Glow plug ECU fault diagnosis and repair

Locked
User avatar
FalcoColumbarius
Site Admin
Posts: 5983
Joined: Tue Jan 22, 2008 1:55 pm
Member's Photo Album: http://www.delica.ca/Photos/index.php?cat=11103
Vehicle: Delica; Chamonix GLX ('92 P25W)
Location: North Van, BC, eh?

Glow plug ECU fault diagnosis and repair

Post by FalcoColumbarius »

Regarding the ECU.

Original thread: http://www.delica.ca/forum/rebuilding-t ... -4139.html


Morganizer
Image
Joined: 20 Feb 2011, 20:15
Posts: 92 Member's Photo Album: http://www.delica.ca/Photos/
Vehicle: L300 Turbo Diesel Low Roof
Location: Vancouver BC

In the morning in Port Moody, Delica wouldn't start. It cranked over fine, but the engine just wouldn't go. I noticed that the wire to the glow plug bus bar was really loose: I could swivel it with my fingers. Tightened it but no change.

I looked on the forum, and it said the ECU has likely failed, and to keep cranking it, to build up heat in the cylinders to start it without glow plugs. Bad advice. Sure, it may build up heat in the cylinders, but also builds up heat in the Glow Plug Relay and the ECU. Quickly reached my troubleshooting limits with minimal tools, called CAA and got it towed home to Kitsilano. $112 to renew membership (including $25 fee for doing it just when I need service) and $50 for over 5km. Let the games begin. I'm trained as an Electronics Technologist, so if it's the ECU, it's a home game.

Image

Image

I took out the ECU. It's located in the RHD driver's side B-pillar, behind the seat belt mechanism. Not hard to get out, once you know where it is. Opened it on the bench and a trace to the relay is burned.

Image

Image

Image

Popped the cap off the relay, and it's fine in there. Cleaned the contacts for good measure. Tinned the segment of the trace that had been heated enough to lose its coating, and soldered a 22 gauge wire to bridge the trace that was burned.

Image

Image

I replaced the largest capacitor, because even though it measured fine, I didn't like that it wasn't glued down. Large capacitors tend to crack their solder joins when they bounce around, and I had a hunch it might do that in a Delica. Put it all back together, reinstalled it, then thought, "there must be something else…"

Image

Image

Took out the Glow Plug Relay; it's supposed to measure 20 ohms across the coil, and I measured a short. It's bolted together, so I opened it up and sure enough the coil is fried (and looking most unhappy).

Image

Called CVI and they said they could sell me a rebuilt for $112 (that number again) but generally they put in a manual switch instead, because the ECU is next. But I'd already rebuilt the ECU; if I could fix the Glow Plug Relay I'd have the automated system up and running again.

I looked around for a comparable generic or American relay, went to Lordco and compared it to a Ford Diesel relay. Bigger lugs, so I'd have to change the terminals on the wires to match. They said go to Armature Electric on 5th, which I couldn't find, but on the way noticed a place called Tokyo Auto Service. A smart Mechanic named Vishal called Rocky Mountain Imports on Vancouver Island, and they can bring in new ones from Japan in a week. Probably expensive. Vishal had a generic 12v relay dusty on the shelf, so I bought that for $10 as a backup plan. Again, larger terminal posts.

I still wanted to try rebuilding the solenoid, so Vishal suggested I go to Ashnat Starters and Alternators on Broadway. The guy there wasn't much help, but sent me to Drake's Universal on Francis Ave. They guy there was no help at all, kept saying, "It's not worth it." He had the wire, but it was "too much trouble" to sell it to me. They sell it by the pound, and I need less than that. He suggested Main Electronics, and I gave them a call and they have 1/4 pound spools for $10.

So I rode my bike to Main Electronics just before closing and got a 1/4 pound spool of 24 gauge magnet wire. It is slightly larger diameter (0.057) than the original (0.051) but for some reason I figured bigger was better than smaller. Not sure on that one.

Image

I pulled off the old wire, and sanded the spool smooth. I improvised a winding jig, and wound about 350 turns on it. I counted about 40 per layer, and 8 layers, but really I just filled it up. It's got to be fewer turns than original, because it's thicker wire. I wrapped the coil in electrician's tape.

Image

Image

I cleaned the relay contacts with steel wool, rotated them 180 degrees to provide fresh contact, and swept out the copper powder that had accumulated. I soldered the coil ends into the Glow Plug Relay housing, and reassembled it. I tested it with my Variable Voltage DC Power Supply, and it clicked very weakly, not enough to close the contacts. Discouraged. I took the Chinese relay out of the package and tried the test with that one, and it clicked confidently. I figured I was going to have to go with that, then realized I had tested the Delica relay upside down, with the relay fighting gravity to close. So I turned it over, and it worked! I checked the coil with the ohm meter, and it read about 3 ohms. The Chinese relay measures 20 ohms, like the Delica manual says the original should.

Image

So I decided to install the rewound Glow Plug Relay and give it a go. I did that and the van started! First try. I measured voltage at the glow plugs and got nothing. Which was disturbing. It was getting dark so I slept on it, worrying that the van started without the glow plugs. It was late afternoon, and had been dry for a few days, so not the same conditions as the original no-start.

In the morning I tried the van again, and it started first try. I could hear the first click, but not the second everybody's talking about on the forum. So I measured voltage at the coil, and got about 5v.
Oddly, I measured nothing on the other side. One would think, with the key on and the coil inactive, a voltage would appear on one side. But perhaps the relay is energized as soon as the key is turned to "ON", so conditions to measure the voltage on an open relay exist only after the ECU has timed out.

So I hooked the meter up to the glow plug bus bar again, this time choosing a different reference point on the chassis. I measured about 10v, which dropped to about 2v after maybe 7 seconds. I'm satisfied the system is working properly.

My theory is this: vibration caused the wire to the glow plugs to loosen. The ECU is programmed to increase the voltage until the correct current flows. The poor connection at the glow plugs would offer a high resistance at first, but when the voltage increased, the poor connection would "break down" and a large current would flow through the normal resistance of the glow plugs, heating up the coil in the Glow Plug Relay and the ECU, before the ECU could react and lower the voltage.

If I had gotten on it and replaced the head cover gasket and half moon seal, I would have taken apart the glow plug bus bar, and tightened it when I put it back together, and avoided all this. But I wouldn't have learned how to troubleshoot a no-start in a diesel, so it's got that going for it.

A few days later, drove out to Port Moody again (this time armed with a multi-meter, spanners, and a length of sturdy wire to jump the glow plugs if I need to) and it started no problem. I noticed the second click sometimes occurs up to 3 minutes after the first, which is within specs, according to the timing chart in the manual.

So, I've apparently repaired the problem for $20...half of which was spent on a generic relay I ended up not using. Took a couple of days tinkering, but right now I've got more time than money. Fortunately this happened when I didn't have any big trips planned.

Somewhere on this site (I saw it once but can't seem to find it) there's a post about the T-shaped connector on the temperature sensor, and what resistances you should measure on each. I use my diode checker on my DMM for a lot of power-off troubleshooting, and was surprised that the temperature sensor checked out like a diode, not like a resistor. That is, it seems to not be a resistive element, as one might expect, but is a pair of diodes. This actually makes sense from an electronics perspective, because diodes respond to temperature changes in a linear fashion, and this property means you can design circuits to interpret the changes to gather temperature info.

Anyway, I think the "lost post" I'm referring to provided resistance measurements, which I'm not sure are relevant. In your PM you said, "the voltage of the coolant temperature sensor". In the manual, it describes testing the voltage applied to the sensor, but says it must be measured at the ECU. I'm not sure if this is relevant to the integrity of the temperature sensor either, since the ECU is what's generating those voltages (which is why the manual says to test at the ECU -- because this is a test of the ECU, not the sensor.

Image

I'm going out to the truck to see if I can duplicate the power-off measurements that convinced me my temperature sensor was ok. At the time, I didn't record it because it pointed to my problem elsewhere.

Also worth noting here, the connector for my temp sensor had some kind of goo all over it. Perhaps a previous owner sprayed it with WD-40 and it dried out. I cleaned this with contact cleaner and a pipe cleaner, and I would advise any Delica owner that's under the bonnet checking the oil or whatever to tidy up the temp sensor contact, now that we've seen what havoc any disruption of the ECU system can wreak.

Okay, the temperature sensor seems to be diodes, which is why there's probably some confusion about readings. However, since not everybody has a DMM with a diode checker, I took a resistance measurement across the terminals, with a known good sensor, installed in a cold engine, wire unplugged, and the ohm-meter read 5.83k-ohms, either way (red/black). I also measured from each lead to the sensor body, and got nothing (infinite ohms, or open circuit.) This is a quick-and-dirty check, and if you're reading either shorted or open, the sensor's probably pooched.

Image

Now, if you have a DMM with a diode checker setting, the measurements are perhaps more relevant. Measured from each contact to the sensor body, with the black lead on the sensor body. The contacts are arranged in a "T" so from the cross-bar of the T the reading is 1.8, and from the vertical of the T the reading is 0.9. If you reverse the leads, you get infinite ohms, or no continuity, which is what you'd expect from a properly functioning diode.

I would say, if you put one lead on the body and the other on a terminal and measure a short or an open in either direction, the sensor is suspect.


I found the post describing the connections to the Engine Temperature Sensor, so I'm adding the link to this discussion:

http://www.delica.ca/forum/viewtopic.php?f=6&t=4328
Sent from my smart pad, using a pen.

Seek Beauty... Image Good Ship Miss Lil' Bitchi

...... Vision without action is a daydream. Action without vision is a nightmare. ~ Japanese Proverb
Locked

Return to “The Engine & Engine Electrical”