Interesting HP-65 card reader failure mode

[John Garza (3665)]

OK, I put the old Teenix CPU in HP65 #1333A06439 (a known good unit).

Writing then reading a program still inserts the STO 4, same as a reading a prerecorded card.

However I did notice something unusual. Hitting a RTN in RUN mode, then switching to W/PRGM the mysterious STO 4 disappears.

Procedure is:
1) load card in RUN mode
2) immediately switch to W/PRGM see 33 04
3) hit SST and see the first instruction from the card
4) switch to RUN
5) hit RTN
6) switch to W/PRGM and see 00 00
7) hit SST and see the first instruction from the card

Looks like the 33 04 is briefly taking the place of the normal 00 00 top of memory marker.
If that's the case, then it's not actually in program memory.
Likely a harmless artifact.

As for the faulty machine, HP65 #1333A09948, it cannot read a card it records. I get flashing zeroes and NOPs in memory.


-J

[teenix]

Thanks for taking the time to investigate the problem John, it will provide another clue.

The faulty machine probably needs a scope to check the data and narrow down the problem.

Probably a non-event, but did the head wires go back in the right positions?

cheers

Tony

[John Garza (3665)]

Double checked the wires - they are correct. First thing I checked.
I do have a scope. I'll try poking around with it this weekend. I presume the RA and RB signals.


One other somewhat related question... is there a way to specify multiple COM ports in the Calcom.exe program? I have it paired with TEENIX65 and TEENIX67, but I see in Windows Device Manager each one has 2 ports associated with it (transmit and receive). Looks like it is trying to emulate a serial connection. It sets the ports automatically in Windows (10 Pro). But I only see a place to set a single port in Calcom.

And definitely thanks to you Tony for all your advice and guidance!

-J

[teenix]

(09-28-2024 02:58 AM)John Garza (3665) Wrote:  Double checked the wires - they are correct. First thing I checked.
I do have a scope. I'll try poking around with it this weekend. I presume the RA and RB signals.


One other somewhat related question... is there a way to specify multiple COM ports in the Calcom.exe program? I have it paired with TEENIX65 and TEENIX67, but I see in Windows Device Manager each one has 2 ports associated with it (transmit and receive). Looks like it is trying to emulate a serial connection. It sets the ports automatically in Windows (10 Pro). But I only see a place to set a single port in Calcom.

And definitely thanks to you Tony for all your advice and guidance!

-J

Yes, RA and RB. There are some 65 scope traces in my Classic Notes in the Card Reader section. You can also check out the WA and WB to see if they are correct when writing.

CalCom can only access one port at a time. It can be a pain, I have mutiple FTDI and Bluetooth modules used for initializing boards and testing. I just remember which is which and change the COM Port in CalCom when I want to use them. If a device has been enabled, CalCom doesn't know about it or what COM Port it is, it just tries to connect to the selected one on the front panel. I can probably make an editable drop down list that you can type a name in and associate it with a COM Port to make selection easier.

Eg
65_1333A06439 = COM 5
65_1333A09948 = COM 8

or similar

cheers

Tony

[teenix]

I made a code modification which appears to still work ok both in the WIN emulator and on the 65 CPU board that I have.

I uploaded v18 test hex file for re-flashing the 65 CPU board here if you can try it out. I'll leave it up for a few days.

cheers

Tony

[John Garza (3665)]

Thanks Tony, I downloaded it.

Regarding HP65 #1333A09948 (the faulty machine),
Both RA and RB are sawtooth patterns during a read. Not the pretty square waves they should be for digitized data.
...my heart sank...

-J

[teenix]

(09-29-2024 02:20 AM)John Garza (3665) Wrote:  Thanks Tony, I downloaded it.

Regarding HP65 #1333A09948 (the faulty machine),
Both RA and RB are sawtooth patterns during a read. Not the pretty square waves they should be for digitized data.
...my heart sank...

-J

Yep, something definitely wrong there :-(

What happens if you read with RA/RB disconnected from the CRC. ie isolate the 2 pins between the sense and CPU boards.

[John Garza (3665)]

Just a final update to close this out.

The problem was a faulty sens/amp chip. It exhibited triangular waves instead of square waves. I do know triangles can be converted to squares via a simple RC differentiator circuit. However, the timing of the pulses would have to be correct for that to work, which may or may not be the case depending on the extent of the damage to the chip. So instead of investigating that possibility, I just swapped in a complete sense/amp board from a 67 parts machine. Slightly different chip (different serial number), but the 65 takes it just fine.

One tip to make it easy on anyone else going this route - the 65 head wires are soldered to the board. Late 67's connect via little plugs. Instead of trying to tin the tips of the head wires perfectly straight, just solder a small section of a clipped component lead (diode, resistor, capacitor, etc) at a right angle to the head wire. You now have a little L shaped straight pin to insert into the plug on the sense/amp board with tweezers.


And thanks everyone for your help!
It's one of the things that makes this Forum great.

-John

[teenix]

Nice :-)

I don't know what differences there are with the sense chips. Maybe they were just improved versions.

cheers

Tony