My Twilight Zone went hard down recently with a DMD display problem.

Blank screen with nothing on the display, although would boot into game mode and would play.  I always heard a surge of power and a loud “thunk” when I first turned it on, but I had assumed that was from the speakers kicking when the sound board was being initialised. I had read that this sound is typical when switching on WPC games of the early 90s. But be careful it isn't masking other problems with the PSU and DMD HV supplies.

At first I considered buying a new DMD glass or an LED replacement. But I’m cheap and so I decided to have a further look first.

I had bought the WPC Theory of Operation and Schematics Manual dated Oct '91 some time ago to learn about Williams WPC games. So I had a look around the DMD Controller board and schematics (especially the HV one which was 4 of 4). Doing some tests I could see there was no +62v available to the DMD display. The +62v section is composed of:

Four Resistors:
R12, R3 both 47K ohm 1/2w resistors. (I could get metal film resistors rated 47K at .6w - available from Maplin, my local electronics store)
R11 a 120ohm 5W cement power resistor (which I didn't have)
R4 a 120ohm 1/2w resistor (I had 120R at .6w again also available from Maplin)

Two Caps:
Capacitor C4 which is 150uF@160v (I didn't really suspect this)
Capacitor C6 which is 0.1uF@500v (I didn't really suspect this)

Plus 3 Transistors:
Q3 an MJE15030 (which I didn't have)
Q2,Q10 both 2n5551 (mpsd52 alternative which are readily available)

Diode D3: a 1n4759a Zener Diode 62 v (which I didn't have)


First thing to do was to read through the Marvin's Repair guides WPC 1990-1999 part 3. Section 3-k let me know I needed to rebuild the HV section and there was a HV kit>/b> available for $5 from Great Plains Electronics but I live in the UK and didn't want to wait for shipping from the USA. I also thought it would be expensive to ship, more than doubling the cost, but I actually spent as much buying the parts locally from Maplin and on eBay.co.uk

The lesson I learned is to be patient, buy the kit from GPE (and pay the International shipping). Go ahead and replace everything in the kit all at the same time! Clay tells you to do this on the Marvin guides, but as I thought I was only missing the +62v I tried to only repair that section of the board. It didn't save time or money to take this approach. Here is a repair log as it may be of use to someone else. I did get it all working in the end, but was a more difficult repair.

Step One. Replaced the following parts (or tested them for correct values):

• 4 x 47K resistors at R12,R13, R3, R6.

Note: You can read R3 and R6 directly in circuit, BUT you will need to remove a leg of R12 or R13 to get a good reading, so I just replaced all 4 of them. R3 and R6 were both within spec. but replacing them meant they all matched. Beauty is in the eye of the beholder.

• R4,R5 were in spec. both close to 120 ohm. Same for the 5W Cement power resistors at R8,R9 and R11. Cost = 0
• Replaced Q2,Q10 with 2n5551 transistors.
• Replaced Q4,Q5 with 2n5401 transistors. These are actually part of the -113, -125v circuit (-103, -115 for a WMS), but are low in cost.
• Tested Q3 and Q7 against each other as they were both MJE15030 (BUV27). This was inconclusive as they both gave similar readings.

Q3 gave an HFE of 27 and Q7 gave an HFE of around 30, so I swapped them over when I put them back. I suspected Q3, but didn't have a replacement. Also note that the HFE readings need to be done out of circuit to be accurate as there are resistor shunts.


Put DMD Controller board back in the game, WITHOUT connecting the DMD Display and took some readings at connector J6:

J6-04-01 Orange -114.5 (-115v ideal)
J6-04-02 Blue -98.6 (-103v ideal) There should be a 12v difference between readings at pin 1 and 2. This is 15.9v
J6-04-03 Key
J6-04-04 Black (Ground reference)
J6-04-05 Black (Ground reference)
J6-04-06 Gray +4.94v (+5v ideal)
J6-04-07 Gray/Yellow +12.12v (+12v ideal)
J6-04-08 Brown +61.7v (+62v ideal)


So this all looks nearly OK, maybe I will need to replace the 12v 1n4742a zener diodes at D6 and D8. But I wouldn't get bent out of shape if the difference pin 1 to pin 2 is larger than 12v without the DMD connected.
I have seen nearly a 20v difference with no load, but it goes down to the 12v difference (under load) when a working DMD is connected.


When I put a known working display (Vishay AGP128 from my AFM) on the Controller and powered up, the display would do a display test.
Well sort of, at least I could see something. The symptom now was this:
It would do the single line test across vertical and down horizontal fine, but any full on dot graphics with the inverted lines would fail.
Same for the all the normal grids and inverse grids. Stopping the display test at various points confirmed the +62v was showing +58v on lines vertical. But dropping to less than 30v on horizontal lines or when doing the full dot graphics.
Running the display in attract mode confirmed this, simple text graphics kind of worked but any animations would fail.


Under load, with a good DMD connected, the voltages looked like this:

J6-04-01 Orange -119.5 (-115v ideal)
J6-04-02 Blue -107.7 (-103v ideal) There should be a 12v difference between readings at pin 1 and 2. This is 11.8v
J6-04-03 Key
J6-04-04 Black (Ground reference)
J6-04-05 Black (Ground reference)
J6-04-06 Gray +4.95v (+5v ideal)
J6-04-07 Gray/Yellow +12.12v (+12v ideal)
J6-04-08 Brown +58v (+62v ideal), but dropping to less than 30v when displaying full dot graphics. Ouch.

The point here is that without the load of the DMD, you can't test the components under any stress, So voltages might still look right.


Step Two. I had a closer look at the DMD Display (Babcock GD-032D128),
and one of the surface mount Bypass caps at C9 looked bad (almost certainly was blown).
BTW- Bypass, or decoupling caps are mounted very close to the IC between ground and and its power rails to filter out TTL switching and other "AC noise".
You see one of these below each of the +5v TTL logic chips on a PCB connected between the power and ground pins, usually small as they only need to be rated at 50v. 100nF (which is 0.1mfd) is a common value. Downloading the WPC-89 schematics to get the Babcock display schematic, I confirmed C9 to be .1uF @ 100v.
100v is needed as the column HV supply is +62v or slightly more.
I had a replacement ceramic cap with the right value and so put that in place to get the display going. Disc caps are useful as bypass caps.
I confirmed that now I got the same results with both DMD displays. At least I had something on the display to work from, and the fault was not the DMD “glass” as I had originally thought.

BTW- The Babcock DMD does need to see the +12v correctly from my experience to work, where the Vishay display doesn't use it.
Sometimes seeing “noise bars” on a Babcock DMD will point to a problem in the +12v section. Check ribbon cables first, then suspect the 12v supply isn't reaching pin J604-07
– use a DMM and test for voltage on the pins. If you see something like +5v there, and not the full +12v you may have found the problem.

Check that 12v is reaching the input to the DMD controller, if not you will need to replace :

D8 (1n4742a 12v Zener), D7 (1n4004 Diode 400v 1A), and possibly R7 which is usually 56 ohms and may be cooked.
This wasn't the problem in my case, so my TZ was going to be harder to fix.


SO AFTER ALL THIS, I STILL HAD TO REBUILD THE HV SECTION

Step Three. Rebuild the rest of the HV section.
I ordered the diodes and power transistors I needed from ebay UK, I also decided to replace C4 and C7 with equivalent parts (220uF@160v) as they were 50p each.


Replaced the following parts:

• C4,C7 Capacitor Axial Elec 220uf 160v 105 degrees C.
• Q3, Q7 MJE15030 Transistor NPN
• Q6 MJE15031 Transistor PNP
• D3 1n4759a Zener Diode 62v
• D4,D5 1n4758a Zener Diode 56v
• D6,D8 1n4742a Zener Diode 12v
• D7 1n4004 Diode 400v 1A Rectifier. I use these on flipper and other coils, so have many on hand.


Step Four. Needed to replace R11 to complete my repair of the HV.


During testing, I found that R11 had gone open at some point after Step One. I couldn't find a 5W wire-wound or ceramic resistor rated at 120 ohm locally.
So I bought two wire-wound cement resistors from Maplin, a 330R @ 3W and a 180R @ 3W.
I piggybacked them in parallel to get 330*180 / (330+180) = 116.5 ohm (which is in spec. given a 5% tolerance)


R8 = 4.77K (4.7K 5W ideal) not changed
R9 = 1.798K (1.8K 5W ideal) not changed
R11 = 116.6 ohm (120 ohm 5W ideal) after replacement it's within 3.4 ohm.


Game is now working and all tests show correct display results and good voltages. I left it up and running in Attract Mode for a few hours to be sure of the repair.

I now plan to order the GPE “WPC-HVP-KIT” from Omaha, Nebraska as a backup and will add a extra 120 ohm 5W resistor or two, so I can finish my repair properly.
I won't bother replacing the SMD 0.1uf cap @100v on my Babcock display right away, as it's working now. I will get the right surface mount part for a few pennies at some point.
Displays aren't supposed to last forever, although the one in my game has been around since Oct ’93 according to the date code on the DMD display board.
I have fixed a line that was out by carefully chipping the glass around a broken pin and getting enough exposed to solder to.


Parts and Costs
Finally here is a list of the parts used and the costs both in GBP (£ Sterling) and USD ($)


Step One:

• 4 x 47K resistors at R12,R13, R3, R6. (Maplin code is M47K at 13p each which is about 23 cents USD at time of writing) Cost = 92 cents
• R4,R5 were in spec. at close to 120 ohm. Same for the 5W Cement power resistors at R8,R9 and R11. Cost = 0
• Replaced Q2,Q10 with 2n5551 (Maplin code is UL36P at 17p each which is about 30 cents USD) Cost = 60 cents
• Replaced Q4,Q5 with 2n5401 (Maplin code is UL37S at 17p each) Cost = 60 cents


Total cost so far: USD $2.12 This was looking cheap if it was fixed at this point. But it wasn't.


Step Two:

C9 on Babcock display replaced with a .1uF 100v ceramic disc cap. On hand, so Cost = 0.


Step Three:

• C4,C7 Capacitor Axial Elec 220uf 160v 105 degrees. 50p each (90 cents each, as we are still keeping score) = $1.80
• Q3, Q7 MJE15030 Transistor NPN 84p each ($1.50 ea) = $3
• Q6 MJE15031 Transistor PNP 84p each ($1.50 ea) = $1.50
• D3 1n4759a Zener Diode 62v 19p each (34 cents ea) = $0.34
• D4,D5 1n4758a Zener Diode 56v 19p each (34 cents ea) = $0.68
• D6,D8 1n4742a Zener Diode 12v 24p each (43 cents ea) = $0.86
• D7 1n4004 Diode 400v 1A Rectifier. I use these on flipper and other coils, so have many on hand.


Total additional cost: $8.18 (OK - but I probably didn't have to replace the 150uF caps at C4 and C7.)

Total cost so far $10.30


Step Four:

• W330R (Wirewound 3W 330 ohm) for 33p
• W180R (Wirewound 3W 180 ohm) for 33p


Total cost was an additional 66p ($1.18)


Total repair cost was $11.48 + royal mail postage costs for the eBay items which were not available at Maplin.
Cost of the GPE “WPC HV kit” would have been around $5 + at least $7 shipping to the UK. Both work.