Pin spacing

edit

We say "40 pin DIL package" but were the pins on 0.1 inch (2.54 mm) centers or on 2.5 mm centers? Would this have fit into a Western 40 pin DIP socket? --Wtshymanski (talk) 19:55, 15 January 2018 (UTC)Reply

@Wtshymanski: As far as I know most U880 were manufactured with a 2.5 mm pin spacing. However, the export versions (80-CPU) as well as most U880 manufactured after 1990 probably had 2.54 mm pin spacing. Drahtlos (talk) 20:01, 15 January 2018 (UTC)Reply
Now to find citations for that so we can put it into the article. Must have been tough, maintaining two production lines, one with 2.54 spacing and one with 2.50. --Wtshymanski (talk) 20:43, 15 January 2018 (UTC)Reply
Both versions came of the same production line, only the metal part differed. A look at the images in the article will make this clearer: For a part with 2.5mm spacing (e.g. UA880D) the narrow part of all pins is in the centre of the wide part. For a part with 2.54mm spacing (80A-CPU) the narrow part is in the centre of the wide part only for the pins in the middle while at the end the narrow part is offset to the side. In other words, the plastic package and the metal inside the package are the same for both pin spacings. And yes, finding a citation is the hard part - I have not found anything so far. Drahtlos (talk) 20:55, 15 January 2018 (UTC)Reply
Here is a picture of 80A-CPU (left) and U880 (right), if you look closely, you can see how the pins on the 80A-CPU are offset as you get further away from the IC's center.
 
Side by side comparison of East German MME 80A-CPU (left) and U880 (right)
Skiselev (talk) 17:45, 23 May 2024 (UTC)Reply

It is not, however, a direct copy

edit

The die looks awfully similar though. Are we sure it isn't just some sort of copying inaccuracy? Something that hasn't been copied with enough fidelity to work properly, or some oversight, some missed gate or other during copying?

Further development - (Post)Soviet KR1810VM1

edit

While it is possible that some early T34VM1 and KR1810VM1 used the East German made U880 dies, later KR1810VM1 likely used domestically developed, or at least modified design.

There is a specific behavior of U880 and derivatives, where the OUTI instruction doesn't clear the carry flag (CF) when HL overflows/rolls over from 0FFFFh to 0x0000h. On most Z80 CPUs, including the original Zilog CPUs, the behavior of OUTI is to clear the CF in this case, even thought documentation says that OUT instruction shouldn't affect CF...

I have two KR1810VM1 CPUs, one dated 9305 and manufactured by Angstrem factory, and another dated 9306 and manufactured by Electronica factory. Both don't test as U880... Where, for example, Romanian made Electronica Bucharest MMN 80CPU, does test as U880, so presumably these guys used U880 dies...

Therefore the claim that post-soviet KR1810VM1 were based on U880 design is somewhat questionable. Skiselev (talk) 17:57, 23 May 2024 (UTC)Reply