MS1v3 - JS4-JS5 signal leak. How did I screw things up?
 

Background: I've got an MS1v3 set up to run parallel on my '97 Miata.

I've modified it in the following ways:
- Joe Perez's "better/inverted spark circuit"
- IAC2A connected to JS11 (with an inline diode) for launch control
- Extra MAP sensor installed and signal to JS5 for constant baro
- SPR4 connected to JS4 for EGT logging (signal comes in to MegaSquirt from an external circuit board made by SymTech Labs)

The unit is also modified in that it's got a USB-serial connector installed (thanks to the previous owner: arkmage), but I don't think that's likely to affect what I'm trying to deal with at the moment.

When I first hooked up the thermocouple, I noticed that EGT readings on MS1 seemed unaffected by actual EGT, so after doing some testing I've discovered that somehow I'm getting signal bleed-over between JS4 and JS5. With the MS hooked up to a stim and the EGT signal disconnected, I can blow into or create vacuum on the baro sensor and watch EGT readings go up and down.

Any ideas how I've screwed things up so much? I've included links to photos of my board.

http://www.somethingsquare.com/temp/ms1/IMG_6992x.jpg
In this first image, I had removed the jumper from the proto area (just used to bring the jumper to SPR4 through to the other side of the board) to JS4. But normally it would be connected.

http://www.somethingsquare.com/temp/ms1/IMG_6997x.jpg

Any suggestions/troubleshooting tips are appreciated. Any namecalling/insults are not appreciated but will be tolerated, especially if you can help my fix this.

EDIT: Oh yeah, and when I try to measure resistance between JS4 and JS5 with my digital multimeter--with the MS1 unit powered down--I get the following results:

When measuring between JS4 and JS5 with COM lead on JS5:
- 1.06M diminishing to 1.02M ohms when testing in the 2Mohm range
- 175K diminishing to 170K when testing in the 200Kohm range

When measuring between JS4 and JS5 with COM lead on JS4:
- ~0.73M (possibly diminishing, but slowly) when testing in the 2Mohm range
- 127K diminishing to 126.3K when testing in the 200K range

Check for solder flux around the socket.

Matt, thanks for the reply. I gently scraped and brushed the area around the sockets to ensure that there was no bridge between them (as far as I can tell), but the result still seems to be the same. I can't really see what's going on beneath the MS1 chip, but it doesn't appear that the is any slop in there to cause a bridge either. I gently scraped that area as well, but again no apparent change in behavior.

When connected to the stim and powered on, I get ~1.75 volts between JS5 and the ground at the proto area. I get ~1.25 volts between JS4 and the ground at the proto area.

Here is a close-up of the area around JS4 and JS5: http://www.somethingsquare.com/temp/ms1/IMG_7118x.jpg I could not take adequate photos of the area under the MS1 chip itself though.

Is there anything else you can think of that would cause this sort of behavior? And for that matter, is there any solution aside from desoldering the chip itself to get a better look/access that you think might work?

Remove the CPU and check the resistance between the two pins. What do you get?

EDIT: OK, when the CPU is removed, there is infinite resistance between the two pins. Similarly, there is infinite resistance between the two sockets on the board. However, once the CPU is reinserted, the same resistance as measured before occurs again.

Anyone have any idea why the chip would allow voltage to pass through it between the two points when it's plugged into the board, but when it's unplugged there is no continuity between the both the points on either the board or the chip?

Don't know anything about this chip, but behavior like this is pretty normal. It could be - for example - that one pin is connected to ground and the other to +5V and when you're plugging it in, the current goes out the ground pin, through obscure and perfectly OK paths on your board and back in through +5V.
Or any of a million other scenarios. The only way to be certain would be to compare to an absolutely identical and known good MS.
If your MS works otherwise, those two pins are likely OK too. Problems like you see are 90% harness or wiring in my experience.

Yeah, I'm thinking that it's almost got to somehow be related to the wiring of the board itself and not the chip, but I'm not sure exactly where I've gone wrong. Since there's no continuity when they are separated, I suspect that somehow one of my built circuits is either wrong or when combined with another circuit creates this odd, undocumented side-effect.

OK, so my brain finally engaged for a brief moment and I thought to disconnect the baro signal from JS5. With the chip plugged in and baro removed, there is NO continuity between JS4 and JS5. The continuity is passing through the baro sensor.

So either:
The baro sensor wiring is screwed up (though it doesn't seem to be, as baro correction functions as would be expected)?
or
MS engineers never anticipated the particular combination of circuits that I have built on my board?
or
Some other circuit is defectively or incorrectly wired on my board?
or
Some fourth possibility that I haven't been able to fathom yet?

I think that I need guidance in determining how/what to test.

I removed the baro signal from JS5, hooked MS1 up to stim, and connected to TunerStudio. All other signals/gauges are steady, but EGT is still actively jumping/seeking around 300F-400F. When tested with my digital multimeter, JS4 to ground on proto sees a jumpy 1.3-1.8 volts, which sort of corresponds to the gauge behavior in TunerStudio.

Since nothing is hooked up to JS4, shouldn't it be seeing 0 volts and the gauge in TunerStudio be steady? Is this indicative of a problem with my MS1 chip?