Re: Friday Vector Trivia #001

From: Kevin Moore <talon.k_at_gmail.com>
Date: Fri Aug 27 2010 - 21:26:27 EDT

Typically I've only seen this problem "Chainlink" because of a DAC. Although
it's normally much more pronounced than what you have. So my money is on the
AD561 for X. D11 or 12 can't recall off hand.

I think I have a picture of the defect in a worse case scenario. I'll see if
I can find it.

Kevin

On Fri, Aug 27, 2010 at 7:07 PM, Jess Askey <jess@askey.org> wrote:

> Hi All...
>
> I have about 10 Asteroids PCB's that are dead. I will post a new one every
> friday with a synopsis of the issue and an attempt at a rationale of the
> issue up front. Im sure some of them might be complex enough that we might
> need to do a few rounds but this will be something fun for the archives
> either way. In 'CarTalk' fashion, I will draw a name from all that made
> correct guesses and send them some free beer schwag from New Belgium Brewing
> here in Fort Collins, CO (my humble employer). If Im correct on my guess at
> the solution, then my name goes in the hat too (the house has to have an
> angle doesn't it?)
>
> 'Ditto' guesses don't count, but you can come up with a different rationale
> that names the same defective part as someone else and I will count it.
>
> Monitor in my cabinet is known good (except for blooming until I get my HV
> diodes from Mark C. :-)
>
> Asteroids PCB #1 (Rev C)
>
> I haven't pulled any chips on this guy yet, but the cool thing about the
> vector engines is that you can tell a lot by just surmising info from the
> test display...
>
> http://www.ipsnd.net/images/asteroids_badimage_001.jpg
>
> As you can see, there is a step in the diagonal, the horizontal line at top
> shows some retrace from defect and if you look ever so closely on the
> vertical lines, there is a slight skip to the left.
>
> The issue is not...
>
> 1. The monitor
> 2. The circuitry after the first buffer op-amp (eliminates the 4016's and
> the final Op-Amp on the PCB)
>
> I don't think it relates to 'Big Blue' as I figured the defect would move
> if it were an analog noise issue (however, this *is* an assumption).
>
> Here is my rationale... but it might have a flaw in it...
>
> 1. the defect seems to show mostly in the X axis and not the Y (Im going to
> dismiss the slight deviation on the vertical line as I just can't seem to
> explain it). If you look close at the image, when moving diagonally, the X
> vector counter seems to count backwards one(maybe two) step(s) every certain
> count.
> 2. There are 15 defects moving along the X count axis. Since the X axis
> counter is 1024 steps, that would mean that it is happening every 1024/16
> (16 spaces resulting from the 15 defects)... or 64 counts, that puts us on
> the defect occuring when bit 7 (one based) is going from low to high.
> 3. Looking at the X counters (IC position on Asteroids PCB's depend on the
> revision and there are G(7)? revisions) this falls into the two lower end
> counters LS191's. There isn't anything complicated here, but it seems that
> there isn't any way that the UNMDACX5-UNMDACX8 counter could cause the
> UNMDACX1-UNMDACX4 lines to change since the lower counter feeds the upper
> counter via the RIP (ripple) output. So, I didn't think it would be a
> counter issue.
> 4. the next step in the circuit is the counter limit multiplexers (LS157s)
> which limit the count range to 0-1023 to prevent wraparound (overcounting).
> Again in this circuit, I don't see a way that bit 7 could cause a change in
> bits 0 or 1 since they are in physically different IC's.
> 5. The next stop is the DAC, since there is lots going on in the DAC there
> certainly could be a small effect when bit 7 flips potentially. Both the X
> and Y DAC use the +5VDAC supply, there is a chance I suppose that if the
> XDAC is messed up, that it might be putting noise onto the VDAC line causing
> that little glitch in the X axis on vertical lines?
>
> So, my guess is the X DAC.
>
> Thoughts? Guesses? Please give a nice rationale for those newbies out
> there!!! Only original rationales get counted, first come first serve based
> on vectorlist timestamp.
>
> have a great weekend everyone!!!
>
> jess
>
>
>
> ---------------------------------------------------------------------------
> ** Unsubscribe, subscribe, or view the archives at
> http://www.vectorlist.org
> ** Please direct other questions, comments, or problems to
> chris@westnet.com
>

---------------------------------------------------------------------------
** Unsubscribe, subscribe, or view the archives at http://www.vectorlist.org
** Please direct other questions, comments, or problems to chris@westnet.com
Received on Fri Aug 27 21:26:30 2010

This archive was generated by hypermail 2.1.8 : Sun Aug 29 2010 - 00:50:01 EDT