My 87 toyota supra hardware build

[via]

Okay, I'll get on that. What is the proper way to adjust dwell time? I'm not sure what the 'T' units are in the fixedconfig2 file, but that doesn't seem like much granularity.

[Fred]

trace should look like the image at the bottom of this page: http://omegaignitions.com/faq.html

You're looking to locate C for several voltages.

Here's a crappy image of the situation: http://www.perfectpower.com/images/_man ... l-time.png

Only get as much flat at the top as you need to realise it's flat. Verify you have minimal peak voltage across the ignitor (it's driving the coil properly).

[Fred]

T is a table macro, the short hand. Those are milliseconds. You can use decimals: T(1.2334) will work. BUT, just leave it table sourced and edit the table live from emstudio instead.

[via]

Okay, I got some graphs:
http://imgur.com/a/KsP5W

All graphs of igniter->coil current
First image - As I last configured the firmware with 2ms dwell. Looks too short
Second image - Reset back to use BRV table, looks to too long, I guess about 4ms is good.

I altered the table around the 12.2 (current BRV) to make it 4. I'm doing this with the understanding that I'm going to have to fine tune these in the future.
Third image - 4ms dwell. Looks like it could be the tinest bit shorter to me, so...
Fourth image - 3.8 ms dwell. How does it look to you?

I don't have a ready means to make anything but the battery output, unless you include 5V, but I don't think any supply I have can source enough current. I think I can borrow an adjustable one tomorrow and set more data points.

That seemed pretty easy assuming I'm doing it right. Does this suggest the igniter is good enough for my coils?

[Fred]

I would assume the ignitor is doing a good job based on those curves. 3.8 seems ideal at 12.2 if 12.2 was your current voltage. See if you can get it on charge and find the 14.4V dwell time, and drain it down low and get a ~10V time too, with those you can get the important part of the curve about right. Whatever values in between you can also come up with, all the better. Nice work. :-)

So as a summary:

A curve that crosses through 12.2V/3.8ms with this ignitor-coil combo could be close to ideal.
A high pulse into the ignitor dwells the coil.

More stuff to consider:

How much current was flowing into the input of the ignitor? How were you driving that? 12v + current limit? 5V? Or what?
How much Voltage was dropped across the ignitor at static peak current with the topology you used?

If the ignitor is fairly normal, then 12v+current limit = correct. If it's logic level, then that may still be considerable as correct, or not.

If the answer to the second question is too high, it indicates that the former is setup incorrectly.

This kinda shoulda been setup/determined before the above testing, however if you determine it was sub-optimal, you can repeat the above tests once it's fixed.

Fred.

[via]

Can you clarify what currents/voltages I'm measuring? Input being the line to the coil, or from the ecu?
The connection from the ecu is using a TC4427. Vdd on the chip I had wired to the 12V supply from the battery. I measured the voltage drop across the line to the coil and ground on the igniter and found it to be about 1V whether 12V was used, or the 5V from the regulator. My current measurement is pretty imprecise, but it looks like the line from the tcc4427 to the igniter is about 30 mA when on (at 12 V).

[Fred]

What current limit resistor did you use with the TC4427?

Sounds like the above results are good, anyway. Nice to see someone doing it properly! A couple of more voltage data points and you're good as gold! :-)

[via]

Output from the tc4427 goes through a 150 ohm resistor to the igniter. I more or less copied ravage but without the xor'ing and selectable voltages/inversion.

Well thats all good. Tomorrow I'll get proper wire for the new harness and start making low side drivers for injectors/fuel pump. My MAP sensor came in the mail, I'll look at finding a place for that too.

[Fred]

Perfect.

"low side drives" for those things are just an autofet and a pair of resistors, pretty simple stuff.

[via]

I rigged up a little potentiometer-adjustable lm4791 circuit with in-series 12V batteries as the supply to make adjustable 5V-15V. I wired that to the B+ on the coil and B+ on the igniter. However, my figures are a little different than what I was getting last night.

I started with 12.2 V again, but found a dwell of 2.6 ms instead of 3.8 from last night.
10 V required 3.1 ms, and 14.7 V required 2.1 ms. What could be going on here? Last night the igniter and coil were directly connected to + on a 12V battery registering about 12.2-12.4 V, and this is through the lm4791. I could understand the latter artificially increasing dwell time, not decreasing it. Maybe my resistor last night for current measurement was more than I thought. Could a couple of tenths of an ohm make such a drastic difference?

Either way, that looks like 31/V = dwell ms, so I'll go ahead and populate the table around that.

Also of note, the ecu was doing a dwell of about 6ms despite the table showing around 4ms for the current BRV (according to the numerical data page). Could the spark be messing with the BRV input to cause that?

Edit:
I loaded my new dwell table into the flash, and ran it with the same resistor as earlier today but with the 12V supply from the battery, and all the dwell times are too short now. So the resistor being different was not it. Could it be the filter capacitor I have on my power supply output I made earlier? 47 uF, a bit on the big side for this.

Another edit:
I changed out the 47 uF for 200 uF and it brought the dwell times closer (not quite all the way) to what they were for the same voltages from battery. With that, I've used last night's data point with today's slope, and I think that'll be good.