UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Reply
 
Thread Tools
Old 1st Apr 2024, 11:20 pm   #1
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Problems using FRAM with retro computers

There were recent posts in other threads recently about problems using FRAM in place of battery backed SRAM or Proms with retro computers. I thought I’d start a new thread with some issues found in the datasheet for FM18W08, attached for reference. I think these would also apply to FM1608, FM16W08 and FM1808.

Note the W indicates a wider voltage supply range 2.7v to 5.5v.

1. The address is latched on the falling edge of chip enable, so on the INS8060 the chip enable should be qualified with write enable or read enable, and not only the address lines.

2. Although the read or write cycles are only 70ns, there is also a required minimum precharge time between each access. Chip enable must remain high for at least 70ns. Probably not a problem for the 8060, but in a z80 system this may need a delay at the start of each memory access.

3. Input logic levels are cmos thresholds, logic high minimum of 0.7 x Vcc might not be reached by some older devices or 74ls memory decoding. This might be fixed by adding pullup resistors on each address and data line.

4. During power fail the processor may write random data to random locations. These FRAM devices do not disable chip enable when Vcc drops.
Mark1960 is offline   Reply With Quote
Old 2nd Apr 2024, 12:01 am   #2
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

There is also a cryptic note under Pre-Charge Operation. “However the device has a maximum CE LOW time specification that must be satisfied.” I couldn’t find where that maximum time is specified, or what happens if it is exceeded. This could be a problem if the processor is running single step mode using the processors wait input, for example the scrumpi system of editing ram.
Mark1960 is offline   Reply With Quote
Old 2nd Apr 2024, 12:26 pm   #3
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,490
Default Re: Problems using FRAM with retro computers

Quote:
Originally Posted by Mark1960 View Post
There were recent posts in other threads recently about problems using FRAM in place of battery backed SRAM or Proms with retro computers. I thought I’d start a new thread with some issues found in the datasheet for FM18W08, attached for reference. I think these would also apply to FM1608, FM16W08 and FM1808.
>>
>>

3. Input logic levels are cmos thresholds, logic high minimum of 0.7 x Vcc might not be reached by some older devices or 74ls memory decoding. This might be fixed by adding pullup resistors on each address and data line.

4. During power fail the processor may write random data to random locations. These FRAM devices do not disable chip enable when Vcc drops.

Yes, I've been meaning to have a look into this again, once I get the time with playing with GAL's and DS1287Y NV-RAM Modules (Where I discovered issues with certain programmers, I'll post about sometime)

I did read many arcade machine owners had success swapping their RAM IC to an FRAM, to eliminate potentially-leaking NiCd Backup Batteries, and surprisingly it seemed these all seemed to work OK. With most concern being about writes endurance, as it seems the current score etc may be continuously written to that RAM IC. And earlier versions of the FRAM's had rather lower (still billions?) max. no. of writes compared to trillions in the later ('B'?) revised versions (That may only be available in SM-packages)

The Address & Databus will often be driven directly by an NMOS Processor, so logic-levels will probably be closer to CMOS and a bit better than TTL. So probably get away with it a lot of the time (although with reduced noise-margin)


Once you introduce any form of NV-Memory, that is writeable by the processor, then it is usually advised to incorporate some Low-Voltage Inhibit (LVI) protection against the processor execution random instructions when powering-down or up. This can be done by using a 'reset IC' with built in voltage detection of its supply / external inputs.

Or inhibiting Chip-select / Write to the Memory, when voltage isn't high-enough for correct operation (some memory devices may have this built-in, especially SRAM that goes to a low-power state when operating on a reduced-voltage battery-supply)
ortek_service is offline   Reply With Quote
Old 2nd Apr 2024, 3:07 pm   #4
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

Pre-charge time is probably less of a problem on the 6800/6502/6809 memory timing as these are generated from system clock with 50/50 duty cycle. Provided the system clock is also used to gate the chip enable and not just the write enable and output enable.

Slower z80 and z80a might also be ok, I might take another look at the timing to see what frequency clock on a z80 might be a problem.

I prefer to use cmos ram with max818 or similar to control reset and write protect, rather than the nvram modules with built in battery that is not easily replaced. I probably need to look at the max818 datasheet again to see if it could be used without a battery to control write enable to FRAM.
Mark1960 is offline   Reply With Quote
Old 2nd Apr 2024, 3:22 pm   #5
Phil__G
Octode
 
Join Date: Mar 2011
Location: North Yorkshire, UK.
Posts: 1,151
Default Re: Problems using FRAM with retro computers

I've only ever used them sucessfully on the PICL where they've behaved flawlessly in heavy use, both Cypress & the ramtron FM16W08's.
Curiously these FRAMs are often referred to on ebay as 'battery eliminators'
Mine came from 'dabbler51'.
BUT I tried them on my 7.37Mhz Grante Searle Z80s and they wouldnt work at all, there are speed concerns as Mark has found
Somewhere in my stash I have a few 2k x 8 frams, 6116 equivalents.

Last edited by Phil__G; 2nd Apr 2024 at 3:32 pm.
Phil__G is online now   Reply With Quote
Old 2nd Apr 2024, 6:35 pm   #6
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

I checked the z84c00 spec for the z80 and this shows one complete T cycle between consecutive memory cycles, approx 130ns at 7.37 MHz, so it could be something else causing the problem with the Grant Searle z80.
Mark1960 is offline   Reply With Quote
Old 2nd Apr 2024, 6:40 pm   #7
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

Grant’s z80 schematic here shows the chip enable is not gated by MREQ, so this would explain why the FRAM is not working in that design.
http://searle.wales/z80/SimpleZ80.html
Mark1960 is offline   Reply With Quote
Old 2nd Apr 2024, 6:55 pm   #8
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,324
Default Re: Problems using FRAM with retro computers

I have recently used an FRAM chip to successfully replace the non-volatile RAM on two resistivity machines for my Archaeology club. The original chips had potted modules containing the volatile memory and a lithium battery, ie non replaceable and with a shelf life of 10 years. The original non-volatile rams are still available at £94 plus VAT from RS.

You have probably seen these type of machines on Time team, they are an H frame with 2 probes which you push into the ground at 1m intervals.
They use a 6502 processor and the FRAM works perfectly.

Peter
Electronpusher0 is offline   Reply With Quote
Old 3rd Apr 2024, 12:05 am   #9
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

Hi Peter, thanks for noting a success using FRAM. Is it possible you have a schematic of the memory address decoding?
Mark1960 is offline   Reply With Quote
Old 3rd Apr 2024, 10:06 am   #10
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,324
Default Re: Problems using FRAM with retro computers

Quote:
Originally Posted by Mark1960 View Post
Hi Peter, thanks for noting a success using FRAM. Is it possible you have a schematic of the memory address decoding?
Sorry, I do not have the schematic for the machine, it was a case of suck it and see. It certainly seemed to work with no issues, it kept all the setup and survey data when powered off.

Peter
Electronpusher0 is offline   Reply With Quote
Old 3rd Apr 2024, 10:13 am   #11
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,324
Default Re: Problems using FRAM with retro computers

I should have said that I used a surface mount device with an SOP28 adapter pcb.
If you want to buy one to try look for sellers offering "Nintendo 64 N64 FRAM Memory PAK Upgrade", this is just an FM28V020-SG for £6.99

Peter
Electronpusher0 is offline   Reply With Quote
Old 3rd Apr 2024, 10:26 am   #12
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,490
Default Re: Problems using FRAM with retro computers

Quote:
Originally Posted by Electronpusher0 View Post
I have recently used an FRAM chip to successfully replace the non-volatile RAM on two resistivity machines for my Archaeology club. The original chips had potted modules containing the volatile memory and a lithium battery, ie non replaceable and with a shelf life of 10 years. The original non-volatile rams are still available at £94 plus VAT from RS.

You have probably seen these type of machines on Time team, they are an H frame with 2 probes which you push into the ground at 1m intervals.
They use a 6502 processor and the FRAM works perfectly.

Peter
I've never seen these NV-RAM that expensive - I thought the DS1287Y 128KB one at £34 + VAT was rather steep at £34+VAT usual price.
You must be able to bulid an equivalent with SMD RAM + Battery-switchover controller IC's much-cheaper, as that size RAM isn't usually very expensive these days. And designing one with a con-cell holder in to, so it can be easily changed would be a much better idea than having to throw the whole module away every 10yrs or so.
- Although that may mean you can pick-up used ones quite cheap, and hack into the casing to replace the the battery with a coin-cell holder etc.
ortek_service is offline   Reply With Quote
Old 3rd Apr 2024, 10:44 am   #13
Phil__G
Octode
 
Join Date: Mar 2011
Location: North Yorkshire, UK.
Posts: 1,151
Default Re: Problems using FRAM with retro computers

The ASC 128k's are dirt cheap and ds1210's can be found for next to nowt, I've used this combo many times:

https://www.mccrash-racing.co.uk/philg/nvram/nvram.htm
Phil__G is online now   Reply With Quote
Old 3rd Apr 2024, 3:28 pm   #14
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

FM28V028 operates slightly different to the 5v FRAM devices. It has a recommended Vcc of 2.0v to 3.6v and absolute max of 4.5v, so may not be suitable for retro computers with Vcc at 5.0v. The ground survey systems are probably battery operated and may be using a cmos 6502 at 3.3v. The FM28V028 does not latch the address on the falling edge of chip enable, but does still require a pre-charge time, internally controlled when the address lines change.
Mark1960 is offline   Reply With Quote
Old 4th Apr 2024, 2:12 am   #15
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,490
Default Re: Problems using FRAM with retro computers

I presume you meant FM28V020, as mentioned by Electronpusher0 - rather than FM28V028, as that doesn't seem to exist!


Despite the pat number, which looks a bit like 2Mbit (E)EPROM etc numbers, these are only actually 256kbit = 32KByte devices. So if the original 32KB NV-RAM's are > £100 then those are ridiculously-expensive.

And yes, these are really only for use on 3.3V etc logic/supply voltage systems, and abs.max ratings are those which exceeding could permanently damage it.
Exceeding its max operating voltage is not recommended, as they won't guarantee it will function correctly above that.


If you do need 5V ones, then might as well just buy used or NOS (as no-longer made in DIL for these) FM1608 (8KB) or FM1808 (32KB) - Also not very logically numbered, for memory devices.
- I recently bought some of these from Shop2943123 Store on Aliexpress for 90p / £1.38 each (Plus around £1.90 delivery on each type), and although they appeared to be 'pulls' as had data already in these (With RICOH_MP-1600L / 1800 mentioned at start of data), they all erased / re-prog'd OK.

Or you can buy the later 'B' more write-endurance version SM ones, mounted on a DIL-Adapter off eBay etc but do pay more for those.

Although if you do actually want a 3.3V / low-voltage 32KB one, then £6.99 for just a FM28V020-SG SO-28W part isn't too bad as a few pounds more at Farnell / Digikey etc.
ortek_service is offline   Reply With Quote
Old 4th Apr 2024, 3:39 am   #16
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,317
Default Re: Problems using FRAM with retro computers

Yes I did mean FM28V020.

I have a few FM18W08 in SOP28, for possible use with a qfp or plcc z80.
Mark1960 is offline   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 12:38 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


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