Tofuball (Jason) and I finally got a chance to test the new megasquirt trailing ignition code on his car today. It works GREAT! I mean it doesn't really affect performance noticably, but it works.

I tested with several different ignition advances and splits with the timing light flashing at both the CAS (to make sure the right trailing coil was firing at the right time) and at the pulley (to make sure the timing was right) and it worked no problem.

The sparks the trailing coils fire are ridiculously strong though... I could hold the wire like 3 or 4 inches from the car, and watch the sparks fly to the chassis.

Details:

Using the wheel decoder with the following settings:

12 + 1 decoder
triggers/returns

trig A: 1,3
trig B: 7,9

trigger angle 60 deg
trigger return cranking
hold ignition 2 seconds

trailing split on
FC type ignition outputs

I'll post a full MSQ once we retune his car (the tune is way off now since he finally got a good fuel-pump and fuel hose, meaning he finally got good fuel pressure).

We also need to redo the PWM idle valve settings (hooked to the BAC valve) so that closed-loop and warm-up idle work right.

Oh, and moderators; thanks for finally creating a megasquirt section.

Ken

EDIT: I forgot to mention that this was with msns-extra firmware 025k, with a dual lm1815 vr sensor conditioning circuit to condition the CAS signal, and 4.7kohm pull-up resistors for the output signals to the ignitors. 025k is the first msns-extra offical version that has working support for rotary trailing.

 

Recently you mentioned I could use one of the V3.0 on-board vr circuits and put the other one in the proto area. Is there any benefit to putting your two vr circuits on the proto area and leaving the on-board one available?

I'm also not sure what outputs this setup will take. Will I still have the four standard outputs availabe with your ignition setup?

Thanks,

Scott

 

I can't think of any benefit to not using the v3 vr conditioner that's built in... and the proto area may be a bit too small to fit the whole dual lm1815 circuit.

do you mean the 4 outputs on the db37? What I'd do is use one of the db37 pins (and a ground) for the second trigger input, and then you still have 3 more outputs on the db37 left for the three ignition signals.

If you want any more outputs out of the megasquirt, you'll need to wire them to another connector or something like that. For the kits I'm building for people on the v2 boards, I've been running the vr sensor inputs into the megasquirt using the 4 extra pins on the db37, and then running the ignition and other extra outputs (vdi/aux ports) out through an extra db9 or db25 mounted either on top of the case or on the front in place of the LED's.

I need to get to writing a complete how-to on modding the megasquirt for the rx7 sometime in my non-existant spare time (my current projects now that trailing works include getting stock rx8 ignition to work, and modding the staged injection code to include a configurable amount of time before the primary injectors cut their pulse-width after the secondaries come on. This will help get rid of the small lean-spot when the secondaries engage).

Ken

 

Ken,

I wanted to use outputs 1 and 2 (X4 X5), output 3 (Pin 15 of the uP (U1)) and output 4 (LED 18) to run solenoids. Will any of these become unavailabe using your rotary ignition?

I also wanted to run the BAC like you were talking about with the PWM signal for controlled idle speed, I think that's on the FIDLE output right?

Scott

 

All three LED's are needed for rotary ignition if you want to run leading and trailing. If you don't mind running leading only, then only led17 is needed for ignition.

If you're trying to pass emissions, I'd suggest just creating a map that runs at 14.7:1 up to about 65-70kPa, and then 13.2-13.8:1 above that just for emissions. I've heard of people going as high as 15.5:1 at cruise for economy purposes on the rotary engine.

As for the BAC valve, you need to get a ztx688 transistor and replace the current fidle transistor with it, and then replace the fidle resistor with a 1/2 watt 510 ohm resistor, and then you can just connect the fidle output directly to the BAC valve.

 

This is GREAT news! I'm working on getting my RX7 running with this code and the dual VR conditioners...is anyone working on a full write-up on this? If not, I'll go ahead and start once I get mine working.

 

I've got a write-up that I wrote months ago. It's like a first-draft and doesn't cover rotary trailing.... but the settings for that are pretty self-explanatory.

Wheel decoder howto

It covers some basic wheel decoder settings, and the logic behind those settings and how to install the CAS to work with those settings. I tried to make it more generic, but I used the rotary as an example. It also has instructions on how to wire up the lm1815 circuit to the megasquirt.

If I have time I'll write more and go over other things that people have questions about.

Ken

 

Let me make sure I get this straight (and I do have a few questions):
I need to solder 3 4.7k resistors from the negative leads to right before the input resistors, correct? And I solder the wires for the coils to the negative lead also?

Where do I get power for the dual LM1815 circuit? Right off the power regulator(U5)?

Where would you recommend I get ground from?

I read someone in that post that I need to put a 1k ohm resistor across the grounds coming in from the CAS, is that correct?

Solder to pin 11 on the ecu...I assume you mean on the bottom of the board and not directly on the chip itself.

I put together my LM1815 board today....tomorrow, pending your response, I will hook it up to the board and test it with a spare CAS.

 

yes, both of your statements above are correct. It's a +5v pullup resistor. The reason for such high resistance is that the mazda ignitors have relatilvely low input resistance to ground, so you need to limit the current draw.

there is a small via just above D9 that you can take +5v from, that's generally what I use.

for ground I usually use the D1 or D2 bottom hole.

Nope, it's not necessary. I didn't do that. It's only really necessary to do that on the toyota vr sensors as they have 3 sensors sharing a common ground.

Yes, I usually solder both the u4 pin 6 and cpu pin 11 wires on the bottom of the board.

 

Great news and thanks for all of your work. I will be updating my MS as soon as possible and installing it after this season ends. It will be great fun to finally set the car up right for the 1/2 bridge. Again Thanks from all of us racers who are desire rich and cash poor.

 

Any fuel economy gain with the trailing hooked up?

 

Actually economy might have gotten worse although I haven't asked tofuball about that.

I noticed some interesting things though... with trailing on at 16 split at low load, the engine seems to need more fuel to run smoothly. We were running 14.7:1 at high vacuum (low kPa) and low-mid revs (1500-3500 rpms) before trailing, and it ran fine, but once we added trailing, it stopped running right at that range of kPa's and rpms.

I did some research, and most people with trailing working are running 13.2-13.8:1 at those ranges... with 14.7:1 in a diamond shape in the middle of the map (basically at cruise ranges). So I guess it won't do anything to his cruise mileage, but it'll probably hurt his high vacuum, light throttle mileage.

Ken

 

Ever played with low or no split in those conditions? Seems if it runs well on light throttle with a leaner mixture and no split, what's the advantage of split in the first place?

 

Yeah, didn't seem to make much difference. It runs well at lighter throttle and leaner mixtures with no trailing ignition at all. With trailing enabled and a lot of split or a little split, it just needs more fuel as far as I can tell.

I'm not even sure why mazda put trailing on there because it sure as heck didn't help power on tofuball's car... at full throttle it seems it didn't make any difference at all (at least on the "butt dyno").

 

I always thought the split timing was just for emissions to reduce the unburned hydrocarbons.

 

it's possible. I don't know the reasons for it, I just know that a lot of people wanted it to work right, so I fixed the code for it at low revs

 

Bah, if that's the case I'll probably forget making it a priority to get split working at all. There is even a device made for FD's called the anti-det, and it's basically just a blind spark plug to fill the hole in the housing where the trailing plugs go. suppedly it reduces the detonation problem with boosted rotaries, but it's all controversial as far as I've read on here.

 

I heard about that... basically they do that because a lot of the spark-retard devices that people use for leading don't work for trailing, so it's safer just to get rid of trailing altogether rather than having trailing fire before leading.

However, the way the MS code is written, the trailing spark is scheduled using a delay from leading, so when the leading is retarded, the trailing will still fire at the same time as, or after leading (depending on what split is set up in the split table)... there's never a chance of trailing firing before leading.

So I'd say it doesn't hurt anything to set it up, and it might help a little, so there's no reason NOT to set it up unless you just don't feel like it

 

http://smrmicro.com/re-ky.pdf , PDF pages 48 and 56 (and then some after each of these) has interesting engineering data on All Things Spark.

-Mike

 

4.7.4 in the above doc is quite interesting as well.

-Mike

 

Hey that's a great find! Wish I had that a few years back.

Thanks Mike.

 

this is an awesome document mike! This has already helped me in figuring out why we're seeing some of the weirdnesses we're seeing at light load (not ignition related).

 

Cool. Glad it's of some help! I have more stuff like that I'll try and get up on smrmicro.com tomorrow, when I get home late afternoon and have full net access - I'm web only right now and can't u/l stuff from my laptop, working off Dad's machine right now, no wireless here and no router here :-( . Ken, please gimme a call at my cell, if you would, when convenient, just need your number - or PM me. My phone was stolen from the fire truck I work off of (scumbags!!) and I lost many important #s. I'll be keeping an actual hardcopy from now on. If you need my number, PM me. As always, thanks for everything, your code kicks *** so far!

-Mike

 

so I'm assuming you're actually running the FD style outputs now in the car? I'll give you a call tomorrow as it's getting a bit late tonight.

 

I've run both FC and FD style with ugly temp wiring and coil(s) mounting for just a trip or two around the block. I think I'm going to go with FC mode. I'm going to mount a 921 where my HEI7 is now for the leading http://smrmicro.com/v-web/gallery/Ma...oject/DSCN2967 and use the FC trailing coil pack to manage trailing. This will keep all high current, nasty EMI inducing current paths farther away from the ECU. Also allows me to get cool air for the intake from the old coils mount holes (FB). Thoughts?

-Mike

PS: tomorrow afternoon will be good for me for a call.