11th Feb 2020, 12:44 pm | #81 | ||
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
Quote:
pin 9 connects to Keyboard connector 8 with 0 ohms reading pin 10 to Keyboard connector 9 pin 11 to keyboard connector 10 |
||
11th Feb 2020, 1:06 pm | #82 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
That last is probably just a bit of brain fade on my part. Will look over everything again this evening.
|
11th Feb 2020, 6:08 pm | #83 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
Quote:
|
|
11th Feb 2020, 7:47 pm | #84 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
The measurements you've been doing so far served several purposes.
-To verify that the intended connections between the 6520 and the 74LS145 were OK -To verify that the intended connections between the 74LS145 and keyboard connector were OK. -To see whether any of the inputs and outputs of the 74LS145 exhibited some anomalous reading which might suggest a faulty IC input, faulty IC output or an unwanted connection from one of those inputs or outputs to something else. So far you haven't found any obvious problems, and in the course of not doing so have also managed to establish the layout of the keyboard connector, which is very useful. Unfortunately, this still doesn't exonerate the 74LS145. Most logic ICs are effectively a box full of similar elements which can be roughly compared with each other by measuring from some common point, such as 0V, to each one in turn. If you do such a comparison and you find that one of four inputs or one of ten outputs on the chip has a very different resistance to that of the others, that would be strong cause for suspicion that the input or output in question was damaged. You haven't found that in this case. However, it is possible for an IC to be faulty somewhere 'inside', in a place where the fault will not necessarily show itself as a measurable change on any of the pins. Unfortunately, it looks like you are going to have to rule the 74LS145 in or out by replacing it. If this had been easy to do (if the IC had been in a socket) we'd have asked you to do this on day one, but it was worth looking, first, to see if some other problem such as an open circuit track or shorted track might be the cause. |
11th Feb 2020, 7:54 pm | #85 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
I see - thank you. Well if it has to be done it has to be done!. I will cut each pin carefully and then take each one out - one at a time. I think I might be okay with that, it's soldering the pins of the strips that worries me. Normally I heath up the wire so that the solder melts onto it and forms a strong bond. I'm worried that if I do that this time I might also damage the delicate board connections. Any tips?
|
11th Feb 2020, 7:56 pm | #86 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
|
11th Feb 2020, 8:12 pm | #87 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
Cut the pins of the original IC high up, as close to the chip body as you can, never ever try to cut them low down beside the PCB / pads, you risk damaging a track / tearing off a pad if you do that.
With the body of the chip removed, you can then desolder the individual pins by melting the solder on just one pad at a time, and when the solder has melted, lift the pin out with a pair of fine pointed tweezers. Repeat until all the pins have been removed. Put a blob of solder on each of the pads, and then use a soldering iron and hand desolder pump (which you will need) to clear the holes. It might seem counter-intuitive to put more solder ON when you are actually trying to remove it, but desolder pumps work more effectively when there is a decent amount of wet solder for them to suck up. If any of the holes don't clear properly first time, put more solder on the pad and try again. Incidentally you need to use proper lead-tin solder, not the modern rubbish, which is awful. If using the strips is making you nervous, PM me your details (don't put them here in the open forum) and I will send you a couple of IC sockets along with a length of decent solder if you need it. You can always just solder the replacement IC directly in, but as you have seen, it is useful to be able to remove and change ICs easily. |
11th Feb 2020, 8:18 pm | #88 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
Hi - I'm really sorry but I don't know what a desolder pump is!! I've just checked on Ebay and there are loads - which one would you recommend?
|
11th Feb 2020, 8:29 pm | #89 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
I'll throw that question about the desolder pump open to others, as I use an electrically assisted desoldering iron myself but I assume you would not want to go to that kind of expense for a one-off job.
Broadly, hand desolder pumps look like a big metal or plastic cylinder with a plunger at the top and a tapered teflon nozzle at the bottom, and a button on the side. You 'arm' the tool by pressing in the plunger, which locks down, melt the solder you want to remove with an iron, whip the iron away, plant the nozzle over the still wet solder, press the button, the plunger flies up, hopefully drawing the solder into the body of the tool. If it doesn't clear first time, put more solder on the pad you are trying to clear, melt it again, repeat until the hole is clear. |
11th Feb 2020, 8:36 pm | #90 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
I've removed the IC but I think I'm going to have to ask someone I know who is good at soldering - There's no way I have the skills to solder the pins - there are 16 and they are so close together
Is there anyway I can check the new chip? |
11th Feb 2020, 8:48 pm | #91 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
Not easily, other than by trying it and seeing if it fixes the fault. You should be able to assume that a new IC is likely to be OK if you bought it from a trusted source.
It sounds like you have someone specifically in mind to ask to help, otherwise I might suggest you find your local 'men in sheds' or 'maker' group, or even consider taking it to a local electronics repair business if you can still find such a thing. if you show them the board, chip, and optionally a socket, they might give you a price and fit it while you wait. Last edited by SiriusHardware; 11th Feb 2020 at 8:56 pm. |
11th Feb 2020, 9:19 pm | #92 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
I found someone! Through a friend of a friend he works for an electronics company and has access to loads of equipment and has agreed to do it for me! However, not until 20th Feb. Im away for a few days anyway so it might be good!
|
11th Feb 2020, 9:29 pm | #93 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
Ok, in that case we'll pause this for now. Fingers crossed, changing the chip will fix your fault. Good luck!
|
12th Feb 2020, 8:07 am | #94 |
Hexode
Join Date: May 2013
Location: Middlesbrough, North Yorkshire, UK.
Posts: 442
|
Re: Commodore PET 3016
Thank you! I will let you know ASAP after 20th. I really hope so-it’s a wonderful machine-I just hope this solves it. Thank you guys for all your help so far-especially you-you’ve been amazing. I’ve learned loads! Mind you-it might not be the end!!
Hope to communicate again next week! Cheers for now, John |
12th Feb 2020, 1:28 pm | #95 |
Octode
Join Date: Nov 2004
Location: Hebden Bridge, West Yorkshire, UK.
Posts: 1,885
|
Re: Commodore PET 3016
I've been following this with interest and I just read another post on another forum that got me thinking. Are we sure the keyboard is playing up and not the video character RAM - I'm not sure how it works on the PET but a BBC Micro was showing similar behaviour in text mode (7) which turned out to be a dodgy DRAM - just a thought.
Can you enter a statement such as PRINT "HELLO" blind i.e. even though the characters look wrong will it execute? |
12th Feb 2020, 2:07 pm | #96 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
From something John said earlier it appeared that the machine was rendering text correctly from RAM, (Prompts from the OS / From software coming up correctly) and that the main problem was incorrect characters appearing in response to some key presses.
I believe John has already removed the 74LS145 keyboard column line driver in preparation for fitting of a replacement, so we may as well see what happens there before proceeding further, if necessary. |
12th Feb 2020, 3:15 pm | #97 |
Octode
Join Date: Nov 2004
Location: Hebden Bridge, West Yorkshire, UK.
Posts: 1,885
|
Re: Commodore PET 3016
Smart, it just got me thinking - I always try every conceivable avenue before removing components!
D |
12th Feb 2020, 4:12 pm | #98 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Commodore PET 3016
It's an interesting thought; but if that were the case, you would expect the relationship between the "wanted" and "actual" characters to be different. Faulty memory would manifest as characters with similar bit patterns but one bit always set or never set; for example, if bit 4 were stuck "on" then you might get @→H, A→I, B→J, C→K ..... P→X, Q→Y, 0→8, #→: and so forth. When the relationship is that the "actual" character is in the same row or column of the keyboard matrix as the "wanted" one, it's more likely that the fault is at the stage of reading the keyboard.
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
12th Feb 2020, 9:06 pm | #99 |
Octode
Join Date: Nov 2004
Location: Hebden Bridge, West Yorkshire, UK.
Posts: 1,885
|
Re: Commodore PET 3016
Fair play. It was just a thought
D |
12th Feb 2020, 9:45 pm | #100 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Commodore PET 3016
It is by no means certain that the column driver IC is to blame although the nature of the fault, in which column N and N+1 of a row produce the same letters, then column N+2 and Column N+3 produce the same letters, then column N+4 and N+5 produce the same letters, etc, makes that an obvious thing to eliminate at the earliest opportunity.
It sounds like the 'KEY A' (ones bit) line between the 6520 and the 74LS145 is either stuck in one state or being read as such by the 74LS145, resulting in the column select line advancing by two, not one, and only on every second step of the four bit count being output by the 6520. We can only wait and see. |