[GreenKeys] Re: TTL and Model 28 Keyboards

[Bob Camp]

Hi

I (obviously) have been giving this a little thought. Here is more or 
less what I'm headed towards.

Each keyboard or tape reader has it's own dry key loop. The RS-232 
trick can also be used with a make / break system rather than a polar 
system.

Each magnet (printer, tape punch, clutch magnet, what ever) has it's 
own solid state keyed loop.

You have an RS-232 i/o or two for computers.

All of this needs to be mashed together in a more or less orderly 
fashion. Take it all into a $4 PIC with enough i/o pins to handle the 
task. Put some real simple code in the PIC so it more or less acts as a 
gate array. Obviously you could use a $2 gate array for the same task.

Let's say you have a nice setup. Two printers, two typing reperf's, a 
set of three tape readers and a keyboard. Maybe it's all in a nice 
large ASR cabinet.

In the ASR cabinet:

You have four "data sources"
You have four "data destinations"
You have three clutches on the tape readers.

Outside the cabinet:

A TU on 30 meters
A TU for tuning around with
A computer with two com ports

That nets out to eight ins, eight outs, and a couple of clutches.

If they all are running at 100 wpm you have a bit over 1,000 things 
going on every second. If you want to keep the distortion low then you 
would need a software loop running at about 16X that. Let's call it 20 
KHz. If the PIC runs at 40 MHz then you get roughly 2,000 instructions 
in the loop. That should be enough to get an 8x8 routing table worked 
out. Probably not enough for speed conversion / buffering / ASCII to 
Baudot though.

Obviously you don't need to route just anything to just anything. There 
are a relatively small number of combinations that make any sense. Set 
them each up with a PC and then switch between them with something 
simple. Maybe add a few more dry key inputs to hook up to some 
switches. You certainly won't run out of pins.

If you are going to do this with surface mount assembly then a flash 
based gate array would do all the routing just fine. A little pic could 
then handle some fancy stuff.

Another alternative would be to put a small ($1) PIC on each magnet. 
Let the PIC run a switcher to come up with high loop voltage only when 
you need it. Run the magnets on 12 volts the rest of the time. That 
keeps things from getting very hot. A dual voltage +/- 12 volt supply 
at about an amp or so isn't all that hard to come up with. You only 
need a few ma off the negative side and all the rest off of the 
positive side.

Of course at some point you spend a fortune on this monster and there 
are only two people in the entire universe who can ever use it.

Here's another trick. Come up with some simple little cards that stack 
up. Stack them on a master board that has some basic smarts and power 
regulation stuff on it. With low power on all the little cards heat 
sinks are not an issue. Various people make stuff like 0.025 inch post 
connectors that are cheap enough to use for this sort of thing. If I 
need eight outs and two ins that's what I buy. They plug in on top of 
each other and I'm in business. That way anybody can match the gizmo to 
their setup.

Here would be the card line up:

1) Main board with the routing (PIC or what ever), computer RS-232, 
control switch inputs, and power regulation
2) 60 ma loop magnet driver card, each one with a simple switcher
3) Keyboard / contact input card
4) RS-232 i/o card

Just for fun let's say you have two stack points on the main board and 
allow an eight high stack on each location. That's enough for the setup 
I started out with. If you do cards with duals on them then a four high 
stack would be plenty. A dual probably makes more sense as keyboard / 
magnet combo or a dual RS-232 than as a dual magnet driver. The net 
effect would be the same either way.

Net result hopefully would be something each guy could configure to 
match their needs and that could handle just about anything.

Boy did you get me going .....

	Enjoy!

		Bob Camp
		KB8TQ


On Jul 5, 2004, at 7:19 PM, AI2Q wrote:

> Hhmmm...good food for thought Bob. I've used Maxim RS-232 chips 
> before, and
> they work very nicely for a variety of applications.
>
> Anyway, I've got two opto-couplers in "series" on my logic board (under
> development), and decided to use the Model 28's internal supply and a 
> 60-mA
> loop, as shown in the Model 28 manuals and diagrams.
>
> On my board, one opto is used to derive TTL from the Model 28, for 
> keying
> the rig (an Omni-V), and the other opto-isolator drives a HV xstr for 
> keying
> the loop from TTL derived in a switched-cap TU that I have here. I'm 
> picking
> off a logic signal from the tail-end of the filter/discriminator 
> portion of
> old Kantronics Universal terminal Unit (UTU). It's an 8039-based TU 
> with
> firmware in EPROM for copying Baudot, ASCII, CW (!), and AMTOR. Using 
> it I
> hope to copy to the glass tube (occasionally) as well as the Model 28 
> (the
> prime mover).
>
> I'd be happy to send anyone JPEG schematics of my board if interested.
>
> Vy 73, AI2Q, Alex in Kennebunk, Maine
> http://users.adelphia.net/~alexmm/ai2q.htm
>
>  .-.-.
>
> -----Original Message-----
> From: Bob Camp [mailto:ham at cq.nu]
> Sent: Monday, July 05, 2004 12:10 AM
> To: ai2q at adelphia.net
> Cc: 'Greenkeys maillist'
> Subject: Re: [GreenKeys] Model 28 Keyboards
>
>
> Hi
>
> What I was thinking about was using a very normal RS-232 transceiver
> chip to hook up the keyboard. Pick any one of the five volt supply /
> internal voltage multiplier parts. from Maxim or Intersil.  Use the
> outputs to drive a nice + and - voltage to bias up the contacts. Then
> use a RS-232 receiver to demodulate the output.  It should work just
> fine to give you a TTL level from a normal 28 or gold  plated 28
> keyboard. No dual supply required. Low current / moderate voltage
> switching just like the Teletype docs call out.
>
> If you have a dual supply available, say from a TU then bias up the
> contacts with +/- 12 or +/- 15 volts. I would still consider using a
> RS-232 receiver.
>
> The only problem with using the RS-232 receiver is that the distortion
> will be slightly higher than if you use a true zero point switch.
> Either way it still will have *way* less distortion than a dry keyed
> single supply setup. The dry keyed setup will always have less
> distortion than it's inductive loaded wet keyed cousin. Of course none
> of that counts ants and grime on the contacts as a source of
> distortion.
>
> 	Take Care!
>
> 		Bob Camp
> 		KB8TQ
>
>