21st Mar 2021, 2:37 am | #1221 | ||
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
|
Re: Non-working Commodore PET 3016
Quote:
Quote:
As when the original 74xx series was first brought out in 74Lxx Low-Power version, this did reduce the speed a lot, hence the popularity of 74LS to bring speed back up a bit again. But Acorn used a 74L42 (despite these being virtually obsolete by then) as well as 74LS42 on some of their System-1 cards, with apparently the reason that slower 74L solved a reflected pulse issue, when driving a long line on the bus. It's possible that more modern 74HCT devices (which seem to have become much more popular than 74ASxx & 74ALSxx) would provide lower power and fast speed, to replace 74S as well as 74LS - And maybe even plain 74xx , that did have better output drive than 74(L)Sxx). Plus possibly often the 74Lxx. But if you've not got any of these to hand, and originals are still available at a reasonably price, then might as well go for those. Given the rather low speed of the PET's clocks, it's surprisingly that 3ns types would be required. So maybe they had a bit of 'dodgy' logic timing issues, they fixed with a 74S device. Although I suspect using a 74LSxx temporarily would still work OK in many cases - and at least gives some outputs to see if it now possible worked OK and there wasn't more completely-dead IC's. |
||
21st Mar 2021, 6:06 am | #1222 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
You forgot to mention 74F, 74ACT, 74AHCT, 74ABT all with their own different tradeoffs, then there are also 74C, 74HC, 74AC, 74AHC, 74LV, 74LVC and probably many more.
HCT is very similar timing to LS in most cases, so no better as a replacement for the 74S. Using a faster replacement for the 74S74 might be no better than the slower LS as we don’t really know why commodore decided on the 74S. |
21st Mar 2021, 9:03 am | #1223 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
The MK14 VDU uses another 'awkward' IC in the form of the 74L86 which Slothie thinks is there precisely because of its unusually slow propagation delay, but in that case it even has a different pinout to the standard '86 so the temptation to use anything else is removed.
|
21st Mar 2021, 10:14 am | #1224 | |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
|
Re: Non-working Commodore PET 3016
Quote:
Alan |
|
21st Mar 2021, 10:38 am | #1225 |
Triode
Join Date: Feb 2021
Location: Kingston upon Thames, Greater London, UK.
Posts: 11
|
Re: Non-working Commodore PET 3016
I haven't read the entire thread but I noticed a recent discussion about 74S74..
Here's something that came up recently in a VCFED thread... "The fundamental difference between a 74LS74 and a 74S74 is output drive. A 74S74 can pull down 20mA whereas a 74LS74 can do less than half that!" This is very significant when driving a large fan-out of DRAM... you may find that a 74LS74 will not be able to reliably drive the memory; they are not the same and it is about drive not speed. |
21st Mar 2021, 11:51 am | #1226 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
|
Re: Non-working Commodore PET 3016
It does sound as if it would be best to get hold of a 74S74.
Alan |
21st Mar 2021, 12:19 pm | #1227 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,798
|
Re: Non-working Commodore PET 3016
I designed with all of these logic families as they went onto and off of the corporate preferred parts lists. There was quite a lot of work just checking the drive and threshold levels to be sure that a 1 was always seen as a 1 and ditto for zeros. On top of that was the effect of propagation delays on timing diagrams, as well as edge-speed effects. I was modifying some high speed programmable divider designs for frequency synthesisers. At first early stages were done in 74H into plain 74 and later on some 74L. The S and LS variants were well received when they appeared.
To determine why a rogue type appears in the PET would require construction timing diagrams for all the interactions passing through it, looking for races, plus calculating the margins by which levels are compatible, and then sticking a scope on one working to look for ringing and undershoots on the track structures of the board. If that isn't enough, it could also have been a manufacturing work around for a parts shortage. A tip-off from the designer, or a copy of his working notes is unlikely to happen. For a one-off fix of a single unit, a try it and see approach should save a lot of work because you don't have to do all the tolerance checking on top. At least you don't also have to juggle ECLII, ECLIII, ECL10K, ECL10KH and all the sub-flavours of 100K. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
21st Mar 2021, 1:08 pm | #1228 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
|
Re: Non-working Commodore PET 3016
Quote:
|
|
21st Mar 2021, 1:15 pm | #1229 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
|
Re: Non-working Commodore PET 3016
Quote:
I do quite like using 74LVC on more modern designs, but (paricularly wth single-gate ones) not all manufactuers have a very wide supply range that goes upto 5.5V / are 5.5V input tolerant with lower supply voltage. I'd only suggested an HCT one might be usable, as it is probably being most similar to a 74S, as being a still-current family then is likely to be more available (and also probably a bit lower power, being (High-speed) CMOS) |
|
21st Mar 2021, 1:26 pm | #1230 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
|
Re: Non-working Commodore PET 3016
Quote:
Thankfully, ECL logic isn't really encountered in consumer products - especially a 1MHz Commodore PET - with it mainly being in older expensive test equipment (and probably some exotic high-performance computers) when speeds of 100's of MHz were required. And use of ECL has rather died out these days, with high-speed CMOS. (Also, at least all the much older DTL families had been superseded by TTL back then) |
|
21st Mar 2021, 1:47 pm | #1231 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Non-working Commodore PET 3016
It's almost as though some designers forgot about AC behaviour (signals don't care if they are supposed to be analogue or digital); and when a circuit misbehaved because they had inadvertently created a mismatched transmission line, instead of adding a simple resistor and capacitor at one end that they probably already had in stock, they just picked a chip from a different logic family instead -- and lit the fuse on a reliability bomb.
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
21st Mar 2021, 3:39 pm | #1232 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
Quote:
The circuit just prior to this IC slices the main clock up into eight shifted phases and signals derived from the differences between those phases are used to clock / reset the 74S74 flipflops, so in this case I think it really is all about the speed / timing. |
|
21st Mar 2021, 5:32 pm | #1233 |
Triode
Join Date: Feb 2021
Location: Kingston upon Thames, Greater London, UK.
Posts: 11
|
Re: Non-working Commodore PET 3016
|
21st Mar 2021, 7:22 pm | #1234 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
I'll wait for the 74S74 to turn up. Should be Tuesday I think from Cricklewood - £2 per 74S74.
I have to say their customer service/speed of delivery has been very good for so many small orders from me. I would certainly recommend them for their stock items. Colin. |
21st Mar 2021, 10:13 pm | #1235 |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
|
Re: Non-working Commodore PET 3016
I used to use them a lot back in the late 80's for spare parts - mainly transistors, as they had one of the largest ranges of these inc. many hard to get elsewhere then Japanese 2S... ones, in the days of paper catalogues and no internet! (I even went to their shop once, whilst passing a few years later).
But I haven't used them for a while, as rarer parts now seem to be relatively expensive (especially with postage / handling) than what you often get via online marketplaces if prepared to wait a bit longer. And can often get several for the price of just 1 from UK sellers, which can be handy as (unless v.expensive parts), then I've always bought some extra spares for next time I need one. I seem to have been quite lucky, and not really had any problems with fakes / not receiving anything. I do still have the odd obscure TTL etc. IC's I need to find for Acorn / MK14 etc. rebuilds so may take a look there again sometime, as they don't seem to have a storefront on the online marketplace, so often forget about them. |
22nd Mar 2021, 4:48 pm | #1236 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
I guess that's why I'm using them still; they've had nearly everything I have wanted so far, and they're not too expensive for me - I get that P&P adds to the price, but they have to bear that cost somehow.
If I knew all the parts to buy beforehand, P&P would be a lot cheaper.... I'm also impatient. Colin. Quote:
|
|
22nd Mar 2021, 8:12 pm | #1237 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Non-working Commodore PET 3016
I've used Cricklewood from time to time over the years and never had a bad experience. They do seem to be a good source of hard-to-find parts at the moment, and cheaper than other places like Littlediode (although sometimes they are cheap. It depends!)
|
22nd Mar 2021, 8:17 pm | #1238 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
If it helps, Colin, the next IC you might need is a 74LS08 and I think you did get at least one of those as part of your mini bulk-buy. Of course the fact that you already have one makes it highly unlikely that you will need it. That's how it generally works: The parts you need are always the ones you don't already have...
|
22nd Mar 2021, 11:57 pm | #1239 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
Yep - I have two of them and 4 14 pin sockets left too. The chips on that side of the motherboard are covered in what I can only describe as verdegris. It seems to have suffered more from damp there.
Colin. Quote:
|
|
23rd Mar 2021, 10:02 am | #1240 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
By the time we get to the far corner I'm expecting... jellyfish.
If UH1 arrives today, can you redo the checks for activity on the two Q outputs of UH1, if there is some, look for activity on the output of the 74LS08 gate that those pins are connected to the inputs of. |