9th Feb 2021, 3:08 pm | #561 | |
Diode
Join Date: Feb 2021
Location: Southport, Merseyside, UK.
Posts: 5
|
Re: Non-working Commodore PET 3016
Quote:
Linux also has xxd which can convert from "hexdump" format back to binary. |
|
9th Feb 2021, 4:32 pm | #562 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
Hello Zed (Welcome to the forum) - I followed your link to 'srecord' but it appears that one of the few things it does not do is convert a straightforward ASCII text dump, although it can output to that format. The closest format it can load is 'ASCII-SPACE-HEX' which requires specific control characters at the beginning and end of the file, but is otherwise similar to a plain ASCII Hex dump.
Colin may have better luck in the short term if he can find a 'simpler' text editor (Maybe 'original' Notepad) to paste into, and inspects the pasted code for any oddball / extra characters or spaces and removes them before putting them through that online converter. I might try modifying the present Arduino sketch so that it outputs Intel Hex, HxD may be able to cope with extraneous characters on the beginning or end of Intel Hex lines. Last edited by SiriusHardware; 9th Feb 2021 at 4:41 pm. |
9th Feb 2021, 5:02 pm | #563 |
Diode
Join Date: Feb 2021
Location: Southport, Merseyside, UK.
Posts: 5
|
Re: Non-working Commodore PET 3016
Sounds like "xxd -r -p" on linux is closer to what is required then:
From xxd -h: Code:
-r | -revert Reverse operation: convert (or patch) hexdump into binary. If not writing to stdout, xxd writes into its output file without truncating it. Use the combination -r -p to read plain hexadecimal dumps without line number information and without a particular column layout. Additional Whitespace and line-breaks are allowed anywhere. Last edited by zedstarr; 9th Feb 2021 at 5:13 pm. |
9th Feb 2021, 5:20 pm | #564 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,821
|
Re: Non-working Commodore PET 3016
I have Ubuntu on his PC too. Maybe it's time for a reboot.
|
9th Feb 2021, 5:55 pm | #565 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
You may or may not know that you can get the identical Arduino IDE software for Linux, if you want to keep it all in one environment - but - I think HxD is a Windows only program so you'd need to locate an equivalent utility for Linux (very likely to exist).
If you want to go large there are libraries - add ons - for the Arduino environment which can -Add the ability to write files to / read files from a FAT32 formatted SD card. Obviously you have to connect the card to the Arduino somehow. -Add the ability to do a ZMODEM or YMODEM file transfer to any PC terminal program which supports those protocols, such as Hyperterminal. The last one would use the existing USB serial connection to send the file to the PC. I have not used that library though, so I couldn't be much of a guide. |
9th Feb 2021, 6:14 pm | #566 |
Nonode
Join Date: Jan 2009
Location: Bristol, UK.
Posts: 2,062
|
Re: Non-working Commodore PET 3016
hexdump ?
|
9th Feb 2021, 6:41 pm | #567 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,821
|
Re: Non-working Commodore PET 3016
I've install Arduino. I'll go look at the libraries - thanks.
|
9th Feb 2021, 6:57 pm | #568 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
Maybe a bit more information? Just googling 'hexdump' gets an awful lot of hits.
To recap, the situation is that Colin has an apparently working PROM reader project which sends its ASCII hex-dump output to the Arduino serial monitor. To get it from there and into a hex viewer, Colin is copying and pasting the output of the reader into a text editor and saving it as a file. The problem is that HxD, the hex editor he is using, does not know how to read simple 'ASCII Hex' so he needs a utility which will convert the ASCII-HEX text back to plain binary, HxD's preferred input format. |
9th Feb 2021, 8:07 pm | #569 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,821
|
Re: Non-working Commodore PET 3016
Quote:
xxd -r -p UF10001.txt dump.txt will convert it to dump.txt as a binary file. Okteta seems like a reasonable equivalent of HxD for Linux - available as a snap for Ubuntu. |
|
9th Feb 2021, 8:45 pm | #570 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
In today's news, Colin's PROMs arrived today and have now been read.
Results: UD6 - matches "basic-2-c000.901465-01.bin", checksum = 3838 UD8 - matches "edit-2-n.901447-24.bin", checksum = FDBF * UD9 - matches "kernal-2.901465-03.bin", checksum = 7C98 * although an 8K device this IC only contains 4K of code, the checksum is of the lower half (0000 - 07FF). Upper half contains all FF. UD7, marked 901465-02 - is dud. Contains almost all 0x80, the first two lines contain Strings of 'BE', 'A2', 'A6' and '82'. It's not a well chip, as we had already feared from Colin's measurements on it when compared to its sister UD6. For replacements, there are a couple of direct-fit options that I can see. One is to use an 8K / 24 pin MCM68764 EPROM with the 4K of code programmed into the upper half. The only pinout difference between the original PROM and the MCM68764 is pin 21: On the original PROM it is an active-high chip select: On the MCM68764 it is A12, the highest address pin, which, when plugged into the socket in the PET, is held permanently high, selecting the upper half of the IC. Another EPROM which seems an even closer match is the Texas TMS2532 - on that IC all of the pins seem functionally identical, although the critical ones are named differently. Pin 20 of the TMS2532 is PD/_PGM, but despite the way it is labelled the datasheet says that this pin should be at 5V unless the chip is to be enabled, in which case it should be brought to 0V. In other words, it performs the same function as the active-low CS1 pin does on the PROM. Pin 21 is VPP (the programming voltage input pin). When in read mode this pin needs to be held at +5V. On the PET circuit diagram pin 21 (CS3) of the PROM is held at 5V. I actually read the PROMs as though they were TMS2532s. Those are my thoughts for possible substitutes - can anyone see any snags with those or suggest anything else which might be drop-in compatible? |
9th Feb 2021, 8:48 pm | #571 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
Quote:
|
|
9th Feb 2021, 9:27 pm | #572 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
OK, looking back to the EPROM source Colin linked to in #462, that provider has both of the types I have suggested with the 2532 being the better priced item. You'd just have to confirm they are Texas branded TMS2532s and provide him with a link to the
basic-2-d000.901465-02.bin ...file, or email it to him. Let's hope he is still there and offering that service. |
10th Feb 2021, 3:54 pm | #573 |
Nonode
Join Date: Jan 2009
Location: Bristol, UK.
Posts: 2,062
|
Re: Non-working Commodore PET 3016
I do have a 1 TMS25P32JL complete with a 'panhandle' logo, also 1 Hitachi branded variant - both used/pulled. I do have a programmer somewhere but no eraser (and ~0 chance of finding / figuring out the software / hardware in any case)
PM if useful BTW 'hexdump' is a pretty ubiquitous linux command (linux-utils package, depending on distro). Re-reading I think it does the opposite of what you wanted. dc |
10th Feb 2021, 4:06 pm | #574 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
I have an eraser and programmer for those, but if you follow Colin's link in #462 the source there is offering them -programmed- for £6 I think it is. £3 blank.
By the time you sent that device to me and then I sent it to Colin we'd probably almost go through that just in postage - but I'm willing to programme it if that is how Colin wants to go. I'm also not too sure about the 'P' 'midfix' in that device label, EPROMs of that period are such a minefield with Single rail supplies / Three rail supplies that it might be best to keep to the 'Plain' TMS2532. Thanks for explaining 'hexdump'. |
10th Feb 2021, 10:09 pm | #575 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,821
|
Re: Non-working Commodore PET 3016
No reply from anyone yet. Are either of these suitable?
https://www.cricklewoodelectronics.c...ircuit-IC.html https://www.ebay.co.uk/itm/TMS2532JL...kAAOSwZexdPv3k |
10th Feb 2021, 11:12 pm | #576 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
I would give the chap on vcfed a little bit longer to respond, he's probably just fallen off his chair.
Have a look to see when he last signed into vcfed, he may just not have had a chance to see your enquiry. Cricklewood's description of their part is sending mixed messages:- Quote:
The part in Birmingham looks right although a little bit beaten up for a 'new' part, but if it works that doesn't really matter. See also the post by Dave Cox, #573. I am not absolutely certain that my programmer can programme the 25P32 as there isn't a specific selection for that variant in my programmer's software. I'd have to look at what the difference between the 'P' and 'non-P' variants are. It might just mean 'plastic' rather than ceramic. If I can't do them maybe someone else here can. Last edited by SiriusHardware; 10th Feb 2021 at 11:22 pm. |
|
10th Feb 2021, 11:45 pm | #577 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
Dave Cox mentioned a possible Hitachi equivalent, I looked and the Hitachi HN462532G does appear to be a direct equivalent with the right pinout but I have just looked and while my programmer does support some of the Hitachi HN-series EPROMs it does not support that particular device, sorry. Someone else's programmer might, though.
|
11th Feb 2021, 1:25 am | #578 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,586
|
Re: Non-working Commodore PET 3016
There is, incidentally, an active user of this forum who also goes by the same handle as the vcfed guy, 1980s_john. Not necessarily the same person, but maybe worth a PM to ask?
|
11th Feb 2021, 12:41 pm | #579 |
Diode
Join Date: Feb 2021
Location: Southport, Merseyside, UK.
Posts: 5
|
Re: Non-working Commodore PET 3016
As a side note, I was worried about the integrity of the data I'd read from my 27C512 directly into the Pi as the Pi's GPIO is all 3V3 (and not 5V tolerant) and the CMOS 27C family is a 5V part. In my setup I'd powered the 27C512 from the Pi's 3V3 bus. To satisfy my curiosity I purchased a TL866II USB reader from eBay.co.uk which has just arrived. I've just imaged the chip again using the 866 and compared with my original Pi-derived version and can confirm they're identical. So for the 27C family at least it looks like there's no need for level shifting logic if all you want to do is read using a Pi...
Code:
[cxf@ux303 EPROM Reader]$ md5sum elektor_scms_rom.bin 64bd11d7cd3a2ffcde307e94fd8fd679 elektor_scms_rom.bin [cxf@ux303 EPROM Reader]$ md5sum TMS27C512@DIP28.BIN 64bd11d7cd3a2ffcde307e94fd8fd679 TMS27C512@DIP28.BIN |
11th Feb 2021, 4:24 pm | #580 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,298
|
Re: Non-working Commodore PET 3016
I’ve read a few times in different fora that one method of recovering data from an eprom that has started to lose its memory due to old age is to vary the supply voltage. If this is correct then reading the eprom at 3v3 might not mean it would work correctly in its original 5v system.
Also if you read an eprom and find it doesn’t match the required content, how do you know if its because you read it at 3v3 instead of at 5v. For fault finding its probably better to read devices as close to the correct spec as possible. |