Some reading materials on reset issues:

https://pinside.com/pinball/forum/topic/wpc-reset-my-voltage-is-55v

https://pinside.com/pinball/forum/topic/random-reset-on-a-bsd

https://pinside.com/pinball/forum/topic/dracula-reset#post-3542008

https://pinside.com/pinball/forum/topic/no-fear-wpc-resetting-woes#post-3182264

https://pinside.com/pinball/forum/topic/bsd-resetting-issue#post-1799061

Maybe one of these will help you find your silver bullet

This may be your magic bullet: http://www.kahr.us/daughterboard.html $30 and your reset issues are likely solved.

The usual cause of the WPC resets is low voltage on the +5V going to the MPU board, which is VERY sensitive to low voltage. Kahr's board basically has a power regulator that taps into the system's +12V to create a new +5V current that goes straight to the +5V MPU board. Otherwise, whenever I get a WPC game, I replace the 5V voltage regulator with this: http://www.ezsbc.com/index.php/products/psu5.html and, while I have the board out, the +12 voltage regulator with this: http://www.ezsbc.com/index.php/featured-products-list-home-page/psu3-12.html. I also replace the capacitors as they only have ~1000 hours of life and most WPC games are in need of those.

Of course, I highly recommend going through pinwiki's breakdown of the WPC reset problem: http://pinwiki.com/wiki/index.php?title=Williams_WPC#Game_resets

However, if you just want to get your game up and running again, Kahr's board is the easiest way to go.

Resets are just low +5V and the watchdog circuit kicking in. The kahr board I would only use in a pinch to keep a game going for whatever reason. It doesn't 'fix' anything, just provides a work-around.

Here is a Readers Digest version of what would be the main culprits in order or probability. If you start changing out components you can easily compromise plated through holes and may even make the problem worse or make the game not even boot-up.

1. Bad J101, J102, J114 (on PDB) and J210 (on CPU). If you have a game with a Z-connector between J114 and J210, that can be an issue too. Best to change both the header pins and connectors. If you reseat these connectors and the game seems to magically appear to be fine. It's just a matter of time before the resets come back, but it lets you know it's a connector issue.

2. There are two electrolytic caps that filter the +5V. That would be the next area to tackle. The first is C5 (15,000uF, 25V, radial) and C4 (100uF, 16V, axial). Be careful with C5 as it's a snap-cap and removing it can compromise the plated through hole to BR2.

3. Voltage regulator Q1. If after doing 1-2 above you are still running low, I'd look at the ezsbc LM323 voltage regulator linked by lyonsden above. It has an adjustable pot to adjust the voltage output (adds resistance between the case and ground) but already comes preset at 5V. It's nice because it produces no heat.

4. BR2 would be the last thing I'd change. Again be careful as to not compromise the plated through holes.

After tightening mounting screws to any TO-3 or TO-66 case regulator/transistor, you should resolder the two leads to relieve any stress on the parts. You don't want to force the top and bottom pads to separate from the feed through channel.

Game is still working?

Wrong. My board fixes the system design flaw in WPC power loading.

mcklit I invite you to read more at http://www.kahr.us/index2.html . The simple truth about WPC is power supply loading was only marginally considered when the platform was first designed and was not reconsidered with each title released. Like so many other design elements of the time, cost and speed to market were the primary drivers. Consider:

- The power demands are grossly unbalanced across the available supplies. The 12v digital has a trivially light load whereas the 5 volt supply is pretty much maxed out even with brand new components.
- The stressed 5v line has is very little "head-room" in the derivation which means the circuit has very little tolerance to components becoming weaker with age/use, ability to absorb surges or sags in load as game conditions change, or for the myriad of other contributors to the emergence of resets (i.e., low line voltage, grounding path,...).
- The components in the 5v line and the watchdog itself have wide tolerances (~10%) for acceptable factory performance (when new!) which statistically translates into certain PDB/MPU combinations only delivering at or slightly above the watchdog trigger voltage under the BEST of conditions. If you are unlucky enough to have such a pairing resets are a certainty over time.
- The distribution of the voltage through the machine was not well thought out (think Z connectors).
- None of the above are independent variables; they interact continually.

My daughterboard intentionally implements a carefully thought out system-level redesign that simultaneously addresses all of these concerns. It reduces the loading on the 5 volt circuit by moving the MPU load to the 12 volt line. This reduces stresses (i.e., from heat) on the components in the 5 volt derivation because they are no longer driven to their limits. It delivers a clean, stable, and relatively precise (within 2%) voltage directly to the point of consumption (the MPU) bypassing the problems introduced by the distribution headers. It's power is sourced from a line with plenty of spare capacity (analysis on my weblink above) with huge "head-room" (12v is derived from 18v) and shared with peer circuitry that is not sensitive to precise voltage levels (lamp columns). It is built using a high efficiency switching regulator that minimally loads the 12 volt line beyond the actual load introduced by the MPU itself.

Think more broadly than about "just low +5V and the watchdog circuit kicking in" when considering the reset symptom and you will come to the conclusion that every WPC should have my system-level improvement installed - whether or not it is presently exhibiting the reset symptom.

There are literally thousands of deployed use cases in commercial and residential environments.

-Rob
-visit http://www.kahr.us to get my daughterboard that helps fix WPC pinball resets or for my Pinball 2000 H+V video sync adapter kit

rkahr
The product you offer to provide stable +5V to the MPU works very well and is a good work around to help a large number of owners that don't have the proper background or skills to properly repair their machines.

While I'm not trying to be argumentative, this truly is a workaround in every meaning of the word as it in no way addresses any of the underlying issues which caused the reboot, these issues will likely still exist even after the system no longer reboots.

I would like to touch on a couple of statements you made that are deceptive/inaccurate.

While many of the components around the 5 Volt regulator have a tolerance of 10% the LM323K regulators that were used on the power driver boards have an output tolerance of +/- 2% and is the only component in the circuit who's tolerance matters as the allowable input to the regulator has a very wide tolerance (7.5-15V). There is a 1% replacement available for this regulator for <$9 but I have no experience with it other than knowing it's available and others have used it.
http://www.ezsbc.com/index.php/featured-products-list-home-page/psu5.html#.WM1enm_yu9J
ebay.com link: LM323K Replacement Switching Regulator 5V 3A

In regards to the watchdog itself, the 10% tolerance you are implying is complete fabrication as the input to the supervisory circuit is tied directly to the 5V bus on the MPU and the supervisory circuit itself has a rated tolerance of 2% or < 0.1V.

Under the most extreme ends of the rated tolerance spectrum the LM339 will supply 4.90V and the supervisory circuit will trigger a reset at 4.7V. In a worst case scenario, having a voltage .2 volts above the reset point cannot be considered to be "at or slightly above the watchdog trigger voltage"

Ref the data sheet for the 323K - http://www.ti.com/lit/ds/symlink/lm323-n.pdf , Here's where the output voltage is spec'ed to range from 4.75 to 5.25 volts... a 0.5 volt range on a 5 volt target output is a 10% range (+ or - 5%):

Wide range of the lm323 listed in TI data sheet

Ref the data sheet for the 34064 under voltage detector - http://www.mouser.com/ds/2/308/MC34064-D-96827.pdf , here's the comparator range of 4.5 volt to 4.7 volt:

wide, and too high range of the 34064

If you've got a bottom-of-range 323 and a top-of-range 34064, your separation between supply drive and reset trigger is only 0.05 volt. That lack of operating space in the design is part of the design flaw. A few more windings on on the transformer would have helped headroom part of the design flaw upstream of the regulator.

Net-net, no deception and really quite accurate... And again, when thinking broader than just about the 5 volt line the daughterboard does great things for the system!

BTW, I've been working with ezSBC for over three years now and every daughterboard I've ever made has one of their regulators integrated (well, except the one hand-etched prototype in my basement that has a muRata regulator). Their PSU5 exists because I asked them to create it. I'm pretty confident I was the first one to point it out to Pinside (ref https://pinside.com/pinball/forum/topic/announcing-new-product-to-help-resolve-resets-in-wpc-era-games/page/4#post-1667232 in my thread announcing my daughterboard). Prior to that product announcement I had pre-release versions that I installed in my Party Zone - I burned chips on my PDB testing them b/c the adjuster could let the voltage climb too high. Even though the adjustment range is tighter now I still hate that microscopic adjuster ... but otherwise it's a great product and I bring them to pinball shows for customers to purchase. I still offer the option to buy the daughterboard with one of their 7812 replacement regulators because the improved regulation efficiency counterbalances the load MPU load introduced to the 12 volt line. All for the good of pinball.

Too bad Pinwiki hasn't kept pace with these contemporary best practices - it still lists the cut-a-trace and install a resistor method to up your 323 voltage... I don't know many people who would prefer to find that over a daughterboard and a PSU5 behind the backglass.

-Rob
-visit http://www.kahr.us to get my daughterboard that helps fix WPC pinball resets or for my Williams system 3-7 sound board potentiometer solutions

I didn't read any further in your post than what I quoted; but you are quoting the spec of the 323N and calling it a 323K, you can also just as easily quote the spec for the 323A portion from the same spec sheet that lists the tolerance as 1%. All specs in that datasheet are irrelevant though as neither one was used in the power driver board. I do see where the confusion is though as Mouser sends you to this spec sheet for the LM323K.

I did notice that the regulator you use in your board was made by the same company that makes the 323k replacement part also makes the regulator that you are using. I did not know about your involvement in thier creation. Like I mentioned earlier, this is a viable option for a large number of folks to get thier machines running. Thanks for your engineering support to the community.

This is very well said. I was going to reply with something similar, but you've stated it perfectly (except for the typo calling the LM323K an LM339 ).

--
Rob Anthony
Pinball Classics
http://LockWhenLit.com
Quality Board Work - In Home Service
borygard at gmail dot com

I think the point that Rkahr makes is still valid. The standard tolerance used on LM323K's is 5%, few people use the LM323AK with 1% tolerance (and is now obsolete). The standard part can range from 4.75V to 5.25V and still be within tolerance.
Williams used the MC34064/MC33064 voltage supervisor. These would trigger at worst case of 4.7V. A simple ground bounce occurance on a nearby IC can easily trigger the reset generator.

I have been an amusement route tech for a very long time. I just want to add a last resort(or first resort if you don't have the time to screw around with this problem) fix for this reset problem. Yes its a hack but its cheap, quick and won't damage anything.
Take out the LM323 you don't need it. Get a old school video game switching power supply. We always used to have one in the service truck. Peter chow anyone lol? Solider a wire from the power supply +5 to the hole where the LM323 regulated +5 output was. Solider a wire from the power supply ground to the LM323 ground. All you have to do now is run AC to the switching power supply and adjust the voltage. You can plug it right into the service outlet if you want to keep it super simple, the CPU would still run if you switch the pinball off though. I would consider this a quick temporary fix but you can run it this way indefinitely. Takes about 30min.

Rickd Great first post! Now help me with my reset issues on my WhiteStar board!

https://pinside.com/pinball/forum/topic/100-stern-playboy-fix-it-challenge-experienced-tech-help-needed