September 12, 2011 - Drag Week 2011, Day 1 and OUT!

[dieselgeek]

Sorry for the delayed response - I'm out in CA preparing for engine masters and in the middle of my own "Jay-style Thrash!"  :-)


OK, so here is how the fuel pump is supposed to work, in software:

IF you have a "priming pulse" greater than zero, the fuel pump control powers up (grounds the relay) for 2.0 seconds.  IF the priming pulse is set to 0.0, you will get NO fuel pump relay control when the ignition is switched on.  That's the only thing that controls the "ignition on" 2 second pulse, or not.

After the 2.0 second pump operation, the EMS turns OFF the fuel pump until the RPM is greater than zero (engine is cranking, for example).

Jay, what I noticed in Topeka was, looking at the LED's that light to indicate tooth pulses being seen on the Tach Input, the EMS seemed to be "resetting" at cranking speed.   What you should see when cranking is LEDs flashing consistently, but yours would flash 3 or 4 teeth then stop.  This explains the fuel pump not being commanded on.

A good way to test this would be to put the wheel on a drill press or mill where you can control the RPMs carefully, and see why we are losing sync at the roughly 125rpm your engine was cranking at.

I checked with the software developers, and one thing they HAVE seen is on high compression motors with missing tooth, wasted spark setups just like yours, is that the missing tooth might be happening right after a compression stroke where the RPM is changing dramatically as the engine is cranking (after all, the actual RPMs when cranking vary a lot, more so with high compression).    They suggested that the missing tooth period is happening too quickly so that the timer that measures the period between crank sensor teeth, isn't able to detect it until you hit higher cranking RPMs.  I believe them.  They suggested re-indexing the crank wheel so the missing tooth is farther advanced than the 45-50 degrees you have it set at now (BTDC).

Unless something changed in the behaviour since topeka, I'd focus on what I believe is the EMS not "syncing" with the wheel.  Although, MANY people run this combo and don't have the issue you do (Bill Fowler, Brian Rock, both with high compression engines as yours, BUT I noticed Bill's cranks around 175rpm).  I am pretty sure the behavior you are seeing with the fuel pump is a result of the EMS not being able to sync up and count rpm properly, at your cranking RPM (i.e., the fuel pump behavior is exactly as expected).   I think the software differences are just that - configuration differences (priming pulse set to zero?)...

What I plan on doing when I get home is testing this on my engine simulator, with the trigger wheel mounted to a variable speed drill press, now I know that the EMS will sync at low low RPMs (one guy with a radial engine cranks at 35rpm and is able to sync up fine, but it's a low comp engine)...  and I'll run your MSQ file on the same version of the box to be sure.

NOW, that being said:  we HAVE seen "firmware corruption" and I can walk you through a field reflash of your firmware.  I have seen it happen usually with a "spark event" (jumper cables creating spark, or shorting out a starter wire by accident).  This is NOT the same as reloading your "tune" it's actually a firmware reflash and worth doing.

I get home Monday and will be on this issue like flies on stink!!  I cannot stand having anyone with confusing issues, we'll figure it out.   Hopefully what I typed above can help a little in the meantime?


Talk to you soon,
-Scott

[jayb]

Scott, thanks for the comments, and sorry to interrupt your Engine Masters stuff.  Everything you said makes sense, so this morning I thought I would head out here to the shop and back the crank sensor off 20 degrees or so.  Before I did that, I wanted to confirm that the gap in the 36-1 Ford wheel is directly over the sensor at 53 degrees BTDC, which is where the timing lead is set in the ems-pro.  What I found, though, is that the center of the gap in the wheel is about 70 degrees BTDC.  Hmmm...  After confirming the 53 degree number in the software, I noticed another software parameter called "cranking advance angle".  This was set at 10 degrees.  Just taking a guess, I rotated the engine so that it was 63 degrees BTDC, and looked at the sensor again.  In this position, the leading edge of the tooth right after the gap is on the sensor, in position to trigger the sensor.

Does any of this make sense?  I thought that when the gap in the wheel was directly over the sensor, the crank position at that point would be very close to the number entered in the software.  That is not the case here.  Color me confused at this point   

What is your advice on next steps?  Should I proceed with backing the sensor off another 20 degrees from its current position and test the setup again for correct fuel pump response?

[dieselgeek]

Jay, the "trigger angle" is the angle before TDC, but remember that we configure "which teeth" are our trigger teeth.  That's why you are seeing the difference, because your trigger tooth isn't the number 0 or 1 tooth.  So, no confusion.

What I suggest is advance the wheel by X degrees, then add X to your "trigger angle" - if the trigger angle exceeds 100 degrees or so, subtract one tooth from each of the trigger points...  and take 10 degrees out of the trigger angle.  hope that makes sense.   


Good news on the EMC front, we found the source of our problems and things are going much better this morning!

-Scott

[dieselgeek]

Jay, check out the spreadsheet I sent you.  A friend made this a few years back, to help un-confuse the trigger wheel setup.   It'll either help dramatically - or confuse further - let me know.  Fighting valvetrain probs in CA this morning, so I have a little downtime.

[jayb]

Scott, thanks for the spreadsheet.  First I tried to advance the crank sensor, but I am physically limited by the brackets to about another 12 degrees.  So I moved it to the far position, reset the trigger angle from 53 to 65, and cranked the engine.  It may have been a little better, but not by much.  Still intermittent flashing of the injector and tach LEDs.

Then I went to your spreadsheet, and set the whole thing up based on that.  My missing tooth lines up with the sensor at 80 degrees BTDC, so starting with that I ran through the spreadsheet and inputted all the wheel decoder settings.  I noticed a couple of things:

- In the Wheel Decoder Settings box, my existing wheel decoder settings had zeroes in all the Trig Return Pos boxes.  I filled these in with the numbers shown on the spreadsheet.
- In the spark settings box, I changed the trigger angle to 60 degrees, as shown on the spreadsheet.  My cranking timing selection had been "Time based", so I set it to "Trigger Return" as shown in the spreadsheet. On the spreadsheet it shows "Spark Output Inverted - Yes".  Mine is set at "No".  I assume I should keep it at "No"?

I tried it like this and still have exactly the same problem.  Go figure...

[Cyclone03]

Jay I have to ask,Would/could these set up problems cause your 100hp drop you had on the dyno?

[jayb]

Good question, but I don't think so.  The reason is that once the engine starts running, everything starts operating as it is supposed to.  The issue is getting everything working properly at cranking speeds.

[jayb]

On Scott's advice tonight I decided to try to crank the engine without plugs, to increase cranking speed.  First I reset the software in the ems-pro to the same as how it was during Drag Week.  Then I cranked the engine, and watched the tach output on the computer.  I got the usual intermittent RPM signal, but when it was working it was 105 RPM.  Next I pulled all the plugs, and ran the same test.  This time everything worked perfectly; all the LED's flashed with no interruption, the fuel pump came on and stayed on, etc.  The RPM reading on the computer was rock steady at 125-126 RPM. 

If an additional 20 RPM during cranking is all it takes to solve this problem, I can see several solutions and explanations for the engine's previous behavior.  On the dyno, I used a Powermaster starter on this engine.  In the car, I used a RobbMc starter.  Maybe the Powermaster starter runs a little faster?  Also on the dyno, I always had my battery charger hooked up to the dyno's battery, on the boost setting, because the dyno's battery has seen better days.  I remember seeing 120+ RPM cranking speed on the dyno, and voltage levels of up to 16V. 

In the car, I could try a switch to the Powermaster starter if it does spin faster than the RobbMc starter.  I could also go to a 16V battery. 

If I do indeed need more cranking speed to make the ems-pro recognize the crank trigger signal more easily, that's not all good, because if I did happen to be down on battery power on the road, the car just wouldn't start without a jump or a boost.  Gotta think about this...

[WConley]

Jay - I'm betting money that the discrepancy in the trigger angle setting on the software vs. the position of the missing tooth is what's causing your power loss.  The timing would be 17 degrees retarded if I'm reading your numbers correctly.

Once you fix the cranking issue and set the software to the correct missing tooth angle, I bet that power magically reappears (after all of the water leak monkey business is fixed!)

[jayb]

I don't know, Bill, the engine ran in 2010 with all the same settings and with the crank sensor in the same position, and made the big power numbers.  I literally did not change anything except for the intake manifold when I first put that engine on the dyno this year.  Same program in the ems-pro, and same position for the crank sensor; I never even unbolted it from the engine.  Nevertheless, all this stuff has got me thinking about anything else that may have changed that I've forgotten about.  But so far I've come up empty.

I've also started thinking that maybe the reason the ems-pro worked so well with the plugs out is that the cranking speed was much more consistent.  This would make the missing tooth much more obvious to the ECU.  The scope traces I have now show the cranking speed varying quite a bit over one revolution of the engine; you can tell because of the pulse width variation from the crank sensor output.  I think this much variation in the pulse width is confusing the ECU; it doesn't know which pulse is the missing tooth, because all the pulse widths vary so much.  I'll try to post a scope plot of the crank sensor output later tonight.

[WConley]

Interesting info Jay.  I've never seen crank speed variation mess up an ECU like that, but obviously it's happening here.  It's also obvious that I shouldn't be betting money on anything...

[jayb]

Here's a scope plot of the crank sensor output.  The variation in tooth width is the result of variation in cranking speed.  There are two missing teeth shown in the waveform,  but I think that with the wide variation in tooth width the ems-pro may not be able to discern which is the missing tooth signal:

[cdmbill2]

Jay, in my recollection my scope pattern (Scott's device conneceted to his laptop) didn't show the amplitude variation I see in yours. I'm using a Proform gear reduction starter at the moment (rated at 1.4 KW) which replaced a Summit Pro-torque unit; both spin the 582/588" at 175 rpm plus. However, subjectively I get some variation in cranking speed which is to be expected.

I also use the older style Aeromotive Fuel pump controller driven by the EMS-Pro's fuel pump output. I disconnected the RPM input when we first converted to the EMS-pro as I didn't want to try and tune around the drop inpressure the reduced supply voltage would cause. The pump has been fine since '07. (not so the MSD alternator, but I digress) I used to also drive an MSD tach signal box for the factory tach which worked fine of that same tach out circuit. I removed it back when we were chasing tach signal noise. It wasn't the problem but I never got around to putting it back.

I get the two LED's lighting off smoothly as you describe in your plugless cranking test during cranking on mine. My missing tooth is set as the seventh tooth or 70* before TDC with the trigger Angle a set at 64. It started even with much lower cranking speeds when the alternator had died and i was on battery only towards the end of Monday's debacle.

Given my recent issues with fatigued connections I'd recheck all of those. My bad sensor signal wire didn't manifest until we hit the rev limiter on the engine dyno. Go figure.

[Kirk Morgan]

That is a very bad pattern. The tops of your square wave look to be broken up. I think the signal gets better from there being a less voltage drop with the spark plugs out rather than the engine speed increase. If i remember correctly when the engine is cranking you should have 1 to 2 A/C volts at the sensor. I would retest both ways with a volt meter hooked up. The 16 volt systems are a PIA. I use a East Coast Electric alt on my race car and i have a constant 15 volts according to my Race Pack data logger. Could you change you timing during cranking to help out? Good luck.

Kirk

[jayb]

I actually used a 10X scope probe when getting the data, so the actual voltage is a factor of 10 higher than what is shown on the plot.  So its more like a 3.5 volt signal, not 0.35.

The fuzz on the top and bottom of the pattern is just electrical noise showing up, and given the sharp transitions at the teeth I'm not too worried about that.  I am concerned about the sine wave superimposed on the signal tops, but really that shouldn't matter to the ECU; it will be looking for a high to low or low to high transition with a comparator circuit.  I'd guess that the comparator circuit trip voltage is set at about 1V or so.

FYI when the engine is running the signal looks just like this, only time compressed and smoother.  In other words, with the engine running all the teeth look to be the same width, rather than varying widths as shown in the scope plot.  I'm pretty convinced at this point that this variation is what is causing the starting problems.

Scott has suggested that I try a Ford VR sensor, which uses a completely different input circuit on the ems-pro than the Hamlin sensor uses.   I'm going to give that a try in the next day or so, and hope it will be more tolerant of the cranking signal variation.