Williams NVRAM Question

Anthony605

Well-known member
Joined
Nov 13, 2013
Messages
1,299
Reaction score
466
Location
HOWELL, New Jersey
There is a seller on EBAY selling NVRAM kits for Williams games. I will post the link below this post. Has anyone used this kit? I have a Robotron with a switching power supply and the lithium battery MOD. It only sometimes saves high scores. If these chips work well I may consider getting one. Also, if the kit works well does anyone sell them on KLOV? Here is the link....

http://www.ebay.com/itm/5114-5514-6...obotron-Joust-Berzerk-and-More-/201648291744?
 
Hi Klov,

I am mostly a pinball guy, but my NVRAMs will work in many other applications including arcade games.

The device FM16W08 is used on all my adapters. It is 70ns rated FRAM (FLASH) chip that acts like a CMOS ram in most situations. It has high enough endurance to last 190 years of read/writes at 1MHz. Holds memory for 10+ years between power cycles.

The main thing you need to pay attention to when using FM16W08 based adapters to replace other battery backed CMOS RAM is that that FM16W08 FRAM chip must have the CE pin cycled for every read/write or it will not work. All pinball applications seem to do this, but i have a report PAC-LAND failed to work with my 6116 adapter. Reviewing the schematic I see why. Midway got lazy with ram and just uses the R/W pin (i think this wastes energy and makes the RAM run hotter).

You can see on the attachment 1 here from PAC-LAND. (2018 is 6116). that the CE pin is just tied to ground. This is not going to work with FM16W08 NVRAM

You can see on attachment two from Bally Pinball 6803 MPU. The CE is is attached to a A LS10 logic gate instead of just grounded. This is what you want to see when using NVRAM.

Any questions about compatibility, I can tell from a schematic in most instances if a NVRAM based off of the FM16W08 is going to work. I didn't even really think about this being an issue because all pinball boards cycle CE for the RAM chips.

As far as Robotron and switching power supplies. I honestly do not know and didn't even think about it. I would assume if the CMOS RAM works with a switching supply, than the FM16W08 based devices would as well. I suppose how fast the rails collapse could effect things. My testers used games with original linear power supplies. I am using the wide voltage part. It has a workable voltage range of 3.3v to 5v, that may help problem boards over say the FM1608 which is a 5v only part.

http://nvram.weebly.com

Andrew
 

Attachments

  • at 1 pacland ram.png
    at 1 pacland ram.png
    21.9 KB · Views: 44
  • at2 bally 6803 mpu.jpg
    at2 bally 6803 mpu.jpg
    56.4 KB · Views: 39
Last edited:
I solved my robotron/switcher reset issue by installing a nvram module. I dont remember where it came from since it was a while ago, but it does work.
 
when using regular batteries, CMOS corruption is normal. I don't really know the logistics behind it, but considering the linear power supplies hold 12 long after the game has been turned off, I'm guessing that has something to do with it.

a switcher by comparison drops voltages backwards from the linear, 12 and -5 will go first, then 5.

that said, I had 4 Williams games on switchers last year and they didn't puke up the CMOS once. I then proceeded to install properly rebuilt linears and remote battery holders.

if NVRAM and switchers work for the rest of you then great.
 
The linear power supply is going to have a very large filter cap on the 12v line to smooth for the 5 regulator (lm323k in pinball). It can take some time for this cap to bleed off.
 
I've put NVRAMs on the troublesome WMS games that have come through my shop over the last year or so. High score table corruption was fixed by doing so.
 
Same here, my Robotron with switcher was losing both settings and high scores every few days until I put in the NVRAM, it's been perfect ever since. I got mine from a member here Klov, I think it was delroy666.
 
I bought the NVRAM that I mentioned in my first post (EBAY Link) and all is good. My Robotron saves settings and high scores perfect now. The seller is also a super nice guy! I would recommend picking one up.
 
I bought one of Andrew's NVRAMs for my Joust and it worked great. The price point is real nice too.

barakandl - Does your 5101 replacement work with Gottlieb System 1 games?
 
No. Unfortunately 5101 adapters will not work in gottlieb system 1 pinball games.

The 5101 in that setup has Data in and data out pins separated. Most of the time Do and Di are hard tied together in 5101 applications. This is the only setup I know of where they are not.

I would need to do an extra special adatper for gtb sys 1 to deal with data in and out and latching them with some kind of logic. Probably not going to happen.

These games actually work ok with no 5101 ram installed. It just does book keeping, no game play ram stuff.
 
I wonder if people having setting / high score issues when using a switching supply is because the CMOS RAM CE high signal is derived the power on reset delay circuit, which uses the unregulated 12v. Perhaps this voltage comes up and/or collapses at a rate where the CE is invalid and the RAM contents get corrupted. In a linear supply, the 12v unregulated goes through a regulator to create the +5v supply. In a switching supply, they are not related in the same way.

The 5114 RAM Vdd is the same the as the systems Vdd. Perhaps a pullup resistor could fix by holding the CE high if the 12v circuit isnt started up yet.

CMOS RAM or Flash RAM having trouble. Try a 4.7k resistor between pin 18(Vdd) and pin 8 CE of the 5114 RAM.
 
I bought an NVRAM adapter for my Robotron which uses a switching power supply. I did this because with batteries I was having issues.

Unfortunately, the NVRAM sort of worked for a while, but it got corrupted a few times and it was a major pain in the butt to reset them. I finally had to pull it and go back to batteries. I've since then added a large capacitor across the +12VDC, and haven't had an issues since (and I've been booting up the game almost nightly as of late). I'll try switching in the NVRAM again since it appears the large capacitor may have resolved things, but the fact that they're such a pain to reset after getting corrupted, I may just stick with batteries.
 
Your reply is helpful. That helps point that the issue is with the unregulated 12v supply. Common connection is the RAM CE High Signal (which HIGH protects the RAM from being written too) comes from the power on reset delay which is also uses the 12v supply.

In a linear supply the 12v rail is going to collapse much slower than the 5v, the filter cap holds it high enough to probably keep the CE high at power down until the 5v rail is down all the way.

How large of a capacitor on the 12v supply did you use? 1000uF 4700uF might be big enough. WMS probably has a 12000uF cap on a linear power supply.

" I may just stick with batteries."
Batteries kill good CPU boards. Be careful.
 
Last edited:
Sure that makes sense. How much capacitance that put on the 12v rail is going to vary switcher supply to switcher supply. May not need even need 12k uF.

Thank you for the reply.
 
My Robotron has been working perfect with an NVRam bought from a member here, although it already had a 4700uf capacitor on the switcher 12v when I bought it so maybe that's helped out.

My Defender likewise has been working great with an NVRam for a few months now but finally got a corrupt high score entry the other day. It's a PC power supply in that one so there is no 4700uf capacitor on it like my Robotron had. I took a measurement at the ram chip (I have the +5v ram mod on both my Robotron and Defender) and on Defender noticed it was high at 5.34v. It made me wonder if it's being that high means that in some rare circumstances the NVram write protect maybe wouldn't trigger in time and hence still lead to corrupt high scores on a rare occasion. I since tweaked it so it's now at 5.03v measured at the ram chip, hopefully that will sort that issue out.
 
I put these in several of my Williams games (Joust, Sinistar, Joust 2, and even in my Inferno), and so far, they work great. I even helped a friend install them in his Sinistar and Joust uprights just this past weekend, and again, they worked just fine.

The only reason I didn't put them in my other Williams cabs (Robotron, Stargate, and Bubbles) is that they are running the JROK board ;) .

Jon
 
Back
Top Bottom