MK14 SN74S571 PROM Power reduction test - Page 2 - UK Vintage Radio Repair and Restoration Discussion Forum

Buzby123 · 20th Nov 2022, 12:14 pm

In the thread about the waalsdorp sc/mp ( https://www.vintage-radio.net/forum/....php?p=1514816 ) there was a link to a 1976 publication showing a SC/MP system.

Also in that mag was an article titled 'High Performance, Low Power Memories from Inexpensive Parts!'

It's just another circuit to stop your PROMs getting too hot !.

ortek_service · 20th Nov 2022, 12:47 pm

Yes, I recall previously seeing this - Probably when I first discovered these "Compute newsletters" a while ago.
It does seem rather reliant on these PROM's being happy with no power supply, whilst still in-circuit. Many IC's would have got parasitically-powered from the data/address lines etc that could result in latch-up when the power was applied and possibly not working getting rather hot (I once encountered that with PLL IC's, where the supply didn't rise quick enough compared to voltages on input pins, as supply was a simple Shunt zener stabiliser and series R + a few uF's capacitance across the zener caused an RC time-contant delay. So I had to put some resistors in series with the inputs, to restrict the I/P current below that which could cause latch-up)

Many years ago, I heard about a RAM IC that died after a few months use, and they then discovered (Fortunately only on prototypes) its power rail had never been connected (The Mentor CAD system put power-rails as a separate block, that was probably on another sheet, and it seems any DRC warnings were ignored / disabled)

SiriusHardware · 20th Nov 2022, 1:20 pm

The article makes this happy behaviour of the DM (National Semiconductor) devices out to be a distinct virtue of theirs - you would instinctively expect that if you had a 'dead' (unpowered) device sitting on your data lines you would be in for all sorts of trouble. It seems unlikely that pin compatible devices from other manufacturers (AMD, Philips/Signetics, Tesla) would behave in such a similarly convenient fashion, but it may be worth reading up on.

To be honest I don't know which is the more damaging, to keep them powered on - Bipolar PROMs which are working perfectly normally run alarmingly hot and you have to think that can't do them any good in the long run, but on the other hand it's often the case that if a device is going to fail it is at the point of power-on, so powering them on thousands of times a second doesn't seem like a good idea either.

I do wish there was a pin compatible low-power modern equivalent which we could just drop in instead.

ortek_service · 23rd Nov 2022, 9:50 am

Quote:

Originally Posted by SiriusHardware

>>
I do wish there was a pin compatible low-power modern equivalent which we could just drop in instead.

Yes, I have just seen these interesting range of ready-made adaptors, to use more modern FLASH / OTP-(E)PROM's: http://callanbrown.com/index.php/fle...ement-adapters

But they've had to use much-larger (capacity) ones, to match the high Bipolar PROM speeds, so the adaptor is a bit unwieldy and may prone to getting knocked out of main board's socket.

SiriusHardware · 23rd Nov 2022, 10:07 am

In our application(s) the EPROMs used to substitute them don't actually have to be as fast as Bipolar PROMs. You can use an old 2716 to substitute for both PROMs, suitably wired in, in the MK14 for example.

There are some PROM applications which absolutely did depend on the higher speed of PROMs vs. contemporary EPROMs, such as TV pattern generators and so on.

ortek_service · 24th Nov 2022, 3:23 am

Yes, before I bought some Tesla PROM's for an MK14 replica I was assembling, I'd planned to initially use the old favourite (and usually relatively slow upto 450ns) 2716 in an adaptor in one PROM socket (and 4 wires for the other nibble into the other one, as the upper & lower nibble memories are unusually not adjacent to each other on this).
And having an 8bit EPROM for each nibble isn't very efficient, that you'd end up with if using the 4bit PROM version of those universal ready-made adapters.

However, for those without a good collection of old now-obsolete EPROM's - Although places like this still have them not too expensive: https://www.laser.com/index.php/cate.../cPath/288_290
then a more modern, higher capacity, one may be preferable as not that much large in DIL28 etc. / can get some in smaller Surface Mount packages
- I recall Psion Organiser 2 had 27C256 OTP-EPROM's in small SO28W, but smallest parallel FLASH seems to be in PLCC packages.

Plus it seems that many cheap PC-Programmers now don't have very high Vpp voltages / Parallel-FLASH (or EEPROM) IC's don't require high Vpp or UV-erasing if you want to amend contents.

As it happens, the first 'high-speed' (120ns? CMOS) EPROM's I bought were for an ETI TV Pattern Generator / Testcard project by a certain Paul Stenning, that I recall you had to send to him with a small fee to get programmed.

20th Nov 2022, 12:14 pm	#21
Buzby123 Heptode Join Date: Oct 2011 Location: Culcheth, Cheshire, UK. Posts: 649	Re: MK14 SN74S571 PROM Power reduction test In the thread about the waalsdorp sc/mp ( https://www.vintage-radio.net/forum/....php?p=1514816 ) there was a link to a 1976 publication showing a SC/MP system. Also in that mag was an article titled 'High Performance, Low Power Memories from Inexpensive Parts!' It's just another circuit to stop your PROMs getting too hot !. Attached Thumbnails

20th Nov 2022, 12:47 pm	#22
ortek_service Octode Join Date: May 2018 Location: Northampton, Northamptonshire, UK. Posts: 1,435	Re: MK14 SN74S571 PROM Power reduction test Yes, I recall previously seeing this - Probably when I first discovered these "Compute newsletters" a while ago. It does seem rather reliant on these PROM's being happy with no power supply, whilst still in-circuit. Many IC's would have got parasitically-powered from the data/address lines etc that could result in latch-up when the power was applied and possibly not working getting rather hot (I once encountered that with PLL IC's, where the supply didn't rise quick enough compared to voltages on input pins, as supply was a simple Shunt zener stabiliser and series R + a few uF's capacitance across the zener caused an RC time-contant delay. So I had to put some resistors in series with the inputs, to restrict the I/P current below that which could cause latch-up) Many years ago, I heard about a RAM IC that died after a few months use, and they then discovered (Fortunately only on prototypes) its power rail had never been connected (The Mentor CAD system put power-rails as a separate block, that was probably on another sheet, and it seems any DRC warnings were ignored / disabled)

20th Nov 2022, 1:20 pm	#23
SiriusHardware Dekatron Join Date: Aug 2011 Location: Newcastle, Tyne and Wear, UK. Posts: 11,546	Re: MK14 SN74S571 PROM Power reduction test The article makes this happy behaviour of the DM (National Semiconductor) devices out to be a distinct virtue of theirs - you would instinctively expect that if you had a 'dead' (unpowered) device sitting on your data lines you would be in for all sorts of trouble. It seems unlikely that pin compatible devices from other manufacturers (AMD, Philips/Signetics, Tesla) would behave in such a similarly convenient fashion, but it may be worth reading up on. To be honest I don't know which is the more damaging, to keep them powered on - Bipolar PROMs which are working perfectly normally run alarmingly hot and you have to think that can't do them any good in the long run, but on the other hand it's often the case that if a device is going to fail it is at the point of power-on, so powering them on thousands of times a second doesn't seem like a good idea either. I do wish there was a pin compatible low-power modern equivalent which we could just drop in instead.

23rd Nov 2022, 10:07 am	#25
SiriusHardware Dekatron Join Date: Aug 2011 Location: Newcastle, Tyne and Wear, UK. Posts: 11,546	Re: MK14 SN74S571 PROM Power reduction test In our application(s) the EPROMs used to substitute them don't actually have to be as fast as Bipolar PROMs. You can use an old 2716 to substitute for both PROMs, suitably wired in, in the MK14 for example. There are some PROM applications which absolutely did depend on the higher speed of PROMs vs. contemporary EPROMs, such as TV pattern generators and so on.

24th Nov 2022, 3:23 am	#26
ortek_service Octode Join Date: May 2018 Location: Northampton, Northamptonshire, UK. Posts: 1,435	Re: MK14 SN74S571 PROM Power reduction test Yes, before I bought some Tesla PROM's for an MK14 replica I was assembling, I'd planned to initially use the old favourite (and usually relatively slow upto 450ns) 2716 in an adaptor in one PROM socket (and 4 wires for the other nibble into the other one, as the upper & lower nibble memories are unusually not adjacent to each other on this). And having an 8bit EPROM for each nibble isn't very efficient, that you'd end up with if using the 4bit PROM version of those universal ready-made adapters. However, for those without a good collection of old now-obsolete EPROM's - Although places like this still have them not too expensive: https://www.laser.com/index.php/cate.../cPath/288_290 then a more modern, higher capacity, one may be preferable as not that much large in DIL28 etc. / can get some in smaller Surface Mount packages - I recall Psion Organiser 2 had 27C256 OTP-EPROM's in small SO28W, but smallest parallel FLASH seems to be in PLCC packages. Plus it seems that many cheap PC-Programmers now don't have very high Vpp voltages / Parallel-FLASH (or EEPROM) IC's don't require high Vpp or UV-erasing if you want to amend contents. As it happens, the first 'high-speed' (120ns? CMOS) EPROM's I bought were for an ETI TV Pattern Generator / Testcard project by a certain Paul Stenning, that I recall you had to send to him with a small fee to get programmed.