Williams NVRAM adapters

delroy666

Well-known member

Donor 2011
Joined
Jul 30, 2009
Messages
1,189
Reaction score
26
Location
Ottawa, Ontario, Canada
Rating - 100%
29   0   0
I had a few people ask me about these lately, so I made another small batch of NVRAM adapters for Williams vid boards. These are intended to eliminate the problem of intermittent settings and high score loss/corruption that sometimes occurs when using a switcher instead of the original linear power supply in Williams games such as Robotron, Sinistar, Joust, Stargate and Splat. (not Defender) Details:

- It uses an ST M48Z02 internal battery backed SRAM chip which write-protects itself when the power supply drops below 4.75V, so it's unaffected by the CPU writing out garbage during power off. According to the datasheet, the internal lithium battery should retain the memory for > 10 years. It also eliminates the need to use the original AA battery holder (or lithium coin cell) on the game PCB.

- To install it, just replace the original CMOS chip (e.g. 5114, 5514, 6514) with the included DIP socket and plug the adapter in to the socket.

- The 2-pin header is for a jumper which can be used to switch between the lower and upper bank of memory, so two sets of high scores and settings can be stored and accessed on the same chip. For example, install the jumper and set the game for "tournament" settings; remove the jumper and set the game for "marathon" settings – you can then use the jumper (or wire up a toggle switch) to easily switch between the two modes and the high scores associated with each.

- The 4 extra pads on the adapter are routed to data bits D4-D7 of the NVRAM, so this adapter could be used on a Bubbles board if you wire these pads to the corresponding data pins of the second CMOS RAM socket on the game PCB.

They are $30 each including shipping to the U.S. or Canada. This includes the adapter, NVRAM chip, a jumper and an 18-DIP dual wipe socket.
 

Attachments

  • Williams NVRAM adapter.JPG
    Williams NVRAM adapter.JPG
    54.9 KB · Views: 277
Last edited:
These eliminate the problem of intermittent settings and high score loss/corruption when using a switcher instead of the original linear power supply in Williams games such as Robotron, Sinistar, Joust, Stargate and Splat. (not Defender)


I dont want to harsh on your FS thread but that part of the thread is not correct. The issue with the switching power supplys has nothing to do with the RAM settings and score corruption. It has to do with the way the data is written and the timing of the power on/off.

That said, this thing is great. I have one in my Robotron and it works to save your settings and having two high score/options settings is awesome. Worth every penny and more IMO.

-Ken

PS Have Dokert rebuild your original PS and you'll be much better off than with a switcher anyway.
 
I dont want to harsh on your FS thread but that part of the thread is not correct. The issue with the switching power supplys has nothing to do with the RAM settings and score corruption. It has to do with the way the data is written and the timing of the power on/off.

That said, this thing is great. I have one in my Robotron and it works to save your settings and having two high score/options settings is awesome. Worth every penny and more IMO.

-Ken

PS Have Dokert rebuild your original PS and you'll be much better off than with a switcher anyway.

You're right, Ken. I didn't mean to imply that this is a problem with every switcher. It totally depends on when the 12V goes down compared to the 5V during power off, and therefore whether or not the CMOS gets deselected soon enough by the monitoring circuit. I've updated the wording in my original post.

And yeah, if you have the original linear supply, rebuild it! Unfortunately many cabs have been stripped of the power PCB and/or transformer assembly though.
 
Is this a viable option for people like myself who have williams power supplies but want to replace the AA batteries with something better?
 
Are these fine with the linear power supplies to make setting and high score saving more reliable? Need something or my Defender and Sinistar.
 
Are these fine with the linear power supplies to make setting and high score saving more reliable? Need something or my Defender and Sinistar.

With my Dokert rebuilt power supply and this adapter I have had a solid score/settings table for well over a year.

I switch it between tournement mode and marathon mode on occasion to keep the score tables seperate. I ran a jumper to the inside of the coin door so I swap the jumper and hit the settings button and off I go.

-Ken
 
Are these fine with the linear power supplies to make setting and high score saving more reliable? Need something or my Defender and Sinistar.

It will work in Sinistar regardless of which type of power supply it has, but not in a Defender. Defender uses a different CMOS RAM chip (5101) than most other Williams vid boards.

I know that Lindsey makes 5101 adapters for pinball boards, (http://pinforge.com/) and I don't see why one of those wouldn't work in a Defender. The Ramtron FRAM part he uses doesn't appear to have an automatic power-fail deselect feature like the M48Z02 has, but if you're using the original linear supply it shouldn't matter. Maybe send him a PM to see.
 
Crazy... I was just about to make some of these. Glad I checked here first. Might make some for myself anyway.

FWIW: My 5101 adapter does work with Defender.

EDIT:
The RAM going write protected below 4.75 should help with corruption but you'll want to jump around the blocking diode so you get the full +5V to the new RAM. That diode could cause as much as .5V drop causing problems with the NVRAM.
 
Last edited:
Crazy... I was just about to make some of these. Glad I checked here first. Might make some for myself anyway.

FWIW: My 5101 adapter does work with Defender.

EDIT:
The RAM going write protected below 4.75 should help with corruption but you'll want to jump around the blocking diode so you get the full +5V to the new RAM. That diode could cause as much as .5V drop causing problems with the NVRAM.

I did the PCB layout for an adapter using a Ramtron part a while ago, but never got it made since I didn't know how well it would work with a switcher on these boards if it lacks a built-in power-fail deselect. It seems it would still be susceptible to data corruption during power down if the 12V did not drop first. For those running the original linear supply though, the Ramtron is ideal. It's cheaper, doesn't need a battery, retains data longer, is rated for tons of write cycles, etc.

I thought the same about that diode (D9) too, but so far I haven't heard of anyone needing to jump it. On my Joust, the 5V supply is pretty low (4.90V) and I still didn't need to bypass it. I only measured a 0.08V drop across the diode, so the voltage at the NVRAM is 4.82V which is still above the VPFD range.
 
I did the PCB layout for an adapter using a Ramtron part a while ago, but never got it made since I didn't know how well it would work with a switcher on these boards if it lacks a built-in power-fail deselect. It seems it would still be susceptible to data corruption during power down if the 12V did not drop first. For those running the original linear supply though, the Ramtron is ideal. It's cheaper, doesn't need a battery, retains data longer, is rated for tons of write cycles, etc.

I thought the same about that diode (D9) too, but so far I haven't heard of anyone needing to jump it. On my Joust, the 5V supply is pretty low (4.90V) and I still didn't need to bypass it. I only measured a 0.08V drop across the diode, so the voltage at the NVRAM is 4.82V which is still above the VPFD range.

Cool. Looks like you did some good work on these.

I'm going to make some using Ramtron parts so I'll report on how they work out. I've got a Joust with a switcher to test with. I've got a crap-load of Williams boards to sell next year. I'm going to make a single EPROM board for the ROM boards too. Mainly because I can really use all those damn 2732s for other projects.
 
Got mine and it looks great!
Even included the socket.
I should be installing this bad boy tonight and we'll see how it goes.
Jeff
 
I have a joust that has a linear power supply with new 4164(?) ram with new cables but still doesn't have free play or credits.

Will this work even with a linear ps and have FP plus HSS?

Thanks
 
Yea boy!

All's well with my Bubbles, no more "AA"s.

Thanks
Jeff

Cool, I'm glad to hear it worked out.

I have a joust that has a linear power supply with new 4164(?) ram with new cables but still doesn't have free play or credits.

Will this work even with a linear ps and have FP plus HSS?

Thanks

Joust normally saves high scores and pricing selection settings (free play) to its CMOS SRAM at location 1C on the CPU board. If yours doesn't, and is using the original linear supply, it's possible the SRAM chip is bad, the AA batteries are dead, or the AA battery holder is corroded. You could likely just get a replacement SRAM chip and/or battery holder (AA or lithium coin cell) and that would solve your problem. If you didn't want to do this for whatever reason, the NVRAM adapter would also be a solution. It does work fine with both linear and switching supplies.
 
Cool, I'm glad to hear it worked out.



Joust normally saves high scores and pricing selection settings (free play) to its CMOS SRAM at location 1C on the CPU board. If yours doesn't, and is using the original linear supply, it's possible the SRAM chip is bad, the AA batteries are dead, or the AA battery holder is corroded. You could likely just get a replacement SRAM chip and/or battery holder (AA or lithium coin cell) and that would solve your problem. If you didn't want to do this for whatever reason, the NVRAM adapter would also be a solution. It does work fine with both linear and switching supplies.

Well I have a lithium coin upgrade on it already and I have several of those Dallas ds1220AD NVSRAM, but I do not have them installed yet!

can i use this Dallas and solve my problem or do i need your NVRAM adapter?
 
Last edited:
Back
Top Bottom