UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Reply
 
Thread Tools
Old 4th Feb 2020, 10:06 pm   #1
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Commodore PET 3016

Hi, I wonder if anyone can help? I have a Commodore PET 3016. It boots up fine and will even load some games from a cassette tape! However, the keyboard behaves very oddly. Not all the keys work and it isn't a dirty keyboard. For example QWERTY comes out as 'blank blank RRYY'. The shift key makes no difference-some keys work, some don't and some return the same letters. When a game loads all the letters that should be there to make up a sentence are - so that's not an issue. For example, I can't type 'PRINT" - it comes out as 'PRINY' but it happily displays 'Print' if it is asked to by the program! Very strange.

I've changes the 6502, 6520s and 6522s. I've even changed the 2114 video RAM etc. nothing.

I live in the North East UK if there is someone kind enough to help? My electronic know-how only goes so far! I'm desperate to get this fantastic machine working again - help!!

Best wishes,
John
John Earland is offline   Reply With Quote
Old 4th Feb 2020, 10:57 pm   #2
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

While we await the arrival of the real Commodore experts, I wonder if you can post a table of which characters actually appear when each of the numeric and alphabetic keys are struck? This may reveal a pattern which could point to the underlying fault.

For example, it seems that when you hit 'T' you get 'Y', so,

What character appears when you hit 'G'

What character appears when you hit 'B'?

In case you haven't guessed, I am wondering if the keys in the column T-G-B are being read as Y-H-N, the keys in the next column to the right.

Last edited by SiriusHardware; 4th Feb 2020 at 11:03 pm.
SiriusHardware is online now   Reply With Quote
Old 4th Feb 2020, 11:07 pm   #3
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

This diagram here looks like it is for the PET 2001 but I wonder if the 3016 employs a similar keyboard scan regime? If so the 74LS145 (or whatever does the equivalent job in the 3016) may be the cause.

http://www.zimmers.net/anonftp/pub/c...N/320349-3.gif
SiriusHardware is online now   Reply With Quote
Old 4th Feb 2020, 11:27 pm   #4
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

Further, this page has some information about PET keyboards, but unfortunately not their circuit diagrams.

http://www.6502.org/users/andre/petindex/keyboards.html

In the section headed 'Keyboard Scan' it describes the key scanning operation as it would be in the diagram linked to in the previous post, and that does use a four to ten line decoder, ie, the 74LS145, to scan the columns. It's not clear whether the 'Business Keyboard' also employs a 4-to-10 line decoder to scan the columns as that is not explicitly mentioned in the following section, 'Business Keyboards'.
SiriusHardware is online now   Reply With Quote
Old 5th Feb 2020, 1:00 pm   #5
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Hi, thank you for the replies. I will post some pictures when I get home. It’s really frustrating because I don’t think it’s a million miles away from being fixed! But, what do I know!! Thanks for the suggestions. I’ll get some images when I get home tonight. Best wishes,
john
John Earland is offline   Reply With Quote
Old 5th Feb 2020, 6:10 pm   #6
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Hi folks.
This is an image from the PET. Underneath the READY remark is the return from pressing the top line starting with the @ sign (on the keyboard that is). The second line is the whole alphabet and other characters (not the number pad) starting with QWERTY etc.
Click image for larger version

Name:	D5C01FA0-A055-4F4A-952D-64CE2EEF61E6.jpg
Views:	98
Size:	46.4 KB
ID:	198474
John Earland is offline   Reply With Quote
Old 5th Feb 2020, 8:17 pm   #7
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

If I understand you correctly you produced the characters in the second line by typing letters Q-P followed by A-L and then Z-M.

If that's a yes, then here's a table for what you press and what you get, assuming I interpreted your input correctly.

Code:
Q=R	W=R	E=Y	R=Y	T=I	Y=I	U=P	I=P	O=^	P=<
A=F	S=F	D=H	F=H	G=K	H=K	J=:	K=:	L=V
Z=V	X=V	C=N	V=N	B=,	N=,	M=?	<=?
We can see from this that pressing either of the first two letters in a row generates the fourth letter in the row

Code:
Q, W =R
A, S =F
Z, X =V
Pressing either of the third or fourth letters on a row generates the sixth letter in the row

Code:
E, R = Y
D, F = H
C, V = N
Pressing either of the fifth and sixth letters on a row generates the eighth letter / character in the row

Code:
T, Y = I
G, H = K
B, N = ,
So there's a definite pattern to this.

It would be helpful to have the exact circuit diagram for the Pet 3000 series keyboard section, but in the meantime, as you have already tried replacing all the 'big' ICs associated with the keyboard, try to see if there is a 'four line to more lines' decoder between the I/O IC and the keyboard and if you can, change that, as you may as well rule that out before going further - assuming it's in a socket and is easy to change of course.

Some good pictures of that area of the PCB may help us to make further suggestions / compare it with any info which can be found online.
SiriusHardware is online now   Reply With Quote
Old 5th Feb 2020, 8:27 pm   #8
TonyDuell
Dekatron
 
Join Date: Jun 2015
Location: Biggin Hill, London, UK.
Posts: 3,574
Default Re: Commodore PET 3016

A couple of comments...

I think all PETs used a decoder IC to select the keyboard columns. It should be easy to find by tracing the connections back from the keyboard connector, some will go to a PIA/VIA IC, the others to the decoder IC.

Secondly, a lot of problems in PETs are caused by bad IC sockets. I have fixed several such machines with odd symptoms just by replacing all the (dubious quality) original sockets with good turned-pin ones. This is normally the first thing I do to a PET. It may not cure the fault, but at least you then know there is a 'real' fault to trace.
TonyDuell is offline   Reply With Quote
Old 5th Feb 2020, 8:33 pm   #9
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

Just found a nice PET repair video by the 8-bit guy here:-

https://www.youtube.com/watch?v=uHbhH7ISL_Y

The section at 2:42 onwards is of particular interest to you because it has a nice explanation / animation showing how the keyboard is scanned, and the various methods he uses to narrow down the fault.

Although his fault is different (no keys working at all) the diagnostic process still applies. At the moment my money is mainly on the 74LS145 decoder being faulty on yours since you've already changed the 6520.
SiriusHardware is online now   Reply With Quote
Old 5th Feb 2020, 9:13 pm   #10
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

In the light of what Tony said above about sockets on these, try just removing, inspecting, cleaning and reinserting the 74LS145, assuming yours has one. It's also possible that the last person who inserted that IC may have folded one of the pins over by accident.
SiriusHardware is online now   Reply With Quote
Old 6th Feb 2020, 4:58 pm   #11
JohnBHanson
Hexode
 
Join Date: Aug 2009
Location: Worthing, Sussex, UK.
Posts: 488
Default Re: Commodore PET 3016

Check for shorts between the outputs of the 74LS145 would be my first choice. This could cause the effect you describe. Looks like a short between the lines that drive the A and F keys or the keyboard cable not inserted correctly into the mother board. If you have a diagram of the keyboard matrix it should be easy to find.

keyboard matrix is given at
http://www.6502.org/users/andre/peti...rds.html#graph

Last edited by JohnBHanson; 6th Feb 2020 at 5:03 pm.
JohnBHanson is offline   Reply With Quote
Old 7th Feb 2020, 11:22 am   #12
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Thanks everyone, sorry for the late replies -I've been at work!
I really like the ideas of a pattern of key strokes and returns - hadn't thought of that!!

So here is the pattern:

Pressing QW, AS, ZX returns nothing - not even the cursor moves

Pressing EDC produces RFV - so the next characters along

Pressing RFV actually produces RFV

Similarly Pressing TGB produces YHN and YHN produces YHN

Pressing UJM produces IK, with IK, producing IK,

Pressing OL; produces P:? with P:? producing P:? as expected

All of these things happen whether the SHIFT key is activated or not - the characters remain uppercase

On the number pad moving in a line downwards you get:

Pressing 7410 produces 852. - the next column along although just occasionally - like when I first tried this this morning, the 1 returned a 1 followed by a . Now it just returns 852.

852. returns 852.

and 963- returns the expected /*+=

So there is a definate pattern!
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 11:24 am   #13
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Quote:
Originally Posted by TonyDuell View Post
A couple of comments...

I think all PETs used a decoder IC to select the keyboard columns. It should be easy to find by tracing the connections back from the keyboard connector, some will go to a PIA/VIA IC, the others to the decoder IC.

Secondly, a lot of problems in PETs are caused by bad IC sockets. I have fixed several such machines with odd symptoms just by replacing all the (dubious quality) original sockets with good turned-pin ones. This is normally the first thing I do to a PET. It may not cure the fault, but at least you then know there is a 'real' fault to trace.
Thanks Tony, its definately worth a try!
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 11:25 am   #14
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Just found a nice PET repair video by the 8-bit guy here:-

https://www.youtube.com/watch?v=uHbhH7ISL_Y

The section at 2:42 onwards is of particular interest to you because it has a nice explanation / animation showing how the keyboard is scanned, and the various methods he uses to narrow down the fault.

Although his fault is different (no keys working at all) the diagnostic process still applies. At the moment my money is mainly on the 74LS145 decoder being faulty on yours since you've already changed the 6520.
Thanks for this. I would tend to agree with you. these are soldered in and I am a bit nervous about taking these out. I tried to do this once on an old Roberts radio and ended up frying the circuit board!!!
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 11:49 am   #15
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

The pattern I mentioned above is what happens when I am running the machine in BASIC 4, running it in BASIC 2 (I think it is) produces this:

SHIFT key pressed stops all keys returning anything except:
@ produces a + and pressing 0.-+ returns graphics

Without SHIFT pressed you get:

QAZ gives WSX with WSX giving WSX
EDC = RFV with RFV = RFV
TGB=YHN and YHN=YHN
and so on

but in BASIC 2 I can press SHIFT and RUN STOP together tp load a programme in from tape
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 11:55 am   #16
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Pressing SHIFT and RUN STOP in BASIC 4 freezes the PET!
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 2:05 pm   #17
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

Thanks for the more precise breakdown of keys pressed and characters returned. I obviously didn't completely understand which keys you had pressed and in what order originally, so the clarification / corrected table is helpful.

As you aren't keen to try to remove the 74LS145 (yet), there are other avenues to explore first but they may eventually lead you back to the 74LS145.

Can we ask what you feel your technical 'level' is - don't want to insult you by aiming advice at too low a level or confuse you by aiming too high. What do you have in the way of test and repair gear and tools?

-Soldering and desoldering tools?
-Multimeter?
-Logic Probe?
-Oscilloscope?
-Any other microprocessor items which could be pressed into diagnostic service, such as an Arduino?

While you ponder that, one thing you could try is the 8-bit Guy's trick of unplugging the keyboard and using a wire link to connect one of the row pins on the mainboard's keyboard connector to each column pin in succession.

If that gives you a nice progression of successive letters in the same order as on the keyboard row without repeating any or missing any out, that would suggest that the main board is OK and focus attention on the keyboard.
SiriusHardware is online now   Reply With Quote
Old 7th Feb 2020, 3:05 pm   #18
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Please do not worry about pitching it too low! Although I have some rudimentary skills and knowledge it is just that! I can solder fine - just a little bit nervous about unsoldering! Although I agree with you that I may have to do it! I have a multimeter but not an oscilloscope - though I could get hold of one. I think I have an old Arduino kicking around - I tend to pick these things up and then forget about them! I definately think the 8-but guy is something I should be watching avidly. It's frustrating because I don't think it's too far away from working properly - but what do I know!
Many thanks for your help.
Best wishes,
John
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 5:29 pm   #19
John Earland
Hexode
 
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 296
Default Re: Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Just found a nice PET repair video by the 8-bit guy here:-

https://www.youtube.com/watch?v=uHbhH7ISL_Y

The section at 2:42 onwards is of particular interest to you because it has a nice explanation / animation showing how the keyboard is scanned, and the various methods he uses to narrow down the fault.

Although his fault is different (no keys working at all) the diagnostic process still applies. At the moment my money is mainly on the 74LS145 decoder being faulty on yours since you've already changed the 6520.
I've used a multimeter on the 74LS145 and I am getting some reading but to be honest - no idea what they mean!!
John Earland is offline   Reply With Quote
Old 7th Feb 2020, 6:11 pm   #20
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 3,604
Default Re: Commodore PET 3016

As I mentioned earlier, disconnect the keyboard from the main board and try simulating key presses by using a piece of insulated wire with tinned ends to momentarily link one of the row pins to each column pin in turn.

Depending on the results of that test, we'll work out what you should try or measure next.
SiriusHardware is online now   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 1:15 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2020, vBulletin Solutions, Inc.
Copyright ©2002 - 2020, Paul Stenning.