My ECM problem is identified...

J

Jeff

'87 GN (for sale)
Joined
Mar 27, 2002
This week I noticed a problem with my ECM when the car got warmed up. It would intermittently go to Open Loop mode and all the codes got completely screwed up, the tool said the tranny was in D/R when it was really in P/N, the RPMs were WAY off, speed was way off, the car was running like crap, etc.
Today I found the problem...it turned out that this was happening when the coolant fan kicked on. I pulled the switch and relays and they are toast. All kinds of sticky black stuff around the terminals. So I started the car with them removed and it hasn't done the weird thing yet. Ordered new relays and a switch, they'll be here Monday.

A question...should I suspect my ECM after submitting it to that kind of abuse for who knows how long? My O2s are WAY off, but I suspect old injectors for that one.
Thoughts?
thanks,
Jeff
 
I hate to tell you this, but those new parts aren't gonna help. What you have described is a common problem with the relay back emf when it turns off. I looked at this problem a few weeks ago and here is what I found. I use diacom, and when the low speed fan turns on, the diacom readings go wacko. I got my scope on the ecm fan output and you can see that the fan is being turned off every 30ms or so for a few microseconds. This produces a positive voltage spike of about 40 volts on this line. Now, it's not the rising edge of this pules, nor is it the 40 volt level that causes the problem, but it's the negative edge from 40 volts to 0 volts that is causing a negative spike to show up on the ALDL serial stream. Someone said that the fan is PWM'd. This is not correct. Although by looking at the fan signal it appears to be a PWM, it doesn't fit the definitin of a PWM signal. So the fan is not being PWM'd, it's just being pulsed on and off. This is a bug in the ECM program. I fixed mine by adding a 1uf 50v cap in the ECM on the fan output line. You can also mod the relay, but for me, the ECM was easier to fix.

Dave
 
whoa, time out...

I'm a mechanical engineer, not electrical ;)

First question...does this actually cause a problem for the ECM or just the scan tool? The car runs differently when this happens, but it may just be because I'm tapping into the ALDL.

Second question...can you describe in a little more detail the fix you implemented? If this is a problem for the ECM I'd like to fix it. Does everyone have this problem?

Thanks

Jeff
 
This only affects the scan tool. Not all scan tools will get messed up by this ALDL spike. I've also got the free winaldl program and it is not affected.


My fix was adding the cap to the fan control line inside the ECM. the relay itself can also be modded by adding a diode across the relay coil. Although, I didn't actually test the diode fix on the relay, so I don't recommand that one. The cap fix has been tested and there are no down sides. And yes, all ECM's will do this, but you will only notice this with certain scan tools.

A little more detail on the fix....if you open up the ecm, locate CR10. This is a diode located about in the center of the board down by the big ecm connector. Connect a 1uf 50v cap across this diode. The cap I used is a surface mount device and is actually soldered on the non component side of the board.

Dave
 
Geez...dumb me! Where are my manners??? If you feel you or someone you know can solder this cap in place, send me an email with your name and address and I'll send you a couple of caps.

Dave
gnjones231@mchsi.com
 
How come every one does not get this back emf?

Very good diagnosis gnjones231 !!! Put that engineering degree to use.
 
Everyone does get this back emf to some degree. In my particular case my diacom readings would screw up only when the low speed fan was on and the brake applied! It's the low speed fan relay that is the root cause. And don't forget, some scan tools are not affected by this, and who knows what other kind of conditions may aggrevate this problem.

And to go just a bit deeper as to why this happens. In the ecm is a quad driver chip. Inside this chip, there are two pairs of diodes. The cathodes of these diode pairs are tied together and tied to an external zener diode (40v). It's the capacitance of these diodes that is letting that negative edge of the back emf pulse to interfere with the ALDL data. The low speed fan relay and the ALDL data are on one pair of these diodes. Clear as mud, right?


Dave
 
Originally posted by gnjones231
Everyone does get this back emf to some degree. In my particular case my diacom readings would screw up only when the low speed fan was on and the brake applied! It's the low speed fan relay that is the root cause. And don't forget, some scan tools are not affected by this, and who knows what other kind of conditions may aggrevate this problem.

And to go just a bit deeper as to why this happens. In the ecm is a quad driver chip. Inside this chip, there are two pairs of diodes. The cathodes of these diode pairs are tied together and tied to an external zener diode (40v). It's the capacitance of these diodes that is letting that negative edge of the back emf pulse to interfere with the ALDL data. The low speed fan relay and the ALDL data are on one pair of these diodes. Clear as mud, right?


Dave
Click to expand...

Took da words right outta my mouf...
Jeff
 
Originally posted by G-Force
Could your ECM been the reason why you failed your emissions?
Click to expand...

I think it could have been...but hard to know unless I replace it and see...
jeff
 
Just wanna let everyone know, I had this EXACT problem (the Original problem Jeff was having) and GNJONES sent me the Capacitor and several EXCELLENT quality pictures of excatly how to solder it to the board, well.... DAMMED if it didn't FIX THE PROBLEM!!! GNJONES - You are the MAN!!! -- Thank You!:)
 
Great stuff!

If you have pictures and such, I'd be happy to get that out on the GN/TTYpe web site for all to use as a resource.
 
Goober is no dummy. Has and had some neat ideas in his lifetime. He should post more often.

ECM fix would be good to publish with permission of course.
 
gnjones231, that's neat info. would you by chance have a (good quality) JPG file of the ecm schematic? When I get a moment I'll scope that line too. Again, very interesting. thanks
 
Interesting, about 6 months ago I fought the same problem. I use
autoxray to rough in the cars tune then finnish up with direct scan.
The autoxray is an aldl scan tool and on one of my buicks, I have 3,
it would go crazy when the low speed fan would come on. If you
would also step on the brake it would get worse. So I scoped the
fan and fan wiring to see what was going on, I had a theroy about
ground noise getting into the aldl signals. Putting the scope fan
supply and ground showed a lot of noise about 2 volts of it riding
on the + and ground. It looked like a sloppy square wave with a
nasty .6 v spike on the rising edge. So I decided to try beefing up
the fan wiring starting with the ground to decrease this noise.
That's all it took was adding a additional ground from the fan over
to the gnd bolt near the charchol canaster. The a quick re-check
with the scope and the noise now is less than .5V. I haven't had
a scantool droupout since. Now I will look at the cpu signal since
I've read this very interesting thread as I started to fix the problem
at the fan end and look backward, mostly from all the ground
problem talk with these cars.
 
Gotcha Steve. :)

Email him directly with your request ya never know when he's gonna be around. He'll answer you directly I'm sure.
 
Might be having this problem, so I'm bringing this up again.
Does anybody have a pic of this mod? It's an old post and a lot of the posters haven't been on-line for years.

Paul
 
You must log in or register to reply here.

Similar threads

D
runs with stock ecm, not with xfi 2.0
2
Replies
20
Views
845
DGC
D
H
GREMLIN CONTINUES TO BE PRESENT MAKING DRIVING UNSAFE
2
Replies
29
Views
1K
Ttype6
T
J
spal fans and shroud
Replies
4
Views
355
jlat
J
S
Need big help no injector pulse!!
Replies
9
Views
473
salvageV6
salvageV6
1986grandNash
1986 grand national
Replies
8
Views
464
Chuck Leeper
Chuck Leeper
Top