Re: Stumped by my WG 6101

At 09:32 PM 6/26/97 -0600, you wrote:
>Hi All,
> I have a WG 6101 (Tempest) monitor in an old Omega Race cabinet that I
>use for my Atari multi-vector setup. Basically I have completely
>re-wired it with all Atari stuff at this point.
> Anyway, the problem that I have had for about 6 months now is that I
>have "noise" in my vectors. It is in both the X and Y axis and seems to
>be high frequency since I can see the noise repeat in a nice long
>horizontal line about every inch or two.

<snip>

I've seen this noise on many WG color vectors so I spent a weekend at Bill's
and together we finally tracked down where it was coming from. If you look
a the regulator built into the HV on the WG monitors you'll see that it's a
very high gain error amp. This regulator will turn on when it's output
falls below it's threshold, and when it turns on, it can turn on pretty
hard. It can place a pretty good spike on the power supply.

Atari's attempt at limiting this spike are the 2.2ohm (3.9ohm on later
models) resistors on the HV regulator. If your's are 2.2ohm you might try
increasing them to the 3.9ohm, you might have to increase the wattage of the
resistors since 3.9 ohm resistor are dropping more voltage they're going to
get hotter.

The biggest reason the spikes are there is because the low voltage regulator
used by Atari is the biggest piece of crap ever designed by a game company.

It's a joke, destined to fail. Every revision of the schematic had
something patched in here & there to try and keep it from failing. Things
like Zener diodes use to protect transistor from over voltage during On/Off
transition -- except they forget to put current limiting resistor on the
zener to protect them. Well forgot? They actually had them there in one
revision, but if you place the resistor in series with the zeners, the
zeners can no longer shunt the over voltage surges, etc.

The design is not really a regulator, but a capacitance multiplier.
Together the circuit looks to the input of the deflection circuits as a
filtered supply with a much higher capacitance than the 4700uf supplied.
When it works.

I've done two things with the WG's I own, not simple I'm afraid.

Fix #1:
The easiest, and the one I personally prefer, is to remove all low voltage
circuitry. The X/Y amps have current feedback built into them and don't
need regulated voltages. This is what they did on the Amplifone monitors,
there is no low voltage regulators on these monitors. This will also speed
up the slew rate of the monitor bringing up to the slightly faster speed of
the Amplifone.

Unfortunately just doing that will burn up your HV since it will now be
supplied with +/- 33 volts instead of the +/- 25 it was design for. The
small transistors used in the HV can't dissipated this extra power, I
watched them melt like an ice-cream cone the first time I tried this.

Also unfortunately you can't simply place a couple of +/- 24 volt regulators
on the input of the HV. This works great for the HV, but the 78xx/79xx
regulators were not design to care about the regulation of there inputs,
just the outputs. These regulators will jerk the input line harder than the
HV alone, and your noise problem will just get worse. BTW: Amplifone solved
this problem by changing the way it regulated the high voltage. The
Amplifone HV regulator consists of a specially designed tuned circuit
(including the HV transformer) and it places a fairly constant load on the
power supply, so there are no "pulses" needing filtering.

I, at this point, found bought some surplus 56vct transformers (28 volts on
each winding) and using a bridge diode, a 7824 and a 7924 (put them in the
now unused sockets where the low voltage regulator caps went), a couple of
filter caps (1000uf) and built a +/- 24 volt supply just for the HV cage.

This gives you a monitor that runs cooler, has a faster slew, *much* more
reliable, nice clean vectors. It also is no longer a drop in replacement
for the original since it now needs an AC line for the HV.

Fix #2:
Beef up the low voltage "capacitance multiplier" to be a real regulator.

By removing the low voltage section of the WG and replacing it with a couple
of regulators you can run the WG as a drop in replacement.

Unfortunately they don't make 5amp three pin voltage regulators (that I can
find), but you can use standard "tab" regulators and use the existing low
voltage pass transistor to beef up the current handling.

By placing the external pass transistor in parallel with the three pin
voltage regulator you can have it take the brunt of the regulation current.
It's a very simple design only needing a resistor, the regulator, and some
bypass capacitors to be added to the existing pass transistors. Ascii art
being what it is I'm not *even* going to try and draw the schematic. If
anyone is interested I'll figure out someway to get in on the net. The
circuit is right out of the Motorola Voltage regulator handbook and/or databook.

Bill P. made up a small PC board and etched it in his garage. It makes the
upgrade fairly painless (much easier than the original ESB/Star Wars
conversion kit -- No Clay this is not a dig! Just a comparison! :^). I'm
not sure whether he wanted to go into mass production of these things, it's
probably not worth the time and effort considering the quantity sold. It
not that hard to wire up on a onesies and twosies scale.

This gives you a *rock solid* display since the regulators are much faster
than the spikes presented by the HV cage. Everything runs cooler (we used
adjustable regulators and regulated at +/- 27 volts, though the monitor
worked perfect at +/- 24 volts using fixed voltage regulators -- though the
slew rate might suffer slightly and might be noticed on something like Major
Havoc, designed for the Amplifone.)

Neither one of these are easy fixes, but it shows you the problems I went
through to get rid of those HV cage noise lines once and for all!

I have seen WG working (properly, without noise) with unmodified Low Voltage
sections. Though most of the time the WG I've seen in operation do not have
their regulation sections working properly. Cracked diodes, bad trasistors,
if the thing fails properly the monitor runs, but the regulator isn't, so
you end up with noise on the Vectors. Yours might be running in this state.
One problem is, I believe, the replacement transistor used to replace the
obsoleted ones used in the regulator section have higher gains than the
original. Since there is no feedback in the regulator circuit, this higher
gain keeps the transistors from running in the linear range the original
designed intended (they most likely saturate). Some re-spec'ing of some
resistor values probably needs to be done.

-Zonn
Received on Fri Jun 27 10:58:38 1997