Asteroids proper grounding (followup to 11/08 postings)

- Finalized wired grounding scheme as follows:
  * EHT grounds remain as referenced above.
  * video + PCB grounds back to stock.

Back in 11/08 I posted the below about grounds in an Asteroids using the
G05-801 monitor. After more tinkering, I changed my mind on a few
things and just would like to post a for-the-record update,

My final configuration has:

  - the power brick tied to the monitor per a dedicated, heavy guage
wire (disconnecting the original, weak ground in the harness at the
monitor end of connector).

   - the CPU board-to-monitor video board grounds (via video cable) back
to stock configuration. I reverted to this because I ended up agreeing
with the original design that it provided the best video reference
ground, and because a physical ground loop does not necessarily mean
there's a problematic ground loop. Sufficient resistances must exist in
the loops to create a problem.

   - a reworked grounding of the EHT unit for improvement by separating
the grounds on one side of the flyback from the grounds on the other
side. The grounds on the primary side are grounded as normal per the
power harness wire. The grounds on the secondary side are now grounded
via dedicated wire going back to the main chassis ground point.

Previous post::
    I'd been reading John and William's discussions re grounds, and
decided to investigate an Asteroids I'm restoring since I was right at
the point of redoing the harness anyway. I found some interesting
things along the way that lend support for what John says about grounds.

    My Asteroids uses the G05-801 monitor. While examing the grounding
scheme, I found an error in both the schematics and factor monitor
wiring harness which I'll discuss below.

    In this machine, you have a power brick where the AC comes in ties
earth ground to the chassis. The brick puts out an unreg 10VDC and
72VCT, both with a common ground (but floating -- not earth grounded
here), as well as a 65VCT and 6.3VAC for the monitor.

    The 10VDC goes to the Atari regulator board, gets regulated down to
5VDC for the PCB. The 10VDC grounds, now named the +5 Return lines,
pass right through to the PCB. This 10vdc/5VDC power is still floating
(not connected to EARTH) at this point, and if you don't have the
monitor video cable plugged in, you can measure from 5VDC (or GND) to
power brick chassis and get both an offset VAC of around 9VAC and a much
smaller VDC offset.

     The two wires of the 72VCT (hi and lo) pass right to the PCB, and
since that supply is common GND with the 10VDC, the center tap wire is
not needed.

      The monitor is a different story. All 3 wires of the 65VCT supply
go to it, as well as an EARTH wire, in the P100 connector on the DC
regulator board. CT and EARTH come in at P100, and are immediately
tied together there. This ground then passes the whole length of the
regulator board, goes out at P101, and gets tied to the monitor's metal
frame.

      While tracing out P101, I found an error in the factory manual.
The schematics throughout the G05 service manual consistently show that
the L100 inductor filtered power comes out of pin 1 of P101, and that
unfiltered power is at pin 2. Yet, if you turn the DC regulator board
over, you'll see that the exact opposite is true -- the filtered power
is on pin 2. Further, if you check the P101 wiring harness, you'll see
the fat +25VDC wire going to the deflection board is wired correctly
according schematic, thus wrong because the board is mixed up. The thin
+25vdc wire going to the EHT unit is receiving the unfiltered power; the
fat wire is getting the filtered. I have two of these G05-801
monitors here, and both have the same flaw. Easy to fix by reversing
the two wires in the P101 harness. Now, back to grounds.

       The metal frame grounding point has a few other things tied to it
-- the aquadag strap, and the main ground for the deflection board.
Tracing the def board ground, I saw that it enters the board, heads
south, and provides the ground for the video in connector. From the
PCB, each of the 3 video signals (X-Y-Z) has its own ground wire, and
all 3 are tied to essentially earth/chassis ground at P703. 3 nice
huge ground loops.

     Now, if you've ever done audio work, you'll look at this grounding
scheme as really a mess... although functional. More grounds are not
always better -- it's how you do them that matters. And what matters
most is proper grounding to begin with.

    It's about impossible to go through the entire asteroids and convert
it to a properly done star grounding arrangement, but I did make a few
changes that support John's comments.

    First, I lifted the flimsy earth ground wire from the monitor's P100
connector, and ran a fat wire directly from the power brick chassis to
the monitor's frame ground point. The advantage here is that chassis
ground doesn't have to occur through the entire DC regulator board
(which is very fragile and has thin copper traces). This reduces the
chassis-2-chassis resistance by a few tenths of an ohm.

     Second, I lifted all 3 video grounds from the PCB end... leaving
them in place and still connected to the def board P703 (grounded at one
end and twisted, they'll still provide shielding). This lifting
removes the EARTH grounding of the 10/5/72 volt power supplies from
occuring through the video lines.

    Third, I provided an EARTH ground directly to the PCB (front, lower
corner has a nice hole for it near the X-Y video outs) using a fat wire
from the brick chassis to the PCB.

     The end result here is that all major ground loops have been
removed. Note that I did not run another ground to the Atari
regulator. Why? Because this would have re-established a major ground
loop. The regulator has no need of a chassis ground because it has
no separate chassis or other 2nd ground plane. The stock 2 ground
wires from it to the PCB should be sufficient, but if you wanted to beef
it up, you'd simply replace those 2 as well as the power lines with
fatter copper.

     Prior to these mods (but still with clean edge contacts and edge
connector), if I measured VDC and VAC between the various grounds with
all cabling in place, eg. PCB ground and the brick chassis or the
monitor frame, I'd get readings in the tenths of a volt... 0.30 VAC for
instance. After these mods, all grounds are solidly even.

     With far less stray AC and DC currents flowing around where they
aren't needed or wanted, this has got to improve reliability of any game
+ monitor. But I caution, it's not simply a matter of stringing
grounds around. You've got to examine the schematics and actual wiring,
and see what makes the most sense to do if anything.

  JS

---------------------------------------------------------------------------
** Unsubscribe, subscribe, or view the archives at http://www.vectorlist.org
** Please direct other questions, comments, or problems to chris@westnet.com
Received on Tue Jan 13 21:40:13 2009