MEP-804A won't start

[maddios]

Hi Guys, I'm new here, helping a friend of mine with his MEP-804A generator.

The machine ran great and started really quick then suddenly one day it stopped starting.

What's happening is when you turn the knob to START it'll bump the starter for a moment (maybe 1/2 second or less) and then cut off. Very rarely the machine will crank a bit longer before cutting off the start and it'll sometimes almost start but sputter out. But that's so rare that I can't reproduce it right now.

Here's what I've tried so far:
I've checked the MPU, it's putting out 2.5VAC while holding the dead crank switch to crank.
Checked the MPU ohms, they read 1050 or so. Seems a bit high (spec is 800-900) but I'm guessing it's close enough.
The starter itself spins the motor very fast so that's not the issue.
Measured the same 2.5VAC at the MPU terminals of the woodward governor controller.
Tried swapping the the K15 and K16 (the 2 start and start cutoff relays) with other relays, no change.

A random thing I tried which yielded a "result" was to back the MPU out until it read something like .9VAC (at full crank speed) and the machine would then crank and almost start a couple times but not really.

Does this mean the woodward control panel is dead or should I just try a replacement MPU just in case even though mine diagnoses just fine.

I've taken this video trying to show the issue

I appreciate any and all input to try to fix this machine for my friend.

Thanks!

 

[Guyfang]

Thats not a MPU failure. Something is telling the set to stop. Pull out the plug on the fault indicator and see if it turns over. What and why is there all kinds of "extra" wires, or non standard wire in the control cube? I do not see the Quad MWO installed on the set. Might want to do that.

 

[maddios]

Yean, at some point in the past the machine had some overspeed relay installed on top of the existing woodward controller, I tried disconnecting all its wires and it made no difference.

I'm not sure what a Quad MWO is, can you be a bit more specific?

As for the fault indicator plug. Do you mean remove on one of the connections on the woodward controller? Which wire is the fault connection?

Thanks!

 

[maddios]

Oh, I just found what you mean by Quad MWO, I think this machine actually had this fault. The whole generator head blew up a few months back and was repalced by a mechanically matching one, but it's no longer 3 phase, so none of those wires/gauges do anything anymore, The control panel just operates the motor itself and the frequency knob lets you adjust engine rpm. But the voltage controller is no longer used.

After this gen-head swap the machine worked great for many many hours, I think my friend ran his whole house off it for about 15 hours straight during a recent power outage. But then when he went to start it a week or so later the machine refused to fire just like it's doing now.

 

[Guyfang]

The fault indicator is the box where the red under voltage light is on.

 

[maddios]

Does the fault indicator feed back into the governor controller? Because the controller itself triggers the starter cutoff. If so maybe I could just disconnect/bypass the fault indicator thing to at least get the engine running and make sure the governor itself is working fine.

If so, do you know which terminal would be the fault connection?

 

[Guyfang]

The A2, (Fault Indicator) shuts the power off to the Speed controler if it "sees" a fault. It should turn on a light, but the A2 is known to fail. So try it.

What wire? You need to look at the schematic.

 

[maddios]

So It's "fixed" now, but I'm not sure why or how it worked before.

I noticed that the frequency selector switch (50/60Hz) was set to 50, i flipped it over to 60 and it fired up O_O but was reading low frequency, so i just dialed up the FREQ dial on the governor up to 60 and now it's running nice.

But I'm really confused why this worked on the 50Hz setting at all since the gen-head is single frequency and can only output 60Hz at 1800rpm.

Guess I can just chalk this up to gremlins.

BTW: I tried unplugging the 18 pin connector on the back of the fault indicator panel and it made no difference, the machine still just did that bump crank and nothing more.

Thanks for focusing my attention in the right area, MPU was distracting me before.

 

[maddios]

My friend swears he never even opened the control panel since I dropped off the machine at his house after swapping out the gen-head. lol, who knows.

 

[maddios]

So seems like the "no-start" issue didn't really go away, it's just so intermittent that it "fixed" itself for a while.

Since april my friend used the generator almost daily until last week when it stopped starting again, just click and nothing.

At this point I went over there and ran some more diagnostics, I removed the addon overspeed relay, swapped the start/overrun relays around with other relays. No change, it will sometimes do that momentary crank and not start and other times will start but then shut down within a few minutes.

When it does the momentary crank the max MPU voltage goes up to maybe .5-1.5v (it literally "cranks" for like 1/2 second).

When it actually starts and runs for a while, I tried to measure input voltages to the woodward and the input power stays at a steady approx 26v the whole time the machine is running, MPU input reads around 9VAC (i think, but it was very steady as it was running). and output to the fuel regulator was around 12.5vdc. As far as I could tell when machine shut down the MPU voltage was still reading the same and dropped off with engine rpm drop, but the fuel regulator voltage suddenly just went to 0 as the engine was killed.

By the way, through all this testing I've left the fault indicator panel's 18 pin connector unplugged.

At this point I'm leaning toward a faulty ECU, maybe some cold solder joint or something like that. Does that make sense or am I missing something stupid obvious?

Thanks

 

[maddios]

Just did some more careful reading of the schematics and noticed that the LMU is basically directly connected to the governor at pins 11 & 12 and 13 & 14 go off to the paralleling connection.

Could the LMU being fully disconnected (it's in there and powered but not getting any input from the gen-head) be sending whacky inputs to the woodward causing this condition? Basically what are pins 11-14 actually responsible for? I can't find a woodward manual. Does the governor even look at the lmu input during start?

 

[Guyfang]

Just did some more careful reading of the schematics and noticed that the LMU is basically directly connected to the governor at pins 11 & 12 and 13 & 14 go off to the paralleling connection.

Could the LMU being fully disconnected (it's in there and powered but not getting any input from the gen-head) be sending whacky inputs to the woodward causing this condition? Basically what are pins 11-14 actually responsible for? I can't find a woodward manual. Does the governor even look at the lmu input during start?

No problem ever fixes itself.
Lets start anew. You have never said that you have checked out the S1. Have you done that?

 

[maddios]

That's a good question, what's the best way to test that thing?

I guess we can just put a jumper to pin 3 from 2,4,6,8 to simulate "RUN" and then a jumper to pin 7 for start signal. I'll see if my friend can try that and get back with the result.

 

[nextalcupfan]

That's a good question, what's the best way to test that thing?

I guess we can just put a jumper to pin 3 from 2,4,6,8 to simulate "RUN" and then a jumper to pin 7 for start signal. I'll see if my friend can try that and get back with the result.

I'd shut off the dead crank switch and test for continuity across the terminals with a multimeter

The blue plastic switch's are pretty famous for failing in weird ways.

 

[maddios]

I tested the switch like that and it checks out, the failure is very sporadic. The machine sometimes just clunks and no crank, other times it'll crank for a second and not start, other times start and run for like 10 minutes then die on its own.

I'm guessing the switch is possibly iffy and goes in/out. I'll see if bypassing it fixes anything.

Oh, I bypassed the dead crank switch already (well, tried to and put it back when it made no difference). I don't think that's the problem.

 

[Guyfang]

The S10 is about as nuke proof as any switch I have ever seen. Mis wired yes, bad, Never. Try this. proper way to test the S1.

 

[maddios]

These tests would only apply if the starter clunks and immediately disengages right? I'll try to go over these anyhow just in case.

This doesn't address the shutdown after a few minutes problem does it? I'm really confused what correlation a short crank and shutdown after a few minutes of running have in common, other than maybe a faulty run position of the S1.

My friend at this point has disconnected the 4 LMU feedback pins 10-14 and no difference.

 

[Guyfang]

Have you checked the canon plugs on the control cube? For tightness. For damage to the pins.

 

[maddios]

you mean the 2 coming up from the load measurement area? I've taken them apart, sprayed some electrical parts cleaner and put some dielectric in them. Didn't make any difference.

My friend bypassed S1 today, made no difference, machine started well, and ran for about 4-5 minutes then shut down on its own, 2nd test, same thing but it shut down after about a minute or two.

I'm thinking next step is to just run 2 wires directly from the battery to the governor controller, to bypass all of the wiring/relays. If that fixes it then I'll at least narrow down the fault to power distribution. Otherwise I'll know it's signals/controller, maybe the tach signal is flaking out sometimes. Then I guess lastly it'll be the governor controller itself, but hopefully not since it's like $1600

 

[Guyfang]

Here is another trick. It feels to me, that you are losing a voltage someplace, and it could be in the control cube. Could be a loose wire, a wire that is rubbed through. I had several time a problem just like this. different time Intervalls before it shut down. So I waited until dark. Then opened the control panel, and turned off all lights. Then started the sets up. Then just stared into the control cube. When the set shut down, I saw a spark. Turned out every time to be a wire rubbed through. Vibrations caused it to arc out, and sometimes the wire stayed touching the wall, sometimes not. Cant hurt. Cost nothing. Might work out for you.