I think I finally got this problem solved today. Earlier in the week I had found out about the start retard setup in the ems-pro; there is a crank angle entry for starting in one of the software menus, and that actually sets the timing that the ems-pro gives the engine when the RPM is below the cranking threshold (in my case 300 RPM or less). This was set at 10 degrees, which is a 10 degree retard from my normal idle timing, so the timing was retarded somewhat already. When I was running the MSD I used the built in MSD 25 degree timing retard, so I thought it would be worthwhile to try to retard this timing a little further. I tried it at 0 degrees and -5 degrees, which were 20 and 25 degree retards respectively, but didn't get any improvements. I was hoping that with the further retarded timing the engine would crank a little smoother, and this would allow the ems-pro to sync up to the signal from the crank sensor more easily. No joy on that one.
Scott and I have been discussing this offline, and the only other two options seemed to be going to the VR sensor to see if that helped the problem, or possibly going to a 60-2 wheel to provide more resolution during cranking. I was actually sick for a couple of days this week and didn't spend any time in the shop, but Thursday night I felt well enough to go out and look at putting the VR sensor I had tested on the dyno in place on the engine. What I found out was the bracket that I had used for the VR sensor on the dyno wouldn't allow the sensor to fit in the car, because the shape of the lower radiator hose was different in the car and got in the way of the sensor. So, I was going to have to make a new bracket.
I made the bracket on Friday night. It turned out to be a fairly involved process, where I started with a 1/2" piece of aluminum plate and had to cut it to shape on the bandsaw, then after some test fitting machined it on the Bridgeport for the final fit. I screwed around with that thing for about 3 hours before I had it right, but finally at the end of the night I had the Ford VR sensor mounted at the correct angle with a .035" airgap to the trigger wheel.
This morning I came out and changed out the ems-pro to the one I had originally purchased. This one has been modified with some smaller capacitors for use with the VR sensor circuit, and all the new ems-pros are getting this modification because the VR sensors seem to work better with it done. The two spares I have don't have the modification made though, and I wanted to be sure that this didn't pollute the results with a potential circuit issue, so I made the change. I also changed the DIP switch settings on the ems-pro to work with the VR sensor, and then I installed it in the car.
After getting my laptop in the car and hooking up the oscilloscope to the sensor so I could watch the signal along with the ems-pro readouts, I cranked the engine to watch the sensor signal. The VR sensor is a purely AC device; it consists of a coil of wire wrapped around a magnet. As the target spins by, the field from the magnet is attracted to the teeth when they are present, and it not attracted when a gap between the teeth is present. This means that the magnetic flux lines through the coil of wire changes as the teeth and gaps between them spin by, and a change of magnetic flux through a coil of wire will induce a current in that wire (see Maxwell's equations for more information on this
). The faster the target spins, the bigger the current; this is a time related phenomenon, and the faster the flux changes through the coil, the more current is produced. Since the coil of wire has an electrical resistance, the induced current will also induce a voltage on the ends of the coil, and this voltage appears like a sinusoidal waveform on the scope. When I first looked at this waveform as I cranked the engine, I wasn't real encouraged. It was fairly low in amplitude, around a volt or so, and the period of the sine wave varied widely as the engine cranked, pretty much like the square wave of the Hamlin sensor had varied.
One of the issues with using the VR sensor with the ems-pro is that you have to adjust the triggering voltage of the VR circuit in the ems-pro to get the correct signal. There is a 15 turn clutched pot, R28, on the ems-pro board that must be adjusted to set this triggering voltage. You start with the pot at zero turns, and usually it will be in the range of a few turns before the signal is correct. You know the signal is correct when watching the RPM gauge on the ems-pro software; you will get a nice steady RPM reading during cranking when R28 is adjusted to work with the VR sensor correctly. I adjusted the clutched pot to the zero turns position by turning it counterclockwise at least 15 turns, and then flipped on the ignition.
The lights came on and the fuel pump went through its normal key-on cycle. But after about 5 seconds, it did the same thing with no further input. Then it did it again! I looked at the RPM gauge on the ems-pro software and it was blipping up to 200 RPM periodically when the engine was just sitting there. Every time it blipped up the fuel pump turned on, because it thought the engine was running. I turned the key off and thought about this for a couple of minutes; I hadn't seen this behavior on the dyno. Then I remembered that on the dyno I had controlled the fuel pump with a switch, so if this was happening I wouldn't have known about it. I figured that the triggering voltage on the VR circuit was so low that any system noise was triggering the circuit. So I flipped the key back on; the behavior was still there. I started turning the R28 pot slowly clockwise, and sure enough after about a half turn the behavior stopped. Now when I flipped on the key the fuel pump went through its normal cycle, and then was quiet. The RPM gauge stayed at zero.
Next I cranked the engine while watching the RPM gauge on the Tach LED on the ems-pro board. The Tach LED seemed to be flashing intermittently, and the RPM gauge was all over the place, up as high as 1200 and down as low as 0. I went to 3/4 turn clockwise on R28 and repeated the crank test; no change. I continued on in quarter turn increments, and gradually the Tach LED seemed to smooth out, but the RPM reading on the ems-pro software never really settled down. However, when I got to two full turns I suddenly noticed a different behavior in the LEDs on the ems-pro board. There are like 7 or 8 of these things, including one labeled INJ-1 and another labeled INJ-2. I assumed that these were for the fuel injector signals. What I noticed at 2 turns on R28 was that suddenly these two LEDs started flashing alternately; INJ-1 would flash, and then INJ-2 would flash a second later. This suddenly made all kinds of sense to me. I had set the EFI system up for bank fire mode, which means that one bank of injectors fires and then the other bank fires. During all of my other cranking tests with the Hamlin sensor, I had never seen these two LEDs behave like that; they had always seemed to flash in the same pattern as the Tach LED.
Then I looked at the RPM gauge on the computer while the engine was still cranking and sure enough, it was nice and steady at 104 RPM! It sure hadn't started out that way when I turned the key. I turned the key off and started again. Sure enough, when the engine started cranking the RPM gauge was all over the place, and the Tach, INJ-1 and INJ-2 LEDs were flashing in unison, but after about 4-5 seconds the RPM gauge fell into a steady 104 RPM signal, and the INJ-1 and INJ-2 LEDs started flashing alternately. It looked like it was taking the ems-pro a few seconds to recognize the VR sensor's signal, but once it did it started working properly. I could crank the engine for 20 seconds and it would continue to work correctly. Fuel pressure was on, RPM signal looked good, INJ LEDs were flashing alternately, etc.
So it looks like I've got a combination that works now, and the only downside is that it takes a few seconds for the ems-pro to see the VR sensor's signal and start acting properly. I think that maybe the 60-2 trigger wheel may help with that. Since they are only $35 from 034 Motorsports I ordered one, and will try to adapt it to the engine next week to see if this improves the situation. I also remember at some point in the past that someone, probably Scott, told me that I could use the VR sensor input signal with a digital output sensor like the Hamlin sensor. I think it might be worth trying that also, to see if maybe the ems-pro will sync up more easily with the digital output of the Hamlin sensor on the VR input circuit.
I plan to spend another couple weekends screwing around with this stuff, while at the same time getting my shop cleaned up. Then I'll pull the motor out of the car and tear it down, to figure out how to address the water leak problem. I'll post any further updates I have on the EFI struggles here. Thanks for all the comments on this issue guys, I appreciate it.