• Please review our updated Terms and Rules here

SWTPC TVT2 Under Repair

Yeah I'm a bit puzzled myself. Used to Apple II keyboards.. simple switches. I'm still learning this stuff.. I'm trying to understand why if the strobe was jammed it was sometimes able to produce characters. That one time where it kept cranking out O's, it let me type other previously nob working keys. Then it just stopped. And then it only allowed N's, and now nothing except the odd random smattering of N's when I power up.
 
I can think of one reason why the keyboard would produce spurious characters--and it has nothing to do with the keyboard itself.

Have you checked your +5 supply for excess ripple? I can see, given the configuration of PNP transistor logic with RTL ICs where that could get greatly amplified. RTL has pretty low output current and really awful noise immunity.
 
Last edited:
I've not.. I think you need a scope for that right?

The power supply would be a good suspect.. it's putting out a really nice stable picture and the other pieces of the unit like the cursor control board work. The previous owner also wired up the user defined keys on the keyboard to do cursor functions and they do work as intended, but I notice they have jumper wires installed and may not have anything to do with the strobe?
 
Yes, a scope would be useful for this and other things. :)

The problem with trying to use a logic probe on the keyboard is that it's not your standard TTL levels. The encoder matrix uses diodes and PNP transistors and an R-C network to create the strobe and key encodings. The output buffers are RTL, not TTL (different logic levels and noise margin). Don Lancaster wanted to create the simplest encoder using commodity parts, even though RTL had sunsetted by the time he wrote the article. To make matters worse, we're driving TTL with RTL--and the outputs of a couple of those RTL inverters are wired-OR.

Maybe we should back-track a bit and see if the TVT itself is working correctly. Let's try by leaving the keyboard off. On the TVT ground keyboard inputs bit 6 and bit 7 (pins 11 and 12 on J9) and attach a test wire or probe to "KEYPRESSED"(pin10). Every time you touch the other end of the test probe to ground, you should get some "?" characters on the screen. If you ground pin 1 on J9 (bit 1) you should get ">" characters instead.
 
Peachy! So we've determined that the TVT itself works fine. We can set that to one side and concentrate on the keyboard without having it attached to the TVT, which is exactly what we want to do. Divide and conquer.

So, if you have the keyboard on its own power source without the TVT, you should be able to read each keycode on on the output as long as a key is depressed.

Thus, pressing ">", would yield lows on bits 6 and 7 and 1, with all the other outputs high. We'll tackle the KEYPRESSED strobe later. You ant to run through as many keys as you have and verify the output. e.g. Pressing the space bar (and leaving it depressed while you check) should give all outputs low, but for bit 6. With no keys pressed, I think all outputs should be high.

(One confusing thing here is that SWTPC numbers bits from 1, not 0, as would be normal practice.
 
Okay... I'll have to find a way to get a 5V PSU into it... it's soldered to a whole much of things on the control panel.

Interesting discovery while messing around with the keyboard still connected though.. remember I cut the keypress/strobe pin on IC2? If I bridge the broken pin back together randomly with my logic probe, it shoots out keys onto the screen.. mostly Ns but also other random characters.
 
Last edited:
Another note... randomly triggering the strobe with my logic probe produces some interesting results. Some keys produce the correct character - ie. if I hold down K, it produces J. If I hold down H or J, they produce the correct characters.

So I'm thinking we do have an issue with strobe, but also maybe with one or both RTLs not triggering the right code?
 
Further discovery. J and K are separated by a single bit, bit 1. Bit 1 according to the schematic runs through pin 14 -> pin 1 on IC1. For J, it should stay low. For K, it should go high. If I check IC1 while pressing K, it doesn't go high. In fact, there is nothing at all on Pin 14, and a low on Pin 1.
 
Well, let's look at the bitcodes for the characters you cited:

H: hex 48 or 100 1000
J: hex 4A or 100 1010
K: hex 4B or 100 1011

So the question is are we producing bit 1? Try pressing "G" (hex 47 100 0111) and see what you get.

But what do you mean "nothing at all on pin 1"? Grab your DMM and measure the voltage. Remember that I said that RTL uses different voltage thresholds than TTL, so your logic probe probably isn't giving the true picture.
 
G does not cause anything with Bit 1. It stays low. If I manually trigger the strobe while holding G, I get F.

Getting 2.4ish volts at Pin 14.
 
I should also note that when I trigger the strobe, it doesn't just produce a character, it always produces a space also.
 
So, pin 14 is always a logic "1". Check the value of R13 - it should measure about 4.7K (or measure pin 14 to ground; same connection).

Space doesn't really mean anything because the TVT doesn't fully decode all characters. With no keys pressed, what are the outputs of bit1 to bit7? I'd expect that bit1 would be low.

That being the case, we could have a stuck key or shorted diode in D3-D11. I'd check the diodes first. If no luck there, then it could be that IC1 is bad.
 
Last edited:
The bits at power up are as follows:

Bit 1 = 0
Bit 2 = 1
Bit 3 = 1
Bit 4 = 1
Bit 5 = 0
Bit 6 = 0
Bit 7 = 1

It's that way every time. Bit 1 never goes to 1 no matter what you press.

I can't see the resistor labelling and don't have a map of the correct PCB, but I get 2.5k ohm off that resistor, or pin 14. All the others are 4.7.
 
Well, according to my schematic, that's not right--it should be 4.7K. I'm thinking that you've got a bad IC1--you might want to pull it and install a socket in its place, so you can check the levels without the IC in place. Bit 7 is correct--it should be the inverse of bit 6 with no keys pressed. Bits 2 and 3, check--and bit 1 should be also, but it's not.
 
I figured since I have new ones coming anyway I'd just snip pin 14 for now (the shears I have are so fine I can actually put the leg back together. I then noticed that one of the user defined keys has a wire that is soldered in to the same trace (ground) that resistor is on, so I disconnected that too. But from one end of the resistor to the other, still 2.5ish.
 
I went digging through my parts bin and found (by my DMM) a 4.7k ohm resistor. I'm thinking I'll remove the suspect one and sub this in.. is there a particular direction these need to face (thought bands on them meant something). Anything else to be aware of like wattage, etc?
 
Back
Top