New Life Games Tech Forums

from the old NLG site...

Ok here's an interesting a problem i just had on a *pair* of
model 40x dotmation WMS slots.

Got a batch of dotmations, and i could tell they were the
"bottom of the barrel".  that is, others had tried to fix them
and failed. parts were missing, and every-little-thing was
wrong with them.

Of course the 12 pin power supply "cans" were dead, and
missing their switching power supply board from inside
the "can" (common on these games). So i did the cable
modification (see prior posts), and added a single computer
power supply for both the lower power module and the upper
dotmation power.

but the two games would not boot consistently. you could
just tell something was wrong, if you could get them to boot
at all (most times the Fail LED would slow flash on and off repeatedly.)
If they did boot, the lamp matrix lights (payline lamps, button lamps,
etc.) would flicker badly. (This was the key to diagnosing the problem.)

The problem turned out to be the lower power module.
Even though i replaced most of the voltages with a
computer power supply (+5, +12, -12), there are still
two voltages created by the power module "can".

already removed the switcher part of the original power supply, which
provides +5, +12 and -12 volts DC. But the lower power module (LPB)
still supplies 18 volts DC (unregulated) for the lamp matrix, and
+5 volts DC aka "+5vi" (the center LED on the I/O driver board). 
And in addition there's a HPCL-3700 chip (8 legs) on the
power supply that gets 16 volts AC. what's that all about?
And why a separate +5 volts supply, different than the main
+5 volt supply?

Looking at the schematic, you can see 16 vac coming from the
transformer through the 12 pin J10 plug. This 16 vac then goes
past a 47v MOV and to a 8 pin HCPL-3700 chip. The MOV prevents
over voltage (the device shorts if more than 47 volts passes by,
blowing the fuse).

The HCPL (mis-labeled on the schematics as HPCL)
http://www.avagotech.com/products/product-detail.jsp?navId=C1,C7189,C4982,P88908
(Mouser part number 630-HCPL-3700-000E)
is a voltage/current threshold detection optocoupler. What this
does is sense voltage/current levels. It allows a micro-processor to interface
to this chip, so the processor can easily tell if a certain voltage is
present. So why is this needed on the power supply?
The answer to that in a second...

Schematic:
http://www.pinrepair.com/slots/wms/wms_40x_ps_sch.gif

The next mystery is the extra 5 volt (+5vi) supply. Again this
comes off the 16 volts AC from the transformer and thru
fuse F1 (10amp slow-blow), goes thru a small
transformer that knocks that voltage down to 8 vac, and thru
four 1N4004 diodes (a mini-bridge rectifier), and then
to a LM7805 (1 amp) voltage regulator and some 470 mfd
filter caps. So why this separate 1 amp 5 volt power supply?

Lastly, the 16 volts AC also goes to a 200 volt 35 amp bridge
rectifier and FIVE 15,000 mfd caps (wired in parallel). That's
like a 75,000 mfd cap on the 18 volt DC line. Not sure why
you need so much filtering, but that's what they do. This voltage
is used for the lamp matrix. (no mystery there.)  Originally i
thought the lamp matrix flickering was caused by an open
diode in this bridge rectifier. But that was not the case.

So back to the question, why the HCPL chip and why the LM7805
powering a different 5 volt supply? The answer is that the microprocessor
must monitor the power supply. That's the HCPL chip's job. And the
7805 provides a "quick" 5 volt supply, much quicker than what
the switching power supply gives. But why is that needed?

Say someone spins the reels, and then pulls the power cord,
turning off the game in the middle of a spin. What does the
game do? How does it know its current state for the next power-on?
Well the battery RAM stores the state, but how does the game
know it's getting shut off?

This is what the +5vi does. Switching power supplies are notoriously
slow for power up, and slow for power down. But a simple 7805 linear
voltage regulator is REALLY fast in comparison. So fast, that the
microprocessor can "see" the +5vi die thru the 7805 on a power down
by reading the HCPL chip, store the game's state to RAM, and do this
before the switching power supply "dies" and kills the game.

So why all the concern about the slot saving it's current
condition at a sudden power lose? because if WMS didn't do
this, no doubt someone would find a way to power off the
slot at just the right time, potentially tricking the machine
into free wins. This was WMS' fear, and why they used this extra
+5vi voltage and the HCPL chip to monitor it. According to
Duncan (who worked there), they had a guy that
tried to do just that nearly full time (trick the slots into
undeserved payouts).

So back to our power supply issue. The LM7805 is really robust.
As long as its heat sink is screwed down tight, the 7805 circuit
rarely has a problem. I seriously doubt you'll ever have to deal
with fixing that circuit. It's a good idea to check the four 1n4004 diodes
which rectify the voltage going to the 7805. Use a DMM (set to diode test),
as these 1n4004 diodes are easy to test and cheap to replace.
And while you have the DMM meter out, also test the 35amp 200v
bridge rectifier used for the lamp matrix too.

On the other hand, the HCPL chip is not nearly as robust or
easy to test. And this was the chip that was giving me fits.
Once I replaced it and C5, the games worked fine, and now it could
read voltages and the microprocessor was happy. Luckily the HCPL
chip is available from mouser for about $3.
http://www.mouser.com/Search/ProductDetail.aspx?qs=s%252bgfxR24RQ4WrPoKtCCdwA%3d%3d

More stuff to consider:
While you're at it, also replace capacitor C5 (47 mfd 25 volts),
as it's a cheap part and used in the HCPL circuit. Not a filter cap
either, probably has to do with timing, so a good idea to replace it.
In reality, that may have been the only issue (the HCPL chip may
have been fine, and it was the C5 cap causing all the trouble.)

And check the out-bound 2n3904 transistor at Q1 with a multimeter
(set to diode function). Again used in the HCPL circuit, and easy
to test (cheap to replace too.)

Interesting stuff...
cfh@provide.net

(http://www.pinrepair.com/slots/wms/ps9.jpg)

Thank you for the direction on this problem. I've had it happen intermittently for many years. Anyway, I had a backup lower power supply to work with. The problem was the original was a 1994 version and the newer was from 1996. There is a molex style connector near the rear 110v outlets. On the 96 version this was a 16 pin and on the 94 it was 10 pin. I found this out after swapping out the switching power supply and installing it in the cabinet! Anyway, I decided to do a part by part replacement. As it turns out, after first replacing the C5 Capacitor the problem was fixed. So I didn't need to replace the IC HCPL-3700, maybe the C5 is the first to go and the common cause of this problem? It's a cheap and quick fix with a soldering iron.

Just thought I'd share in the case that your Williams WMS 401 has flashing or flickering lights with the CPU bong stuck in a loop.