traces all test good on the ohmmeter.  I’m tempted to carry on – but this state of affairs really bothers me.  More experienced hands warn it will eventually fail.  So I decide to switch to the second mainframe board I made — that one passes the backlight test handily.  I’ll wipe the black silksreening off it, clean it up, check it, and transfer everything over.  Thankfully the mainframe doesn’t have too many parts installed!

‘Dye-ing’ to Begin My Build

For several months, not much progress has happened on the TVT.  An attention-challenged collector, I’m always being tempted by other things.  I ended up making a few more acquisitions in the last few months, including a second SWTPC 6800 system, a Netronics video terminal (and homebrew Z80 computer), a Jupiter Ace.. oh, and did I mention I scored some original Mark-8 boards?  🙂 I also got wind of another possible project – building a replica of the original SOL Terminal prototype.  I actually discovered I had the proper keyboard to do it!   But I digress.

Part of my problem with getting going on the TVT was knowing where to start.  My logical brain said: go with the mainboard.  But I didn’t have the transformers, nor some other odds and ends, and I kept delaying.  Eventually I realized time was passing quickly, and maybe I could just build what I could and get going, and then fill in the blanks later?

So that’s what I decided to do!  I decided I’d not care what order I went, I’d just make sure I was building the TVT every day I could, a bit at a time.  Whatever I could do, I would do.  That’s the only way to get there.  Otherwise, procrastination leads to a dead project.

One detour before I built though, I decided I wasn’t quite satisfied with the colour of my boards.  They just don’t look vintage enough in my eyes, up close.  The old process of making PCB substrate (the fibreglass/plastic the copper traces are attached to) produced something that was somewhat fluorescent.  That process was problematic and is no longer in use.  Today’s boards tend to be either brown, or yellow, or a very light yellow green.  I chose the latter since that was as close as I could get to the kind of vintage green you see on originals.  Even now, I’m still not wild about it.  I’ve had some crazy thoughts, like getting a piece of original substrate and having it replated with copper and then I could etch it.  But, finding original substrate is almost impossible and the cost to do all that would be exorbitant.  How about dyeing?

Yes, apparently you can dye PCBs.  It’s simple.  You get some Rit fabric dye, a pot your spouse won’t mind you destroying, and throw it in some hot water and boil your PCB until it attains the color you desire.

I decided, looking at my boards, they weren’t green enough.  They were more on the yellow side, so, drawing on my experiences as an art student in the ancient past, thought I’d try using blue dye to green it more.  Royal Blue, to be exact.

First, a comparison of colour.  On the left, my new TVT mainboard.  On the right, an original Mark-8 CPU board.  You can see the difference right away — the bright fluorescent hue.  This is likely impossible to reproduce with dye, but let’s see how close we can get.

The instructions say you mix up the dye separately in a cup or so of hot water, and then add to your hot pot.  I guess this is to prevent clumping of the dye powder.

Next you need to build an apparatus to allow the board to be suspended in the pot.  I don’t have any photos of my mainboard being dipped, but I do have photos of a Mark-8 reproduction board I was working on simultaneously:

I had the oven on high for about an hour.  I didn’t notice any difference on inspection.  I then left it simmering, for an hour, then two, then turned it off and left it overnight.  I still wasn’t noticing a difference, so I decided to add more dye.  Within an hour, there was a striking difference.  Here’s two of my repro Mark-8 boards side by side just to give you an idea.  One is undyed.

Quite a difference with the dye, huh?  And here’s a comparison shot of the dyed Mark-8 board against that same board and an original SWTPC GT-6144 graphics board:

Unfortunately, not the correct color for the Mark-8 boards produced by Techniques, but could be legit as a totally home brew.  And close to SWTPC, which provided the original TVT boards in the kit Radio Electronics offered.

I dyed my TVT boards with this stronger batch of dye.  I ‘blue’ it!

Oops.

Definitely a little heavy on the blue.  I can live with it, but a note for next time: keep the amount of dye relative to water light.  My SWTPC boards went into the dye after I had doubled the blue, and so consequently the blue sunk in much more quickly.  Further, the manufacturer of the boards I used for my TVT is a different one than my Mark-8.  Subtle differences in the substrate result in different colors.  That said, some of the boards I put in after the mainboard benefitted from that lesson, and actually came out pretty close to original colour.  Check out the Timing board here vs. an original P197 SWTPC power supply board:

I mean, that’s kind of as close as I think you’re going to get with new substrate.  Gotta be happy with that.

Lesson learned: more dye = less soaking time, less dye = more.

On to the build!

TV Typewriter Boards – Cutting

With my first test board behind me, and applying everything I learned previously, I embarked on creating the first full set of boards I needed to start building my TV Typewriter.  That very first board was encouraging, however it ended up having many damaged traces because I didn’t have the vinegar/peroxide/salt etching method dialed in and left it long enough for the etchant to start getting under the toner.

After a few evenings of toner transfers and etching, at last I had a set of etched boards!  Now I just had to cut them to the correct size.  To be frank, I hadn’t thought much about cutting, figuring of all the technical hurdles, that one ought to be easy.  Seasoned PCB makers will tell you to use something like a hacksaw.  I tried that but found this FR4 stuff to be super difficult to cut through and keep from flexing while making cuts.  My next attempt was with a jigsaw and special fibreglass blades.  That cut the stuff like butter, but unfortunately jigsaws scratch the heck out of the boards and they’re very difficult to keep on a straight, accurate line.  In fact, my first TVT ‘mainframe’ board suffered an accidental detour into one of the traces.  Ugh!  Luckily I had more than enough board available to make another one, and I had thought about it anyway as I was still getting the etching process down pat when it was done and again, there had been some intrusion of the etchant onto the protected copper.

On the smaller boards, I went with tin snips.  Those were quite effective and accurate, but they had two problems: one, they tended to break the board apart internally, leaving little whitish splotches along the edge.  Also, these boards were fairly large, and the snips would only cut an inch or so before it became impossible to operate them.  Plus, the force required to cut would give a healthy 20 year old instant arthritis!

In the end, a trip to our local Rona store proved to be most helpful.  The gentleman there introduced me to a ‘laminate scoring tool’:

20160710_141625

This little guy has a hard carbide tip.  You use a ruler as a guide and then press and drag the tool tip along it, scoring a groove into the board.  After a couple of those you can forgo the ruler and just keep carefully pulling the tool across.  Eventually it digs out enough of a ‘trench’ that you just bend and snap the excess material off.  Woohoo!  Nice, clean, almost professional looking boards!

20160619_135421

Obviously we have to do a lot of drilling, and for that I got myself a dremel press, which I hooked up my Dremel 4000 to.  That’ll be another story for another time but essentially I’m seeing how difficult it is to keep the bit on target (and not break it in the process).

With the boards all cut out, I decided I did want the parts placement ‘silkscreen’ on the back.  Although I’m pretty sure the prototype didn’t have this, from a practical perspective, given my limited experience with homebrew electronics, I figured this would be really useful.  As mentioned previously, the instruction documents included these, which apart from some ‘shadow’ traces to show the builder where traces on the other side were located, were identical to the silkscreen treatment SWTPC did on their own boards.

I decided I’d make use of Photoshop and the Threshold tool again to make those shadow traces disappear.  Here’s before using the Threshold tool (note the light gray representations of the traces on the other side):

photoshopdemo5

And after:

photoshopdemo4

The Threshold tool can be a bit tricky here — some of the ‘shadow’ traces are really difficult to eliminate without affecting the ‘silkscreen’ stuff I want.  Some of the resulting silkscreen stuff will be a little bit less than ideal quality wise.

Then I just had to mirror the image, print it on my Hotrod paper, and after careful alignment to the front side of the board, iron it on.  Turns out ironing toner directly to the fibreglass of the PCB is even easier than to copper.

After repeating this process for all of the boards, at last, I have a complete set of TVT boards ready to begin building!

20160626_094712

20160626_094824

20160626_094850

I should mention that fifth, smallish board on the bottom right — that is Don Lancaster’s original ASCII encoder.  I believe this is the one that was used in his prototype.  Shortly after the article on the TVT was published, Don designed a second, simpler one.  I chose the older design just to be on the safe side.  It’s quite possible he might have used a prototype of the new design in the prototype TVT, but I’m thinking it unlikely.  Interestingly, the old encoder is exactly the same depth as my MDS/Microswitch keyboard.

Oh, and remember I said photos distort things like colour?  Check this out.  I’ve put a photo of my new ‘mainframe’ board side by side with a photo of the old.   The original is to the left, my new is to the right.

tvtorigvsnew

A good reminder, again, that photos can kind of trick you.  My boards don’t look quite that green in person, but then, maybe the originals don’t either?

By the way, for my ‘silkscreen’ layer, I don’t actually have to leave it black — apparently I can get something called ‘white toner reactive foil’.  Basically you iron/laminate it to your ‘silkscreen’ toner, and it leaves a white residue, instantly converting the silkscreen from black to white!  I’ll have to get some of that and try it out.  The black looks nice and crisp, but white would just bring it that much closer to ‘originalish’ look.

Anyway, now the hard part begins.  I need to secure the remainder of the parts I need (resistors, caps, molex connectors, etc).  And I need to get drilling, which I know now is going to be pretty tricky.  Never mind turning it all on and trying to get it working!  But at least now I have the main bits — and this whole process has inspired me to keep at it.

Making PCBs IV – The Home Stretch

After a long mental debate I decided to forgo the muriatic acid/peroxide homemade etchant route.  Watching videos and considering all the precautions and risks involved, it seemed to me, given the nature of my clumsy self, that this would just be a recipe for trouble.  So I decided against that.  I went instead and ordered a bottle of ferric chloride.  Ferric isn’t exactly cheap – $30ish USD, and it aren’t necessarily a whole lot safer than the acid route, although reading along it appears to be somewhat.

But whilst doing reading one day on the subject, I discovered another method.  This one involved using household cleaning vinegar (stronger stuff than regular table vinegar), peroxide, and salt.  I’m not at all certain how it works, but in the few videos of the process I could find I did discover that it does in fact work.  So I figure, what the hell?  I head to the pharmacy and pick up a bunch of peroxide bottles and then some cleaning vinegar at the grocery store.  Salt, obviously, I already have.  Sweet!  Let’s see what happens!

The instructions, such as they are, suggest a 50/50 ratio.  So that’s what I do.  I get a nice plastic tub, and pour in equal parts vinegar and peroxide.  They say to always add acid to whatever you’re mixing it with, not the other way around.  And although none of these components seem particularly dangerous, I have no idea what I’m actually concocting here.  So I’m doing this out in my well-ventilated garage, I’ve got safety glasses on, full rubber gloves — the whole nine yards.  And I’m pouring like the bartender in Drip Along Daffy — making cobra fang juice/hydrogen bitters/old panther shots for Nasty Canasta.

On first blush, there’s nothing to report.  Apart from a strong vinegar smell, there’s nothing exciting or scary going on.  So I drop my board in.

Nothing’s happening.

Now, I’ve been warned this method tends to be slow.  Basically from what I understand, we’re creating a weak hydrochloric acid here.  After five minutes, the copper is starting to change colour a bit, but it’s clear we’re not getting anywhere significant.  So I go to the instructions again and remember the salt, which they theorize ‘accelerates’ the reaction.  Ahh…. now we have action:

20160618_145928

I keep checking on it every few minutes, periodically adding salt.  At some point in the process, I don’t know how or why, the etching becomes much more aggressive — there’s fizzing, clouds of stuff coming up from the board.  And I’m seeing, finally, some of the PCB being exposed.

20160618_165549

The liquid in the tub is turning a blue color — I think this is copper chloride?  It’s been a long time since Grade 12 chemistry.  Every once in a while the reaction slows and I find I have to add more salt to the remaining exposed copper to keep it going.  Eventually though adding salt proves pointless.  It looks like we got most of the copper etched off anyway:

20160618_170642

The green color looks fantastic and I’m thinking maybe my Chinese FR4 were the right call — but of course the color is being accentuated considerably by the blue etchant.  I also learn by accident that this stuff isn’t all that harmful — although the powerful vinegar smell is a bit off-putting, splashing a bit accidentally onto my skin produces no burns.  I wouldn’t want to try a skin soak with it or anything, but hey.. nice to know we’re not talking about instant chemical death.

Comments: this method is slow and requires constant intervention.  And you need a lot of it.  On one board I tried to reduce the total amount of ‘etchant’ I made.. it wasn’t enough.  You need like two of the 500ml peroxide bottles and 1L of vinegar and a whole bunch of salt to do these big 200mmx300mm boards.  And it’s not cheap — peroxide retails anywhere from $2.00 – $5.99 for a 500ml bottle.  Yikes.  But, it has the advantage of being readily available and far safer than that nasty muriatic acid.

Back to color – the resulting board it turns out is fairly green, which is great.  It’s got a bit more of a brown twinge to it than what I was expecting for a vintage-look PCB.   Putting up against a vintage Digital Group board shows the difference:

20160710_141122

I did end up also etching a small piece of that sample board American Micro was kind enough to send.  And I’m glad I did, because it was way more brown coloured than my Chinese FR4 stock.  It would not have looked right at all.  So that’s a $400 mistake missed.  Kudos to American Micro though for their service and being willing to provide a sample in the first place.  If it weren’t for the color, I would have happily spent the money.

 

Making New PCBs III

So now I’ve got my TV Typewriter ‘artwork’ all finalized and I’m ready to try a transfer.  I don’t have any special toner transfer paper — but I do have magazine stock.  My first attempt is using pages cut out from a Macleans magazine.  The first couple runs are surprisingly easy – the printer takes the sheets and passes them through without fuss.  With one sheet done, I decided to take a run at it with the iron and see how the transfer worked.  And I had to use iron because I didn’t have a laminator handy — there were none available at any of the stores in town and I was just getting too impatient to get going.  Plus, why spend money if you don’t have to, right?

I put the iron on the hottest setting and pressed and ironed for what felt like an eternity but in reality was only about 10 minutes.  My arms ached.

20160619_212637

I then took the hot PCB (with oven mitts, eventually, after discovering by accident how hot it really was), rinsed it in my sink and peeled off the paper.  Damn.  Almost.. but still some traces missing.

20160529_190842

I spent the rest of the evening trying various things without success.  Longer ironing time, more pressure.  Started to wonder about the toner I was using — like many, I skimp a bit on that and use remanufactured cartridges.  I’ve heard these sometimes have less plastic in them (the stuff that melts), and the result is they don’t transfer as well.  But before I go and buy a $200 OEM toner cartridge, I figure maybe I’ll try everything else I can, including changing papers.   It’s possible this isn’t going to work — most of the toner transfers via iron I see demonstrated on youtube involve much smaller PCBs.. far easier to apply consistent pressure and heat.  But I’m not giving up yet.

I get out some acetone and wipe my test board down.  Then I try parchment:

20160604_191558

You’d think parchment would a shoo-in — it’s slippery stuff.  I thought for sure once heat was applied the toner would just jump right off.  Nope.  All it did was smudge a little.  Argh!

As a last ditch resort, I figured I’d try higher quality paper.  By this point I had already ordered Press N Peel, but that’d be a week or more of waiting and I just couldn’t wait that long.  I grabbed a Hotrod Magazine from my pile and noticed the paper was just a bit thicker and shinier than the Macleans stock.  I gave it a try.  I discovered right away though that I had a problem – the traces were not printing at full darkness.  They would on regular paper, but not this Hotrod stock.  After fooling around for an hour, I realized my issue was that I wasn’t giving the printer the proper setting for printing to glossy paper.  What you needed to do was actually tell it you were printing to heavy glossy in Photoshop’s print settings dialog.  Doing that caused the printer to request the sheets via the manual feed tray (I had just been slipping them into the regular paper bin, causing the odd error/jam).  I noticed right away a big difference — set to heavy glossy, the printer pulls the paper through more slowly, reducing the risk of crease and I guess giving more time for the toner to adhere to the paper properly.  The result is fantastic, especially for the really big traces on the motherboard:

20160710_130045

In reading online, I also discovered that a lot of people actually fully immersed their transfers in water right after they finished ironing.  So I filled up a large corn roasting pot.  And after 20 minutes of mad ironing, dunked my ‘transferred’ PCB:

20160619_213041

And voila!

20160612_172825

Now that I’ve got my ‘technique’ sort of down, I can make more of the boards.  With 5 pieces of 200x300m PCB, I’ll make two mainboards, and then on the remaining 3 pieces make 2 each of the cursor, timing and memory boards.

A few days and many sore arms later, here we have it:

20160618_120038

Okay.. now I’m getting impatient.  I really want to etch.  Should I just risk it with the muriatic acid?  Hmmm…. 🙂

Making PCBs II

Have been at a bit of a standstill on etchant.  They don’t sell the stuff anywhere in my town at all, so I have to order online.  It is apparently possible to mix your own using muriatic acid and peroxide; however, remembering my experiences with that nasty  acid cleaning our old concrete pool, I wasn’t eager to risk burns and who knew what else messing with that stuff.

Anyway, let’s focus on what we can do – the latest boards from China finally arrived.  They don’t *look* green to me, but I realize quickly the copper on the one side is distorting the color.  I decide I might as well proceed to the ‘design’ stage and start figuring out how to create the patterns I’ll need to etch with these.

I did consult pretty widely.  Most suggested using PCB cad software to recreate the boards.  I had the original plans via Michael Holley’s SWTPC site, but the low-res scans he’d made available weren’t that great and I was told I’d have problems with broken traces, etc as a result.  Still, I really wanted to use the original artwork.   I mean, yeah, going the redesign route would allow me to produce better quality boards and fix known issues with the originals in advance; but to me that meant going away from the whole purpose of the project.  There was something about using original artwork that just made the thing feel more authentic.

Here’s a sample of the artwork:

http://www.swtpc.com/mholley/TV_Typewriter/page_22.pdf

You can see the problem right away.. the scan produced fades on some of the traces.  That’s going to make it difficult or impossible to avoid etching those away.  But, this is 2016, and we do have this thing called Photoshop.  So let’s see what we can do with these.

First, we need to convert them from PDF into a format Photoshop can use, like JPG.  I downloaded the individual files from SWTPC.com and then used this free online pdf to jpg converter, set at 300dpi to do the conversion.  I then downloaded the resulting jpg files and opened them in Photoshop.  I started with the timing board.  First order of business, cut out the parts placement page.

photoshopdemo1

Next thing we have to do is get it scaled correctly.  There was no scale offered; the magazine of course wasn’t accounting for vintage computer enthusiasts in 2016 downloading these in scanned PDF format.  They assumed you’d have the originals, in which case they’d be at their ‘full’, correct size.

I started by printing as is, placing my ICs at various locations to see if the pins fit, and then using Photoshop’s Image Size tool to increase the size until things fit, making sure to keep proportions the same.  Eventually I got a width of about 8.247 inches and a length just a hair under 11 inches.  Keeping that info in my back pocket, I now knew what I had to size all the other artwork plans to.

The next thing we need to do is ‘fix’ the artwork.  First, I stripped it of color, making it a pure grayscale image.  Through google, I learned of a tool called ‘threshold’ that allowed you to eliminate greys.

photoshopdemo2

Moving the slider to the right, to about 172, created an amazing improvement in the darkness of the traces.  Had to kind of play with it a bit more to get it as close to ‘just right’ as possible — too far in one direction and it eliminates all the drill holes, too far in the other and traces white out completely.

I had only ordered five of the green-looking Chinese FR4 boards, but I realized because of their size (200mm x 300mm), that was actually all I’d need to create two sets of boards.  Of course, this being my first time, I was probably going to fail miserably on the first attempts, so probably not.

Anyway, for the smaller boards, I decided to put two per sheet.

photoshopdemo3

Now, you’ll note I did one final adjustment — I flipped them horizontally.  This is because along the way, I decided I would use the modern ‘toner transfer method’ of PCB making.  Based on reading, it just seemed like the easiest way to get the job done – a splash of modern that nobody would care about, as long as the boards ended up looking correct.

Back in the day, you might have taken these plans, had them printed as-is onto transparencies, and then placed those transparencies over the pre-sensitized PCBs in order to expose them and remove the protection on the copper you didn’t want to keep.  Because you’re using a transparency, you can leave the orientation as is and it’ll be correct.  However, toner transfer is different.  You’re printing onto a medium (ie magazine paper), and then you’re placing that down on the board and trying to, via heat and pressure, force the toner to jump over to the copper on the board.  If you just print the patterns as is, the resulting pattern you iron onto the board will be backwards.  So you have to mirror image it before printing.

Now that I had that, the next step was what to use to print it.