UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Computers (https://www.vintage-radio.net/forum/forumdisplay.php?f=16)
-   -   I found it! A very sorry looking MK14. (https://www.vintage-radio.net/forum/showthread.php?t=170727)

Timbucus 26th Mar 2021 8:48 pm

Re: I found it! A very sorry looking MK14.
 
Nice result on getting it going again and a sensible idea to use Vicky to test.

Mark1960 26th Mar 2021 9:58 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1357664)
I did actually buy a 4.43MHz crystal to put in my issue VI as well because some of the old manual software such as 'Music Box' and 'Digital Clock' depends on the crystal being that frequency, but then the prospect of reconnecting my VDU (and Karen's 'Ortonview' alternative), both of which prefer 4.00MHz, was just too good to resist so I never have changed it.

When i tried to buy a 4MHz and 4.43MHz crystal I ended up with 50 of each.

I think the ortonview would work ok with the 4.43MHz, as its not synchronized to the 8060 anyway. I’ll try it when I get round to building one.

Mark1960 26th Mar 2021 10:02 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1357677)
It would also let you plug in a high quality external keypad or even a keyboard with more keys on it, should the mood ever take you. (The MK14's row / column matrix can support quite a few more than the 20 keys on the original keypad).

Was there ever a suggested connection matrix for adding a qwerty keyboard to the MK14 ?

Buzby123 26th Mar 2021 10:12 pm

Re: I found it! A very sorry looking MK14.
 
The key matrix is 4x10, so 14 wires needed. A 15 way sub-D would work, but I've got plenty 25's, so a 25 it will be !.

Way back in the day, I had access to some neat hole punches, one for each size of sub-D. This time it will be drilling and filing !

SiriusHardware 26th Mar 2021 10:20 pm

Re: I found it! A very sorry looking MK14.
 
Once the SC/MP is running under your program's control it is obviously up to you which row / column nodes you choose to scan and how you choose to respond to them, but I don't think any wiring convention ever got as far as being on paper.

It was arguably the next thing in the pipeline, SOC hinted at their future plans in one or more newsletters which I am sure Tim can effortlessly produce a link to, but instead they made the ZX80, which begat the ZX81 and eventually the Spectrum, so it was probably just as well that they decided to leave the MK14 where it was.

For that reason, I don't think there was ever a 'standard' way of connecting a QWERTY keyboard but that leaves us free to invent one, perhaps utilising an existing standard keyboard such as the ZX81 membrane keyboard, still widely available.

A QWERTY keyboard would only really make good sense with an alphanumeric display permanently fitted, ie, VDU or OrtonView, along with an alternative OS to support them. So once again, we are really veering towards the evolution of the MK14 into a NIBL machine or something like it.

SiriusHardware 26th Mar 2021 10:23 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

The key matrix is 4x10, so 14 wires needed. A 15 way sub-D would work
Ah, you are forgetting about the two wires needed for the reset output from the uploader. That's too useful to leave out, it allows the uploader to force the MK14 back to the monitor if it already happens to be running a program, so it can upload another one.

Also, if you were using a deluxe external keypad you might want to have a reset switch on there as well.

SiriusHardware 26th Mar 2021 10:41 pm

Re: I found it! A very sorry looking MK14.
 
It's a shame you don't have access to your specific hole cutters any more but one trick I used to use when cutting holes for sub-Ds was to find a PC interface card of any kind / sort which had the right size of 'D' connector on it, temporarily remove the vertical metal backplate from it and use that as a kind of stencil with which to draw the desired outline for the cutout, including the positions of the holes for the screw posts, by running an inedible ink marker around the inside edge of the cutouts for the connectors and the screw post holes. Remove the plate and you have exactly the right outline marked on the housing and ready to cut out with the screw post holes marked at the correct relative spacing.

An old motherboard backplate, or one from a serial port card or multi i/o card should have at least one 25-way cutout which you can utilise in that way - for 15-way you'd need to find an old soundcard with a 15-way 'game' (joystick / MIDI) port on it.

Buzby123 26th Mar 2021 10:47 pm

Re: I found it! A very sorry looking MK14.
 
I like the PC-card template idea, I'll try that.

On another subject, has anyone got the full data sheet for the DM74S571 PROM ?. All the ones I can find just seem to be the first 3 pages of a much bigger document.

ortek_service 26th Mar 2021 11:10 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
There's a 4 page Datasheet for the DM74S571, here:

http://theoddys.com/acorn/Semiconduc...iconductor.pdf

And it doesn't look like it end unexpectedly, although there may be a separate section in the original databooks (probably on various archive / bitsavers etc. websites) on programming their PRON's, from what I recall.

Edit: Attached is an edied version of the NS 1977 Memory databook, that has (1 page) programming procedure for these, and also has the contents index, which appears to show from page numbers order that the 74S571 datasheet had 5 pages in that version.

SiriusHardware 26th Mar 2021 11:17 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
PDF pages 202-206 (5 pages) of this National Databook over on Bitsavers include a good description of how to program them. It's an entire book so it takes a little while to download.

http://www.bitsavers.org/components/...y_Databook.pdf

If you are thinking of programming one yourself don't forget there was an official MK14 PROM programmer from SOC, details attached. You seem to have working MK14s lying around all over the place now, so...

Bear in mind however that this and all other programmers which handle the DM74S571 can not program the Tesla MH74S571, the hardware method and timing for programming that device are just too dissimilar.

Also bear in mind that these devices are one-shot, you only get one go at getting it right.

Buzby123 26th Mar 2021 11:34 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
I was interested in find out what this 'testability' is in these PROMs, anyone know ?

ortek_service 26th Mar 2021 11:51 pm

Re: I found it! A very sorry looking MK14.
 
I've never seen that before - I presume it was in National Semiconductor documentation ?

And that it was all for their manufacturing tests, so you wouldn't really be able to do it yourself - Although it might be nice to be able to read these extra test fuses / even program some that weren't blown, to see how good an IC was before programming the data area.
I suppose they had a special test mode that they don't tell you how they entered, in order to do this.

Buzby123 27th Mar 2021 12:19 am

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
And there's more !

Today I found in the attic a box with about 30 databooks that I've not seen for years. One of them I think is the same NatSemi Memory Data as linked to a couple of posts ago.

I just had a quick look in it, and page 11-82 looks very interesting.

We MK14 owners know two things. (1) SoC cut every corner possible, (2) the PROMs get @***ing HOT !.

It looks like NatSemi had this heating in mind, and suggested a way to reduce it dramatically. That's what's on page 11-82.

Have any of you seen these PROMs in commercial products, did they get this hot in those applications ?

SiriusHardware 27th Mar 2021 12:33 am

Re: I found it! A very sorry looking MK14.
 
They always run very, very hot, alarmingly so. Whenever I programme a set for anyone I always warn them about that because they could be forgiven for thinking they were faulty otherwise. I worked out once that the PROMS consume 2/5ths of the entire current drawn from +5V by a fully populated MK14. You could always try putting little heatsinks on them I suppose but that would give them a non-original appearance. I quite often see heatsinks added to Sinclair ULAs, another famously hot-headed series of ICs.

Just lately I experimented with pin-compatible PROMS from other manufacturers, first the N82S131 from Philips / Signetics which also runs quite warm, possibly not quite as hot as the Nat Semi devices.

Then more recently I bought a pair of the AMD 27S13As offered here on the forum by kan_turk and although they still warm up they don't seem to run anything like as hot as the Nat Semi devices. Tim has those now, but I don't imagine he will have run them for long enough to notice.

ortek_service 27th Mar 2021 3:15 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by Buzby123 (Post 1357754)
And there's more !

Today I found in the attic a box with about 30 databooks that I've not seen for years. One of them I think is the same NatSemi Memory Data as linked to a couple of posts ago.

I just had a quick look in it, and page 11-82 looks very interesting.

We MK14 owners know two things. (1) SoC cut every corner possible, (2) the PROMs get @***ing HOT !.

It looks like NatSemi had this heating in mind, and suggested a way to reduce it dramatically. That's what's on page 11-82.

Have any of you seen these PROMs in commercial products, did they get this hot in those applications ?


That's interesting App Note info info, I'd not noticed before - Especially as it was missing from the very abridged 20pgs pdf I'd previously got with the TTL/Bipolar PROM's programming procedure page and just the contents pages (minus the last one, listing the App Notes), and the cross-reference pages.

This 1977 databook 74S571 datasheet on pgs 5-13 to 5-17 also covers the 74S570 Open-Collector version, - as well as seemingly duplicating the Programming procedure for these on page 5-1.
It seems the 4 pages (3-16 to 3-19) later date unknown? one, I'd previously linked to was complete, but that NS had removed the programming procedure duplication from it (so needed other pages from their memory databook), as well as removing details on the internal circuitry - but they had added an extra SM PLCC version, and also that Testability note you'd noticed. So useful to have a copy of both versions.

To be fair to SoC, I don't think it was they who'd cut corners regarding power consumption reduction, as I'd never seen anyone else doing that, but National Semiconductor themselves - As most memory devices have a Chip Select as well as an Output Enable, to save unnecessary power.
If a program was running entirely from RAM, then you could save quite a bit, but if the with the monitor in the PROM's running to update display & scan keys etc. then the PROM's would still be frequently accessed - albeit needing to be selected for rather less than 50% of the time.


But rather than trying to mod the original design to include power-saving, then adding heatsink(s) would probably be simplest to preserve the PROM's a bit, and efficiency is not too important if not left running all day. If D/S thermally-conductive adhesive tape was used, rather than thermally-conductive epoxy etc, then they could be later removed for originality.

Or a metal plate clamped over the PROM etc. IC's, with some thermally-conductive gap-filler silicone rubber sheet underneath (would be easier if they'd laid the board out with PROM's adjacent to each other like others did!) - Just need to make up a suitable non-conductive clamp system, to fit around PCB, without some convenient mounting holes for it.

This does allow you to revert it be just a museum piece / retain resale value, whilst actually being able to use it wit less worries of these IC's failing - Although I'm tempted to just substitute a plug-in EPROM adaptor assembly for the PROM's, and keep the originals safe.

It may also be worth seeing if the bit lower power AMD 27S13A or N82S131 that SiriusHardware has tried could be obtained at a fair price, and are easily-programmable by most 74S571 programmers / supported by some common universal ones.

SiriusHardware 27th Mar 2021 4:22 pm

Re: I found it! A very sorry looking MK14.
 
For absolute originality I suggest an original machine really does have to have 2 * DM74S571 fitted or at least available to be fitted because SOC never, to my knowledge, supplied alternatives. It wouldn't be a bad idea to have a pair of the cooler running types in whenever the machine is allowed off the leash, but if it ever comes to the day when the machine has to be sold on or given to a museum, there should really be a pair of DMs waiting to be put in it.

My MK14 was heavily used for several years and for most of that time it had the SOC supplied New-OS DM74S571s that it still has now, so although they do run very hot, somehow, generally, they keep working (...usually. One of Buzby's original SOC supplied PROMs failed after maybe 12-15 hours of use after decades of inactivity).

Mark's issue VI has N82S131Ns in it, I would be interested to know if they appear to him to run any cooler than the DM74S571s in his original issue IV. They still run quite warm, but to my mind not as hot as the DMs.

Buzby123 27th Mar 2021 4:27 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
Micky is not having anymore experiments done on him, he has earned a decent retirement.

Vicky however is going to take his place, and the power reduction mod looks like it's the first to be tried.

Below is the /OE signal on the 571s while scanning keyboard and display. As you can see, it spends the vast majority of it's time in the dis-enabled state. The /OE pulses are 500nS wide, and the gaps are mostly 3uS to 15uS, so a power reduction of 90% is likely.

I'll try this mod next week, in a new thread.

Cheers,

Buzby

Timbucus 27th Mar 2021 5:28 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1357713)
Once the SC/MP is running under your program's control it is obviously up to you which row / column nodes you choose to scan and how you choose to respond to them, but I don't think any wiring convention ever got as far as being on paper.

It was arguably the next thing in the pipeline, SOC hinted at their future plans in one or more newsletters which I am sure Tim can effortlessly produce a link to, but instead they made the ZX80, which begat the ZX81 and eventually the Spectrum, so it was probably just as well that they decided to leave the MK14 where it was.

For that reason, I don't think there was ever a 'standard' way of connecting a QWERTY keyboard but that leaves us free to invent one, perhaps utilising an existing standard keyboard such as the ZX81 membrane keyboard, still widely available.

A QWERTY keyboard would only really make good sense with an alphanumeric display permanently fitted, ie, VDU or OrtonView, along with an alternative OS to support them. So once again, we are really veering towards the evolution of the MK14 into a NIBL machine or something like it.

I agree - it is something I plan to do to go with the machine when it has a VDU

The effortless link is:

http://www.computinghistory.org.uk/d...Update-Letter/

which talks about the BASIC board that needs a VDU and the 40Key (4x10 matrix) keyboard.

That matrix will mean that the ZX81 and Spectrum ones cannot be used as they are 5x8 but, I assume it would be logical to use a similar physical layout. We really need to try and contact Nick Toop and see what they actually planned in the way of a keyboard.

Out of interest the May 1980 Computing Today contains a Classified ad for an improved monitor with and Alphanumeric keyboard... so at least one was done by a third party.

Timbucus 27th Mar 2021 5:33 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by Buzby123 (Post 1357904)
Micky is not having anymore experiments done on him, he has earned a decent retirement.

Vicky however is going to take his place, and the power reduction mod looks like it's the first to be tried.

Below is the /OE signal on the 571s while scanning keyboard and display. As you can see, it spends the vast majority of it's time in the dis-enabled state. The /OE pulses are 500nS wide, and the gaps are mostly 3uS to 15uS, so a power reduction of 90% is likely.

I'll try this mod next week, in a new thread.

Cheers,

Buzby

Sounds interesting - nice find.

SiriusHardware 27th Mar 2021 6:40 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

a power reduction of 90% is likely.
Certainly an interesting lead, potentially interesting not only to lower the running temperature of the PROMs but also to take some of the heat off the regulator which, if mounted on board as it is supposed to be, rarely has a big enough heatsink.

Slothie 27th Mar 2021 7:02 pm

Re: I found it! A very sorry looking MK14.
 
I have a suspicion that the Tesla MH74S571 dissipates less heat, as I don't recall noticing them getting hot on my original prototype. My new r1.2 prototype has DM74S571's (Kindly donated by a very nice man) and I'm thinking of designing a grill attachment for them!

SiriusHardware 27th Mar 2021 7:08 pm

Re: I found it! A very sorry looking MK14.
 
A couple of things I notice about the document in Tim's link in #268, one is the claim that the VDU is 'full ASCII' which is a bit disingenuous when the reality is upper case only and 64 characters altogether.

Another is the reference to 'adjustment' of issue I, II and IV PCBs to make way for the projected 2K memory expansion (meaning, presumably, removal of the unwanted PROM images on those versions). It's interesting that issue 1 (which I have never yet seen) is mentioned but issue 3 (which I thought I actually had seen) is not.

SiriusHardware 27th Mar 2021 7:21 pm

Re: I found it! A very sorry looking MK14.
 
Code with a side order of eggy toast: What could be nicer?

The Tesla devices may not have appeared until quite a bit later and I think the AMD 27S... devices are also next-gen - those databooks linked to a few posts back have cross references to quite a few other-manufacturer devices that the Nat Semi devices are equivalents of but both the Tesla and AMD devices are absent from those tables.

I wonder if the Tesla clones came about as a result of a cold-war embargo which prevented US technology from being supplied to iron-curtain countries (pure speculation on my part, but it would explain why so few 'western' programmers support the Tesla devices, which maybe didn't appear on the world market until after the fall of the iron curtain).

The AMD 27S13As I got from kan_turk, being the 'A' version, are the faster version which usually goes hand in hand with being more power hungry, so the 'plain' 27S13 may be even more economical in that respect.

ortek_service 27th Mar 2021 8:04 pm

Re: I found it! A very sorry looking MK14.
 
It seems the programming of the Tesla ones is closer to the Texas Instrument versions. So maybe they were a clone of those (that do seem less common that NS ones), and they have similar power consumption.
It's a pity there never seems to have been an 'L' Low power version, like with some EPROM's etc, before (High-Speed) CMOS versions were available.

It seems there was quite a lot of cloning of Western technology in the Eastern-bloc - Often whole Sinclair computers, including their own ULA.
And apparently they had a series of logic IC's in DIL packages with a metric pitch! (Recently found RS sell 2.5mm pitch veroboard, after someone bought it by mistake, when there's very few metric-pitch leaded connectors etc.)
The only Russian etc. made semiconductors that seem to make it over here back then, were transistors in those 5" B&W Rigonda VL100M TV's, popular with Caravaners etc. as had a bolt on the back PSU plugged into TV's 12V DC input.

SiriusHardware 27th Mar 2021 8:15 pm

Re: I found it! A very sorry looking MK14.
 
Yes, it's been suggested before that the MH devices could actually be programmed using the hardware method and algorithm for the SN (Texas) prefix devices - only trouble is, Texas don't ever seem to have made an SN74S571 so it's not as though you can just choose 'Texas / SN74S571' on your 'western' programmer to programme MH74S571s.

Texas did make a SN74S287, and here is an article about programming those - the author points out what may be a critical difference between the SN and MH devices (roughly 8th paragraph down) so they may not be as directly programmer-compatible as it first seems.

http://randomvariations.com/2014/09/...87-programmer/

circuitryboy 27th Mar 2021 11:19 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
I hadn't seen that "Christopher" 2014 posting before.
In 2019 I showed that the Tesla PROMs (MH74-188, -S287, and -S571) had those similar part numbers and used exactly the same programming protocol as the earliest TI (SN74-) PROMs. But TI didn't have a '571.
The TI data shows the '287 programs 'the other way'. I'm not sure why "Christopher" believes the MH74S287 does not. He wasn't using them, just warning to check the datasheet.
Later TI PROMs have different prefixes and require other/more voltages in their programming sequence.
How far back do the Tesla chips date? Certainly to 1989.
Remember that rogue "MH74S571" page in German which confusingly indicated 21V?
It's actually a page of Fairchild 93438 from a databook published in DDR, 1989. The MH74S571 was described immediately before that (without programming details). And somebody didn't notice the change.

Slothie 28th Mar 2021 12:09 am

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by circuitryboy (Post 1358068)
I hadn't seen that "Christopher" 2014 posting before.
In 2019 I showed that the Tesla PROMs (MH74-188, -S287, and -S571) had those similar part numbers and used exactly the same programming protocol as the earliest TI (SN74-) PROMs. But TI didn't have a '571.
The TI data shows the '287 programs 'the other way'. I'm not sure why "Christopher" believes the MH74S287 does not. He wasn't using them, just warning to check the datasheet.
Later TI PROMs have different prefixes and require other/more voltages in their programming sequence.
How far back do the Tesla chips date? Certainly to 1989.
Remember that rogue "MH74S571" page in German which confusingly indicated 21V?
It's actually a page of Fairchild 93438 from a databook published in DDR, 1989. The MH74S571 was described immediately before that (without programming details). And somebody didn't notice the change.

A year ago I translated a data sheet from Czech from the Tesla manual that details the programming procedure in this thread: https://www.vintage-radio.net/forum/...d.php?t=146388 (first post) I have since received details from an actual Czech with experience of programming them that confirms my translation (transliteration?) was broadly correct, along with some designs that I am not at liberty to divulge but I can use to guide my own design.
AFAIK although there are similarities, Tesla 74S571 PROMs need specific programming that is different in the details. One of my many projects in my "in progress" list is a Arduino based programmer for Tesla MH74S571 PROMS.

Slothie 28th Mar 2021 12:12 am

Re: I found it! A very sorry looking MK14.
 
Crumbs I just looked at the date, that was nearly 3 years ago!

SiriusHardware 28th Mar 2021 12:16 am

Re: I found it! A very sorry looking MK14.
 
Quote:

In 2019 I showed...
It was your work I was referring to in #275. If anyone has a Tesla MH74S287 and a programmer which can natively programme the SN74S287 then it would be great if they could try programming the MH74S287 'as though' it was an SN74S287.

Did you ever complete the conversion of the SOC PROM blower code to the MH... compatible version and compatible hardware which you were pursuing in this thread...?

https://www.vintage-radio.net/forum/...d.php?t=160247

I think you got as far as making a working conversion of the PROM editor code but we never did hear whether you also did that for the PROM blower code.

SiriusHardware 28th Mar 2021 12:21 am

Re: I found it! A very sorry looking MK14.
 
Quote:

that was nearly 3 years ago!
A sobering thought, we have now been going on about the MK14 and related matters for longer than the machine itself lasted as a commercial product. ;)

SiriusHardware 28th Mar 2021 12:44 am

Re: I found it! A very sorry looking MK14.
 
If anyone does have a programmer which can 'do' the Texas SN74S287 then it might be possible to programme an MH74S571 in two passes, with a slight adaptation.

The difference between the two is pin 14 (CE2 on 287, A8 on 571). You could tie pin 14 low to programme the lower 256 nibbles, and then tie pin 14 high to programme the upper 256 nibbles, both times with the programmer set to programme the 256 x 4-bit SN74S287.

Unfortunately my programmer does not offer the SN74S287 either, so I can't try it.

ortek_service 28th Mar 2021 2:14 am

Re: I found it! A very sorry looking MK14.
 
2 Attachment(s)
Quote:

Originally Posted by SiriusHardware (Post 1357998)
Yes, it's been suggested before that the MH devices could actually be programmed using the hardware method and algorithm for the SN (Texas) prefix devices - only trouble is, Texas don't ever seem to have made an SN74S571 so it's not as though you can just choose 'Texas / SN74S571' on your 'western' programmer to programme MH74S571s.

Texas did make a SN74S287, and here is an article about programming those - the author points out what may be a critical difference between the SN and MH devices (roughly 8th paragraph down) so they may not be as directly programmer-compatible as it first seems.

http://randomvariations.com/2014/09/...87-programmer/


Yes, it seems TI were relatively late to Bipolar PROMs (Maybe busy keeping up with Intel on EPROM's), and they never made some of the larger ones.
Although it seems National Semi weren't there first either, with it actually being Harris (Even though they don't seem to be as widely used as others, as well as not as many programmers supporting these): http://www.bitsavers.org/components/...is_Bipolar.pdf
"Harris introduced the industry's first bipolar programmable read only memory (PROM) in 1970, and has continued as a leader in the field of bipolar
PROMs. Harris offers a complete spectrum of bipolar PROMs from 256 bits
to 64K bits. In 1982, Harris became the first supplier of a new family of programmable logic (HPL) featuring patented on-chip testability"

So I wonder if all makes of Bipolar PROM do have on-chip Testability, that National mention in their later DM74S571 datasheet.

Most manufacturers have cross-reference guides with comprehensive list of other manufacturer's types. But they all seem to not mention anything about programming compatibility.


I recall seeing that randomvariations webpage before, where it does seem to indicate that Tesla devices are like National ones in one respect that the bits are all at 0 unless programmed, whereas, TI types appear to start out at 1 (like EPROM's). I know National and some other makes do start out at 0, from ones I've checked, but can't recall what Tesla ones were when I once programmed a couple and haven't current got access to that programmer to check. But I do see that you had said that Tesla ones had been confirmed as also starting with bits at 0, in post 17, here:
https://www.vintage-radio.net/forum/...d.php?t=146388

I've had a play on my older Dataman-48XP programmer , to see what it thinks the blank state should be for the PROM's it supports (Lots of AMD 27S and Philips/Signetics 82S types + a few National DM74S and a couple of TI TBP + Harris HM-66xx ones - Although no MMI or Tesla ones at all).
And all have the blank state as 0, including the TI TBP18S (4bit?) types.
However, the TI TBP28(L/P/S)(2/4)2 8bit ones have blank state of 1 (like it seems the SN74Sxxx ones are)

Strangely the ELNEC-designed Dataman-48Pro range, which supports the Tesla ones, amongst many others, only supported two Phiips/Signetics 82S10(0/1) types)


I have just found this webpage on fusible link proms, with a manual-switches type (I presume you can't over-programme a fuse!) programmer schematic for Signetics ones, and some info - that basically says you need to check specific datasheets for each make (but not really anything about 74S571): https://www.retrotechnology.com/rest...prom.html#data


Attached are about the most comprehensive data I could find on MH74Sxxx ones' the 'net, with some of it from Tesla's website. And looks to contain the programming info that Slothie had attached a translated version of.

ortek_service 28th Mar 2021 3:24 am

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1358086)
>>
If anyone has a Tesla MH74S287 and a programmer which can natively programme the SN74S287 then it would be great if they could try programming the MH74S287 'as though' it was an SN74S287.

Well as you got it confirmed that Tesla ones have blank state of 0, and would appear that TI SN74Sxxx ones have blank state of 1, then I certainly doubt trying this would work.

And unfortunately, finding programmers from the last 20+ years that support the TI SN74Sxxx ones, seems even more difficult than for Tesla MH74Sxxx.

I've searched the >900 page pdf latest device list of the ELNEC-designed Dataman 48Pro one, that did a Tesla MH74S287 one OK, and that doesn't list any TI 74Sxxx types at all (or any MMI 63xx-xx ones, with only 2 not a common type of Harris ones - So no way of programming those ones I've got, unless they do program the same as AMD/Philips/National equivalents).

Mark1960 28th Mar 2021 7:33 am

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1357900)
Mark's issue VI has N82S131Ns in it, I would be interested to know if they appear to him to run any cooler than the DM74S571s in his original issue IV. They still run quite warm, but to my mind not as hot as the DMs.

I’ll see if I can find a second power supply so I can set them both up beside each other.

I’ve also added a non contact thermometer to my shopping list in case I feel the need to take more objective measurements.

Looking at the datasheets, the DM74S571 max supply current is quoted as 130mA and typical 90mA at 5.5v, while the 82S131 quotes only the maximum of 140mA at 5.25v. This would suggest the 82S131 would run hotter.

SiriusHardware 28th Mar 2021 10:45 am

Re: I found it! A very sorry looking MK14.
 
That didn't seem to be the case to me, but I will leave that to you to investigate. I do have some more blank 82S131s here but I don't want to programme them for purely whimsical reasons (which probably means they will never be programmed with anything - always the problem - when is code ever important enough, and bug free enough, to commit to one-shot memory devices?). It's a shame no-one ever came up with a pin-compatible Flash version.

Buzby123 29th Mar 2021 1:29 pm

Re: I found it! A very sorry looking MK14.
 
3 Attachment(s)
First attempt at powering Micky's display.

I'm using Vicky to test. Things to note, the display is very dim. ( It looks OK because phone cameras are more sensitive at the red end of the spectrum, but trust me, it's dim ! )

A few segments missing. The display should be '8888 1F'

EDIT : Rusty pins !

Buzby123 29th Mar 2021 2:04 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
I think I need some new digits :(

SiriusHardware 29th Mar 2021 2:48 pm

Re: I found it! A very sorry looking MK14.
 
From experience, you may need a full set because if you replace just some of them the brightness of the new ones will probably be noticeably different. Slothie reckons modern displays produce a lot more light per mA than equivalent displays from 40 years ago.

Going way back to near the beginning of this thread you did say that in your recollection the segments on the big display had always been dim, at the time you seemed to feel that you had possibly not got the display driver design or component values right originally?

Buzby123 29th Mar 2021 3:09 pm

Re: I found it! A very sorry looking MK14.
 
Yes, the display was dim from day one, but I wasn't too bothered at the time. I was fairly certain I could just replace the 150 ohm resistor pack to get a bit more brighness.

However, that's a minor problem compared to the rusted pins.

The displays are FND507, and I'd prefer to replace with the same.

This project is 1 step forward, 2 steps back !. Will I ever get Micky working ?

ortek_service 29th Mar 2021 3:53 pm

Re: I found it! A very sorry looking MK14.
 
I presume you have fitted for IC13, the original 7445 (80mA sink) ?
- rather than the 7442 (16mA sink), that NS used on their Introkit with it's slightly smaller pocket calculator display / keypad or a lower-power version of either.

SiriusHardware 29th Mar 2021 3:57 pm

Re: I found it! A very sorry looking MK14.
 
I think it's possible that Buzby's display driver actually expects there to be a 7442 fitted rather than a 7445 - isn't that what Micky had fitted when it originally emerged from the loft?

It's not just the current output capability, isn't one totem-pole and the other open-collector outputs?

I think, Buzby, that you will still be able to obtain brand new 7-segs with the correct common (Anode? Cathode?) and the exact same pinout but surviving NOS examples of the original FND507 types could very well be dim just due to age even if you can still find them.

I built a Maplin Z80 system in the early nineties, played with it for a while, put it in the loft, got it out ten years later and the 7-segs were all very dim compared to how they had been, despite the system never having been run in the interim. I looked all over for a fault and then finally fitted a set of replacement 7-segs which brought it back up to the original brightness. I have no idea what the failure cause / mode of the original displays was. The replacement ones I fitted are now twice as old as the original ones ever were, and the display is still nice and bright.

ortek_service 29th Mar 2021 4:29 pm

Re: I found it! A very sorry looking MK14.
 
Yes, the 7442 can also source current but not very much - only 0.8mA, compared to its 16mA sink capability. So rather asymmetric, and you'd be unlikely to use it as a high-side driver (Plus it is low-side in the MK14, although you may be able to use some beefier high-side driver gates).
I don't think I've seen a CMOS 74HC42 etc. CMOS pin-compatible version, that would give you better high-side drive. And I presume they didn't bother with an 74L / 74LS in these, if they were design to sink quite bit of current compared to conventional L / LS TTL outputs.


I remember buying some similar looking FND500's, on a Digital capacitance meter EPE project, back in the 80's, and these may be the opposite common polarity. But I never really used it, as never found it worked too well (maybe because they said you had to use the Mullard version of the Display-counter IC, whereas I'd bought cheaper Motorola one - although I did at least mod their circuit to produce a proper square-wave clock into it)

SiriusHardware 29th Mar 2021 6:32 pm

Re: I found it! A very sorry looking MK14.
 
I've just realised that Buzby's original displays effectively have a built-in bezel / red filter so that's understandably why he wants to source some similar ones.

ortek_service 29th Mar 2021 6:37 pm

Re: I found it! A very sorry looking MK14.
 
Yes, the FND500 I got, also had this, looking very similar.
I seem to recall I got them (new) for only around 20p each from Sendz Components (surplus TV parts supplier), but may have been from Greenweld / Cricklewood etc.

SiriusHardware 29th Mar 2021 6:51 pm

Re: I found it! A very sorry looking MK14.
 
I found some of Buzby's original type with an online seller who Tim and I have both independently used before, I've sent Buz a PM about them.

Buzby123 31st Mar 2021 12:53 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1359146)
This repair may cost you a bit more yet, ... It all adds up.

Don't I know it :(

SiriusHardware 31st Mar 2021 6:46 pm

Re: I found it! A very sorry looking MK14.
 
1 Attachment(s)
Another little step forward for the project, a DM74S571AN arrived from Buzby today, so... (image).

To be honest I'm posting this partly to cover my own rear exit, because Buzby's original SOC-supplied IC2 (marked '1o', offset to one side upper right of the picture) died not ten running hours after it was originally working here. So here is proof that the original IC3 and the replacement IC2 were working on this day. I couldn't find a newspaper headline to include in the shot. ;)

I ran these for about an hour and a half in my original MK14, because that machine has the regulator mounted offboard on a substantial heatsink whereas my issue VI replica only sports a skimpy little black number.

One thing I did notice is that the original IC3 (Non-A version, large label on top) runs hotter than the replacement IC2 ('A' version, no label) so I would suggest using small labels like the sticky 'dots' SOC originally used on these new-OS PROMs and placing them at one end of the IC rather than over the middle, which is where SOC originally put them.

We can now add the DM74S571AN to the list of specific devices known to have been tested in the MK14.

Buzby123 31st Mar 2021 7:13 pm

Re: I found it! A very sorry looking MK14.
 
Quote:

Originally Posted by SiriusHardware (Post 1359227)
.. To be honest I'm posting this partly to cover my own rear exit, because Buzby's original SOC-supplied IC2 (marked '1o', offset to one side upper right of the picture) died not ten running hours after it was originally working here. So here is proof that the original IC3 and the replacement IC2 were working on this day. I couldn't find a newspaper headline to include in the shot. ;) ...

Don't worry, I'd have trusted you if you just said 'it works'.

The the original PROMs you tested for me worked fine when I received them back, but one stopped working the next day. That's just my luck, which still hasn't changed. ( Today, the original bubble display I salvaged last week decided to lose a segment !. )

Quote:

... original IC3 (Non-A version, large label on top) runs hotter than the replacement IC2 ('A' version, no label) ...
I didn't notice the new one was an 'A' version. Is there a significant difference ?.

Cheers,

Buzby

SiriusHardware 31st Mar 2021 7:15 pm

Re: I found it! A very sorry looking MK14.
 
Speed.

But the non-A version is fast enough, so the 'A' version is more than fast enough. No difference between them from the MK14's point of view.

SiriusHardware 31st Mar 2021 8:05 pm

Re: I found it! A very sorry looking MK14.
 
I suggest you save your life at regular intervals, and then when bad things happen you can just go back to an earlier save.

Quote:

the original bubble display I salvaged last week decided to lose a segment !
Is that

-One segment of just one cell?

-The same segment in all of the cells?

or

-One whole cell off?


All times are GMT +1. The time now is 5:28 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.