Hazeltine 1500 faulty keyboard

[MeanDumpsterCat]

Hey folks, new guy to the forum here.

So I recently aqquired an old Hazeltine 1500 serial console from around 1977. Most of it works okay (screen is a little glitchy). It's main problem is that only a small number of keys on the keyboard work (1,3,5,7,^,TAB,reset).

I have tried re-soldering all of the keys and stilll made do difference and the main board has also been cleaned. It is set to half duplex mode so that it can display keyboard characters on the screen. As of the moment I do not have an RS232 cable available but I am pretty sure the connection is fine.

But like I said, main problem is the keyboard. If anyone has any suggestions on how to fix or how I could get a refurbished board, please let me know.

Thanks in advance.

I hope I enjoy my time here in the future!

[jay_oldstuff]

You may have to resort to removing each switch dismantling cleaning and re-assembling. There is an excellent video on You-tube by RetroGamerVX where he strips cleans and reassembles the keyboard from a late 70's terminal. Do you have a multimeter? try a continuity check on the pins on the back of each switch just to determine it is the switches at fault.

Jay

[MeanDumpsterCat]

Thanks, Jay.
I'll post some pics of the keyboard but under the keys I can see that it makes contact okay, I will try checking the connection with the multimeter whenever I can.

[AC/HL]

It probably works on a matrix principle, it may just be one of the lines not working.
As well as a photo of the layout, a complete list of either the working or non-working ones, whichever is easiest, may help.

[julie_m]

These keyboards almost invariably operate on a matrix principle. One "row" at a time is energised and the columns read. The switches often have two pins connected to each contact, to use as pass-throughs to simplify matrix design. Look out for dry joints and cracked print between key switches.

[MeanDumpsterCat]

Thanks. I have checked the joints on the bottom of the keyboard and they all seem to be fine. I'll check the switches too when I can.

[JohnBHanson]

Yes it is a matrix drive and all of 1,3,5,7,^,TAB are on the X1 line. The keyboard encoder activates each of the 9 X lines in turn and reads on the 10 Y lines. This is driven from an AY-5-3600 keyboard encoder (U60). (Page 145 and Page 146 B-24, B-25) of the service manual.
It will be worth checking that the pins on the keyboard encoder (pins 32..40 are X8..X0, pins 17..29 are Y0..Y9).

Pin 16 should pulse when a valid key is pressed.

(It is possible you also have some shorted lines which are stopping the keyboard working. With the power off check remove the AY-5-3600 (Assuming it is socketed). Check for each X pin check that it is not connected to any Y pin or ground.

Also check for shorts between X and GND and also between Y and GND.

Hope that helps.

[MeanDumpsterCat]

I just connected this to my PC via a straight through RS232 (db25 to db9) but I cannot get anything to appear on the terminal. I can set the baudrate and stop bits but that's about it. Anyone know how I could overcome this?

[G6Tanuki]

You will need a crossover-cable.

In the traditional days on 25-pin V.24 connectors we wired

1 to 1 [chassis earth]
2 to 3 [transmit to receive]
3 to 2 [receive to transmit]
7 to 7 [data earth]

Then we strapped 4 and 5 together at each end, [CTS/RTS]
and 6-8-20 together at each end [DSR/DCD/DTR].

[mike_newcomb]

Hi G6Tanuki, at one end (I would think the pc) he has a 9 pin socket.

I think your connections may need modifying to take this into account, unless Mean is already using a 25 to 9 pin adapter.

====

Hi Mean, you should be able to also set up:-

Data bits (6 or 8)
Parity (yes, no, odd, even)

Under dos the MODE command was used for this.

Are there any options to change settings (such as terminal emulation) on the Hazeltine?

Regards - Mike

[MeanDumpsterCat]

Im guessing I will need a db25 to db9 modem cable then. I have one but it has female on both ends.

[acollins22]

Hi,
It does sound llike a keyboard problem primarily but if you want to test comms here is what I'd try.

If you know the terminal is set to no handshaking or software handshaking, just set it to full duplex and connect pins 2 and 3 together on the terminal end and try pressing a few keys. You don;t need the PC. You should see what you typed appear on the screen as you've connected RX and TX together.

If there is hardware handshaking, connect 2 to 3, 4 to 5 and 6 to 8 to 20. This should fool the terminal that it's good to go. again you don't need the PC.

These two routes will check the terminal as far as the socket on the back. It's a handy test as you don't introduce too many unknowns such as the PC cables, comms software etc.


Cheers,

Andy.

[MeanDumpsterCat]

Thanks, Andy.
I am pretty sure the comms are okay since the photo's of the terminal in the auction showed a CP/M computer loaded up on the screen.

I just need the keyboard and caps sorted first

[mike_newcomb]

Hi Mean,

when RS232 appeared, there was no definite standard covering the connections, especially as at first upto 25 (+ possibly the screen) were available. (The 9 pin version became common later when it was realised 25 were often far more than required.)

More so, when handshaking etc. was used e.g. DSR/DCD etc.

Thus you cannot assume a modem cable will have those you require. I found it best to check which connections a cable has, by opening its plugs. If this is not possible, metering out may not be easy if any are cross wired.

To connect a dumb terminal, probably no more than 3 - 6 end to end connections are required.

I think G6Tanuki is incorrect advising connecting 1 to 1. We were told to connect the cable screen to pin 1 at one end only. As this is for safety, you can see why. Should one chassis become live, you do not want the other one to do so.

Regards - Mike

[MeanDumpsterCat]

I think what I will do is get a db25 / db25 modem cable with a 25pin to 9pin adapter on the PC's end. Is this a good plan of action.

The cable schematic in the manual matches that of a db25 modem cable if that is of any help.

[Dave Moll]

Just be careful to heed the advice given above about making sure that the needed signals are available on the appropriate pins, otherwise you may need to make up your own cable with the specific requirements of your setup.

[G6Tanuki]

Originally Posted by mike_newcomb

I think G6Tanuki is incorrect advising connecting 1 to 1. We were told to connect the cable screen to pin 1 at one end only. As this is for safety, you can see why. Should one chassis become live, you do not want the other one to do so.

From my days deploying rather a lot of V.24/RS232 stuff on production-lines, you *always* linked pin 1 to 1. And the cable-screen was linked device-to-device by connection to the [always metal] DB25-plug shells not to any of the wired pins. And you always kept these isolated from pin-7 [data-ground].

When you're controlling gear that can spot-weld the entire side of a Transit-van to its chassis in a second you need to be sure of your different grounds. The coming of EIA-485 balanced circuits made life a lot easier when provisioning large production-lines.

[RobinBirch]

Originally Posted by G6Tanuki

From my days deploying rather a lot of V.24/RS232 stuff on production-lines, you *always* linked pin 1 to 1. And the cable-screen was linked device-to-device by connection to the [always metal] DB25-plug shells not to any of the wired pins. And you always kept these isolated from pin-7 [data-ground].

Ditto. We normally wired 1 to 1 and 7 to 7. The screen went from Plug shell to Plug shell. The circuits were then designed to manage signal, power and safety earths separately. On a number of circuits we had to deliberately modify them to split pins 7 and 1 to prevent earth loops giving us noise all over the place.

Robin