PDA

View Full Version : Heathkit H89- TX Low Voltage



CommodoreKid
November 5th, 2014, 10:50 AM
Greetings! I wasn't sure where to post this, so here's my best guess

For some time now, I've been putting off the next stage in bringing my Heathkit H89 back up to working status. The folks over at the SEBHC were most instrumental at helping me revive it over a month or so. After several blown capacitors, a few torches traces, and a reset line being stuck, it finally whirred back to life. Now all I need is the OS disks, but before I mess with that, I want to try using the terminal properly.

However, I've reached an impasse that might require a more generalized retro computer approach: the terminal portion (identical to an H19/Z-19) has low TX voltage. I'm only getting about 2 volts from it, which isn't enough for any of my RS232 equipped gear to pick up. I replaced the line driver for the TX section, and checked all the voltages going to the driver, but no change occurred. Any ideas what the next course of action might be?

If for some reason this cannot be solved properly, I was wondering if it would be wise to use a device to step up the voltage to 5 volts after it leaves the H89. To me, this seems rather kludgy, but I have no problem doing something like that if I must. The idea of actually using a physical terminal is whimsical, and I wish to pursue it.

Thanks!

MikeS
November 5th, 2014, 11:21 AM
...the terminal portion (identical to an H19/Z-19) has low TX voltage. I'm only getting about 2 volts from it, which isn't enough for any of my RS232 equipped gear to pick up. I replaced the line driver for the TX section, and checked all the voltages going to the driver, but no change occurred.Not enough info;

RS-232 is bipolar, so +2V, -2V, or +/-2V?
What is the line driver chip?
Specifically, what are the voltages on each pin?
On the interface connector?
At rest and while transmitting?

The output voltage should not be 5V, but around +/- 12V (+/- 5V is pretty well the minimum).

If the driver IC is good, the supply voltages and input signals are good and nothing is loading the output it has to work.

CommodoreKid
November 5th, 2014, 02:45 PM
Sure, here's the rest of what I remember off-hand. The line driver is an MC1488, and it is powered by +12v on one side, and -12v on the other side. I got the same results with both the original part and a modern replacement.

The signal's leaving the TX line on the H89 are approximately +2v high, and about 0v low. This was measured on the output pin of the chip, and mid-way though the RS232 cabled on its way to my desktop machine. The +/-12v going to the driver have been measured, and match what they should be at. I have yet to measure the input voltages.

A few other things of note. When I fire up hyperterminal (when I've tried connecting to the H89 with Windows), the H89 does display what it has been sent (when I give it the correct protocol/text format). Still haven't figured out the right program & settings to achieve the same goal with Linux yet. The terminal portion is successfully communicating with the computer portion of the H89, which means that it is working, besides the lower than ideal voltage.

If there are any more necessary details about my specific unit, I will do my best to provide them.

kyeakel
November 5th, 2014, 03:20 PM
Schematics show a cap tied to pin 11 (serial data out) of the 1488, I'd check or just replace it.

Kipp

CommodoreKid
November 6th, 2014, 08:10 PM
Seems that my set of schematics omits capacitors on both the 1488 & 1489, which is probably why I didn't think to check them until now. I replaced them this evening, and performed yet another test connection: no dice. I also double checked the voltages to the 1488, and they are about 11.7 & -11.8 volts respectively.

Sadly, the oscilloscope that I was using to perform tests was on-loan, and I have since returned it.

But at least the capacitors were ruled out!

kyeakel
November 7th, 2014, 05:23 AM
Well there doesn't seem to be anything else on the output side of the circuit, I guess I'd isolate the output pin of the 1488 and see if it then swings + and - Also, I don't see that you have verified that the input signal to the 1488 toggles when you send characters? If you have a logic probe, you should see this pulsing when keys are pressed.

Kipp

CommodoreKid
November 7th, 2014, 06:31 AM
I'll be borrowing the oscilloscope again today, and running some additional tests with it after work. I want to see what's going on between the 8250 ACE and the 1488. I also realized that I had not watched the Tx/Rx between the computer & terminal portion.

Thanks for the suggestions -- this whole experience has been one of learning for yours truly. Considering that when I acquired it a few months back, I had fried plenty of stuff during the first couple of power-on tests, I'd say I've come along way. Here's a picture from a few weeks ago showing the unit during my first RAM test.

21379

g4ugm
November 7th, 2014, 08:56 AM
1488/89 are very prone to popping. Are you sure the modern part was an exact replacement not a MAX232 or something designed to offer 5v to rs232 without +/- 12v power rails. Also make sure you are using pin-8 as ground (signal ground) not pin-1 which is frame ground.

Does the H89 need a cross over cable?

CommodoreKid
November 7th, 2014, 09:20 AM
Prone to... popping?

I specifically ordered a pair of brand new MC1488's (not a Max232, 3232, or a cheap derivative/knockoff thereof). If you would like a picture to prove it, I will gladly oblige when I get home today. Since the H89 has demonstrated that it can receive data just fine, I have not replaced the 1489.

I did not think to test based on the signal ground, rather than the frame ground though -- I will use that for my tests this evening.

As for thee RS232 connection I'm currently using, it's basically just ground, Rx, & Tx. My initial cable that handled CTS, DTS, etc. but after it fell apart, I've just been using the bare minimum 3 pins.

g4ugm
November 7th, 2014, 10:59 AM
Prone to... popping?

I specifically ordered a pair of brand new MC1488's (not a Max232, 3232, or a cheap derivative/knockoff thereof). If you would like a picture to prove it, I will gladly oblige when I get home today. Since the H89 has demonstrated that it can receive data just fine, I have not replaced the 1489.

I did not think to test based on the signal ground, rather than the frame ground though -- I will use that for my tests this evening.

As for thee RS232 connection I'm currently using, it's basically just ground, Rx, & Tx. My initial cable that handled CTS, DTS, etc. but after it fell apart, I've just been using the bare minimum 3 pins.


OK new 1488 sounds fine.

Sorry to be a PIA but can I check that the pins are you using 2,3 and 7, not 1,2 and 3. Pin 1 is often not connected to anything...

CommodoreKid
November 7th, 2014, 10:11 PM
Actually, on the RS232 connector, I'm using pins 2, 3, & 5. 2 & 3 being the Tx & Rx lines, and 5 being the signal ground.

I found the problem though!
http://www.db9-pinout.com/db9-pinout/db9-pinout.gif

See that image? That was the reference image I used for the DB9 connector I jury-rigged. As a result, I made my connector with the right pinout... for the opposite gender connector. I've been looking at the DTR pin thinking it's TX this whole time! I was also using DCD as my ground -- boy do I feel stupid. :stern:

I have successfully sent data from the H89 to my modern desktop via terminal. AKA, problem solved! Moral of the story? Double check reference diagrams. Thank you to everyone who replied to give me a hand!

g4ugm
November 8th, 2014, 12:28 AM
Nope, lots of folks have done that. I was equally daft assuming you had 25 pin connectors which is where the 7/5 hash up came in..