Can anyone help me get an early Williams Defender working?

powkesmore

New member
Joined
Feb 25, 2012
Messages
19
Reaction score
1
Location
bridgnorth, shropshire, United Kingdom
Can anyone help me get an early Williams Defender working?

Hi, I can't tell you how excited I am to own a genuine Williams Defender. It was always the game I loved to play in the arcade as a youngster (far too many years ago now) and I am pathetically desperate to get the thing working again (my wife currently has that long suffering look she wears when faced by my video game nostalgia).

When I start the machine up there is no video, no startup sound and all four leds stay lit on the ROM board.

I checked the voltages to the boards and various chips and they are all fine, the supply is clean with no voltage 'ripple' showing on the oscilloscope. The board is the old layout. Based on an internet search of repair logs for 4 leds constantly lit, one problem identified was a problem with the initial reset on startup. I checked the reset circuit which seems to be working OK but when I checked the reset pin on the MPU it stayed low even when the reset button was pressed. I followed the track on the circuit board back to a logic gate chip at 5F. The reset goes in but doesn't come out. Looking at the circuit diagrams I deduced that the other two gates on the chip are used by the Video RAM controller which I hoped would account for the lack of anything being displayed on the screen. The chip is a 7411N, I removed it, installed a socket and ordered a new chip, replaced it but still no change (I was gutted). However if I bridge pins 8 and 11 with a resistor (bypassing the gate) I get a start up sound, all of the leds go out and a second later the middle two leds light up. Still nothing on screen. The strange thing is once this is done the reset switch works normally until the machine is turned off then it reverts to 4 leds on constantly with no change when reset is pressed.

I'm afraid it has been many, many years since I've done any electronics but I was a little confused to see that the logic gate at 5F is an AND gate but reading the documentation it says that the inputs to the gate come from the reset circuit and the watchdog circuit but then goes on to say that input on EITHER will trigger a reset but wouldn't that be an OR gate?

I really want to do everything I can to get the original machine working, a friend of mine advised me to pull out the existing electronics and fit a Jamma board but I can't bear to desecrate it that way!

Anyway I would be very grateful indeed for any advice you might have..
 
I would replace that reset switch first thing and you are correct if either input will cause the reset then it is an OR gate. There could be a condition where either will cause this and that to do the reset but I haven't look at the schematics yet.
Something is holding that in reset and that is where you should be looking. You are on the right track and keep on it.

Hi, I can't tell you how excited I am to own a genuine Williams Defender. It was always the game I loved to play in the arcade as a youngster (far too many years ago now) and I am pathetically desperate to get the thing working again (my wife currently has that long suffering look she wears when faced by my video game nostalgia).

When I start the machine up there is no video, no startup sound and all four leds stay lit on the ROM board.

I checked the voltages to the boards and various chips and they are all fine, the supply is clean with no voltage 'ripple' showing on the oscilloscope. The board is the old layout. Based on an internet search of repair logs for 4 leds constantly lit, one problem identified was a problem with the initial reset on startup. I checked the reset circuit which seems to be working OK but when I checked the reset pin on the MPU it stayed low even when the reset button was pressed. I followed the track on the circuit board back to a logic gate chip at 5F. The reset goes in but doesn't come out. Looking at the circuit diagrams I deduced that the other two gates on the chip are used by the Video RAM controller which I hoped would account for the lack of anything being displayed on the screen. The chip is a 7411N, I removed it, installed a socket and ordered a new chip, replaced it but still no change (I was gutted). However if I bridge pins 8 and 11 with a resistor (bypassing the gate) I get a start up sound, all of the leds go out and a second later the middle two leds light up. Still nothing on screen. The strange thing is once this is done the reset switch works normally until the machine is turned off then it reverts to 4 leds on constantly with no change when reset is pressed.

I'm afraid it has been many, many years since I've done any electronics but I was a little confused to see that the logic gate at 5F is an AND gate but reading the documentation it says that the inputs to the gate come from the reset circuit and the watchdog circuit but then goes on to say that input on EITHER will trigger a reset but wouldn't that be an OR gate?

I really want to do everything I can to get the original machine working, a friend of mine advised me to pull out the existing electronics and fit a Jamma board but I can't bear to desecrate it that way!

Anyway I would be very grateful indeed for any advice you might have..
 
I'm afraid it has been many, many years since I've done any electronics but I was a little confused to see that the logic gate at 5F is an AND gate but reading the documentation it says that the inputs to the gate come from the reset circuit and the watchdog circuit but then goes on to say that input on EITHER will trigger a reset but wouldn't that be an OR gate?

I'm no expert on Defender, and I haven't taken a look at the schematics you're referring to, but...

A reset line (like many things in digital electronics), is "active low." This means that it "does its thing" ...resets the CPU in this case... when the line is pulled low (zero, off). This means that it is normally high when it is inactive, not resetting the CPU.

Logic gates, on the other hand, are usually "positive logic". So, an AND gate requires that all of its inputs (3 in the case of a 7411) be high (one, on) in order for its output to be high (one, on). Conversely, if any one of its inputs goes low (zero, off), then its output will be low (zero, off), because not all of them are high (one, on) any longer. So, in effect, in the "negative logic" sort of way that an active-low signal like reset works, an AND gate works kinda like an OR gate.
 
Last edited:
This sounds like the correct operation of the reset circuit. As Darren pointed out the reset is active low, and so should pulse from high to low at the 7411 when the reset button is pushed. If you are seeing this pulse, it probably means the watchdog circuit is stuck low.

Check the 7402 at 3d. If pin 4 is low, the watchdog is stuck Low. Check pins 5 or 6 of the 7402 if they are low, replace this chip, otherwise check Pin 12 of the 74393 at 3c, it should be pulsing quickly. If this isnt happening, the data bus isnt running and you have bigger problems. If it is replace this chip. Try this and report back with your findings.
 
It sounds like you are referring to the AND gate at 6O - correct? If bridging pins 8 and 11 allows the CPU to come out of reset, then the other input (pin 9) must be keeping the reset asserted low. Can you confirm the logic levels of pins 9 and 11 on the gate?

Pin 11 comes from the watchdog circuit, whereas pin 9 comes from the power-up / switch reset circuit. If pin 9 is stuck low, then you can backtrack from there using the schematics, or the theory of operation doc (which explains the expected power-up sequencing):
http://www.cityofberwyn.com/schematics/williams/DEFcpu1.tiff
http://www.robotron-2084.co.uk/manuals/defender/defender_later_pcb_theory.pdf

My guess is you have a failed component in the power-up reset circuit, and manually pulling the reset high is working around this (and allowing the switch to function).

LeChuck
 
6O is a positive 3 input AND gate, meaning it is normally High. The reset line pulls input 9 Low which forces the output on pin 8 low forcing the CPU to reset. Pins 10 and 11 are tied to the output of the watchdog circuit coming off 5M pin 13.

Pin 13 is forced low by the counter at 5O. If the counter is reset by the watchdog reset circuit Pin 2 the counter is reset and continues to count again.

So when the CPU is not running, it can't reset the watchdog, so you should see pin 11 of 6O pulsing forcing Pin 8 low triggering the CPU reset.

Pin 9 should be pulled low for 1 second when the reset button is pressed and then go high again. If it does not, then the reset circuit needs to be debugged. The first thing is to look at all the transistors directly below the reset button. They are taller than they should be and frequently will be bent causing shorts between the legs. if there are any of the transistor legs that look close together, carefully pry them apart with a small flat bladed screwdriver. Also look for any broken legs on the transistors.

Once you have the reset circuit operating correctly (pin 9 toggling for 1 second after the reset button is pressed) then we can look at the watchdog circuit.

ken
 
Hi guys, remember he has an early revision board, you are quoting chip positions from a later board.

From his description, i.e. The input to the tri input AND at 5f from the reset is toggling, but the output isnt, I'm lead to beleive that there must be a stuck input or output in the watchdog keeping the reset line low.

Just checked the schematics and 5f pin 11 comes from the reset circuit, 9 and 10 are from the watchdog. Problem has to be in the watchdog circuit.
 
Last edited:
Thanks..

Thanks guys for the prompt reply, looking at the schematics the board seems to be the old board layout so I've been using the 'defender_early_pcb_theory.pdf' circuit diagrams. The gate is labelled 5F in this document and 6O on the 'defender_later_pcb_theory.pdf'. Am I right in using the early_pcb document?

I'm away for a few days with work so won't have a chance to investigate further until I return, when I get back I'll take a look and get back to you; thanks again for your help in this.
 
Hi guys, remember he has an early revision board, you are quoting chip positions from a later board.
Good catch, I missed it being an early board.

Position 5F is the same as 6O in the later boards, although the pin purposes are swapped (pin 11 is the power-up / switch reset circuit, pin 9 is the watchdog circuit). So if bridging pins 8 and 11 allow reset to deassert, then it sounds like the watchdog input on pin 9 is stuck low.

The early theory of operations doc has all the info necessary to backtrack and figure out where it's failing (mainly checking the functionality of counter 3C and the NOR gates on 3D:
http://www.robotron-2084.co.uk/manuals/defender/defender_early_pcb_theory.pdf

LeChuck
 
OK, looks like the reset circuit is OK, it's the watchdog circuit that seems to be low on startup. If I 'force' the reset by bridging the pins the watchdog pin and the reset pin are high and then the reset works as expected (both inputs are high until the reset switch is pressed then the reset pin drops low and the output drops low triggering the reset). After I bridge the reset the first time the four leds that stay on are replaced by the middle two being illuminated. Looks like the watchdog circuit needs further investigation. I read somewhere there was a way of disengaging the watchdog circuit for testing purposes?
 
Last edited:
I don't think such a thing exists unfortunately. The chip numbers on the schematics do refer to physical chip locations though. I.e chip 3C is the third row, third chip along. Does that help?
 
OK pin 4 at 3D is low, pins 5 & 6 at 3D are high. I can't check if pin 12 at 3C is pulsing as I borrowed an oscilloscope last time but pin 12 is high. I'll borrow a scope again and let you know if it's pulsing. Thanks again for your help.

Andy
 
Correct, if the reset on pin 12 is high, the output on pin 8 should be low. Pin 8 connects to 3d pins 5 and 6, which are high, so there is either a stuck output on 3c, or something is holding that line high, such as a short to the 5v line.

Check the pins on 3d and 3c for any accidental shorts (stray pieces of solder etc) and if there is nothing there, change the chip at 3c.
 
It's been a while

Hi Guys,

It's been a while but I'm almost there with this damned willo the whisp defender machine! I had to send the boards away for repair in the end (every time I fixed a problem it revealed another). Got the boards back connected them up and found out that the monitor had a fault. It was suffering from a vertical frame collapse, I tried all the obvious stuff (transistors on the heatsink, IC replacement, cap kit etc.). In the end I had a retired TV engineer take a look for me and he eventually tracked the problem down to a single low power resistor on the chassis board that had failed. I got the monitor back this evening, fitted it and it works fine. However I get a message on startup that says "Initial tests indicate a ROM failure" but after a few seconds the game starts normally. However when I begin a game most of the controls don't work. The ship immediately goes to the top of the screen and doesn't respond to the joystick. The hyperspace and smart bombs work but thrust, player 1 and fire don't work. I tried disconnecting the control board and starting the game by connecting the player 2 pin to ground, but the ship still goes to the top of the screen.

Any ideas? This damned machine is driving me nuts!
 
I had the same problem your having now after I had my boards repaired. I had a problem with the wiring connection at the boards. Mainly at the CPU. Game would boot up and have a RAM or Rom error and then would start game normally. Turn on and ship would go straight up. I checked for cold solder joints at the header pins on the CPU board. Finally replaced connectors going to the boards and that took care of the problem. I also checked to make sure all voltages were good coming into the boards. Hope that helps.
 
Last edited:
Back
Top Bottom