Stumbling when rad fan comes on

[rg2200]

I have a MS-II with 2.89 code on a 5.0L Mustang V8 in a Mazda B2200 pickup. I hadn't driven it for any amount of distance (more than 30 minutes or so) until recently. The trip consisted of highway driving, and then a bunch of stop and go city traffic. Once the truck was fully warmed up, I'd be sitting at a light, and sometimes the engine would just die/stall, sometimes it was fine. When I'd restart it, I'd have to feather the throttle for about 10 seconds to keep it running, then it would be fine. After watching more closely, I found that it seemed to do this about every minute or so. If I was driving/cruising and disengaged the clutch, often the engine rpm would just drop (as expected), but then keeping dropping below idle speed, and would just die - I'd re-engage the clutch to start the engine (while coasting), but it would surge for a few seconds, then stabilize. Then, it would run normally again. I got back on the highway on the way home and there was an accident on the highway, so back to stop and go traffic on the highway. The truck started to run pretty lean, then started running very lean (18-21 range),so I pulled over and turned off the engine right away.

I let it sit about 20 minutes (until the traffic backup somewhat cleared up), and started again - it was still running a little lean (16-17 range), but not dangerously so. The MS-II is in the glove box, and I opened it and the MS case was quite warm (to the point that I probably couldn't hold my hand on it for 20 seconds). Once I got back to highway speeds, then the problem seemed to go away. As I got back near home and back into some more stop-go traffic, it started to mildly act up again.

Unfortunately, I didn't have my laptop with me, so I didn't get a chance to datalog it.

Today, I wanted to verify that the rad fan was coming on, so I checked and it was. To be clear/sure when I'm driving that it's turned on, I wired a small light on the dash to the s1 (along with the wire going to the relay for the fan), so I could see visually when the fan kicked on. In testing the light, I used the spare port settings screen to adjust the variable to force the fan on (so I could verify that the dash light was working etc.), and once it did the engine started acting very much like it did before. This is when the engine is cold. So, I'm deducing that the problem wasn't necessarily because the engine was warm, but rather because it was triggering the fan to turn on. I was starting to think heat soak problems etc., but it doesn't seem that way.

So, I datalogged it this time. I've attached the datalog and msq. The datalog is very short - just starting the engine, turning on the fan (where I have to give it a little throttle just to keep running), then it settles down a little, then I turn the fan off, then stop the logging. It's very obvious when the fan is turned on - the graphs spike all over the place, for example, the coolant temp reading drops dramatically, which seems to cause it to feed more fuel, then it's like a roller coaster with smaller and smaller lumps on the graph. I have a volt meter in the truck, and I didn't see any sudden wavering of the voltage when the fan kicked on.

I have S1 going to a 30 AMP relay, the load of the relay goes right from the battery (via a fuse) to the relay and from the relay to the fan.

Any ideas, or things that I should check, or anything obvious?

Thanks in advance.

[rg2200]

I've done some additional testing and datalogging. Matt Cramer had replied via email - Matt, I've taken some timing out of the low RPM's - attached is my updated MSQ as well.

I've attached 2 datalogs:

1. In the 'first datalog', I have the engine running and then using the spare ports screen, I manually enable the spare port (about 928sec) that goes to the relay for the fan. You can see the rpm drop etc. I then unplugged the power source to the relay (so the spare port would still trigger the relay, but no power going through the load part of the relay to the fan.) This did not seem to have the same effect. You'll see the TP spike - the throttle was never physically depressed like this, although I did give it a little throttle to keep it from stalling.
2. At some point, when I had the power to the fan relay disconnected, I noticed the light on the dash that I wired to the spare port (s1 on the relay board) was flashing on and off - on for <1 sec, off for about 1 sec, then on again, but not in a totally consistent pattern. So, I put power back to the fan, started the truck and then datalogged the result. This is a lot more like what I had experienced on the longer trip before. At about 1348sec I disabled the spare port, and it settles a little. I have the hysteresis set to 5 - it almost seems like it's set to .5 - the way it cycles on and off like it does. You can see the TP wavering all over the place - the throttle pedal was physically in the same spot until about the 1348sec position.

Thanks.

[rg2200]

Thinking that maybe since I'd messed so much with my tune over time that it might be just generally out of whack, I started a fresh tune, using the MSII_base_code as a starting point. I very quickly got the basics set and got it running. I was able to manually turn the fan on (by changing the spare port trigger from >195 to <195) and there is a brief drop in rpm (which is somewhat expected as a sudden load from the fan occurs), but it did not seem to go into the 'roller' coaster in the log viewer. Once I let the engine idle until it started to hit 195 (with the spare port setting back to >195 as the trigger), it started this stumbling/revving 'roller coaster' ride. Attached is the msq and datalog (after the CLT hit the 195 range). You can see the CLT spiking from the low 180's to 195ish, which would obviously trigger the hysteresis to cycle the fan on and off. Until it hit 195, and the fan triggered on, the CLT temp was steady (this isn't on the datalog, since I started the logging after this problem started.) It seems that when the fan kicks on, it's causing the readings to be off, which causes the fan to kick off, once it kicks off, the readings are ok, and it triggers the fan back on. Note that while this is happening, the dash light I installed (hooked to S1, so I can visually see when the fan kicks on), is turning on (very briefly, more like flashing), then off and cycling to match the cycling in the datalog.

[Matt Cramer]

There's some pretty bad electrical noise, not just on CLT but on most of your other analog inputs. I'd recommend taking a look at this guide to noise troubleshooting:

http://www.bgsoflex.com/megameet2008/me ... gnding.pdf

[rg2200]

Matt,

I re-traced all my grounds and found that I had the fan ground on the same lug as all the sensor grounds. So, I moved the fan ground over to the other side of the engine bay, and so far the problem seems to have been resolved. It isn't warm enough out tonight to get it very hot driving around, but I was able to let it idle in the garage long enough for the fan to kick in, and no drop in rpm, no fluctuating etc. So, I think it's resolved.

Do you sell, or know where I can get, a good terminal block/ground block? Right now, I have a heavy gauge wire going from a lug on the sheet metal (a very heavy gauge wire direct from the - on the battery shares this lug, so it should be a good straight path to ground), and this wire is about 8" long and terminates with a ring connector on a bolt with all the sensor grounds and the MS-II ground also all with ring connectors on this same bolt. All the terminal blocks I've found aren't a common connection, they're more like the one in this link: http://order.waytekwire.com/productdeta ... L%20BLOCK/ . I figured that a good grounding/terminal block would be a better (and neater/tidier looking) place for all the grounds than the bolt that I currently have. Your thoughts?

[trakkies]

It's better to run all the sensor grounds back to the DB37 connector. And run a couple of heavy ground wires from the DB37 to the engine block. Running all the grounds to some remote point means the high currents flowing in the MS ground will also be flowing in the sensor ones.