Dummy Ignition - VEMS wiki www.vems.hu

DummyIgnition (2004-11-18 00:41:52)

Dummy Ignition means that the ignition driver is a simple (dummy) but fast switch, and the uC (which is smart) takes care of all the timing. This approach can be applied to a distributor-setup, wasted-spark or DIS-style

turning on the [IGBT]
leaving it on for a while (for a certain time, called dwell: which should be higher for lower VBatt) to let the coil charge (current climb) with magnetic energy
than turn it off at the right time, which gives a spark at the secondary side of the coil.

This is a cheap and very good solution (despite the name).

advantages of IGBT-based ignition:

IGBTs are cheap (currently 3..4 Euro in the WebShop). Since dwell-time is controlled by GenBoard/VerThree, there is no need for smart current-limiting (annoying when it limits lower than you'd like) ignition drivers.
you control the dwell (normally in function of VBatt, maybe other variables), so you can adapt easily to a large variety of transformers (eg. I use brand new Lada coil for about 10 Euro) and have total control over spark energy. This saves you from buying the expensive coil that just matces the ignmodule.
no problems with (special ignition-module) connector
GenBoard/VerThree has onboard support for 8 ignition banks (upto 16 cyls with wasted spark).

disadvantage:

you need additional components (measure current) if you want failure indication (but it's true with any other systems)

How can one set the dwell to his coil?

Note: The acceptable dwell range is huge for transformers with coil primary resistance > 1.5 Ohm. Some cavemen used to time the dwell mechanically from the rotor, so the dwell time was 10 times longer at 600 RPM compared to 6000 RPM.

However lower resistance coils can be fried relatively easily, so take care.

measure current consumption

If you want (say) 120mJ spark, at 6000 RPM (100 crank rotations per sec) that means P = 120mJ * 200 / sec = 24W for a 4 cyl (2 sparks per crankrot). Preferrably do it on table first, as on OnlineCourse/OffEngineTesting. No need for all 2..4 coils of course, but consider this in the calculation: calculate with actual_number_of_transformers * actual_spark_freq (h[2] table comes to mind) instead of the 200 / sec above.

With 85% guesstimated efficiency and 14V supply you'll want to adjust dwell until you measure

I = 24W / 0.85 / 14V = 2A average (mean) current (DVM in 10A .. 20A mode).

go from low dwell to higher
first fine-adjust dwell14 at 14V VBatt
after the 14V is set, fine-adjust dwell6 at 7..10V VBatt
check again at 14V
make sure you have some reasonably small spark-gap (<2 mm gap) in any case so the clamping IGBT is not stressed (the IGBT will dissipate all the energy with too big sparkgap): I use bent header-pins, but a few 2W 100k resistors in parallel can be just as good; oldschool teaches to short the plug-wires to GND (eg. when removing from the sparkplug, or on table): that will dissipate all the energy in the coil, which is acceptable, and is a simple solution.

It's wise to calculate max current:

Imax = (Imean / num) / dwell * period * 2

(Imean / num) is the mean current for just one IGBT ( num is the number of drivers you use in the experiment).

Period is related to (60000 / RPM) msec (and consider number of cyls). If the max current is above what the IGBT can handle (12..14 A generally) start to worry.

Just to be safe, possible to check IGBT temperature soon after shut down. You can die from electric shock if you touch it while it's running. Above 70C it burns your finger (but IGBT can survive 125V junction temperature).

Other methods

I try to decrease the dwell until I get misfire (also take care), so I know what range is good (eg. multiply the misfire-threshold by 2). You can also look at the spark (be careful!).

Too much spark energy will result in:

faster sparkplug-wear
warming coil(transformer)
more energy consumption

These are not major things, but I don't go for skyhigh spark energy.

If you have a scope to look at the signal, you can connect a (say 0.1 Ohm) series resistance in series with the coil (eg. at the 14V side), and measure current, and tune it as you like. (take care of high voltages, it can damage your scope or you!!! When measuring voltage, measure only the primary side of the coil, and use voltage-divider from megaohm range resistors, check the range of scope and cable etc.. etc..).

Some more wisdom from Marcell on finding correct dwell:

you must tell the dwell in config. The firmware cannot find out by itself what dwell you need. Depending on your module: 800 usec .. 5000 usec. The lower will not fry the coil, but make a lower energy spark. There is some (light) relation between input resistance and L. There is a direct relation between L and output Energy: E = L*I*I / 2

What can we do (besides hunting specs):

measure DC resistance (DVM, ohmmeter) of primary (0..20 Ohm) and secondary

(0..20k) winding

you can measure many things by applying 50..60 Hz AC to the transformer.

I applied 230V, 50 Hz AC from the wall (this is dangerous) to the secondary of my coil, and measured about 2V AC on the primary. This tells me that at 23kV output I can expect 200V on the IGBTs. You can connect 5..20V 50..60 Hz AC with a limiting R (220 .. 1000 ohm) to the primary, and measure (AC) voltage on the primary and the resistor (the resistor voltage tells the real AC current). knowing the freq (50 or 60 Hz) we can calculate the L of your primary winding.

Many recent transformers for electronic-ignition systems have very low impedance primary winding, many times below 1 Ohm. These like a short dwell, peak current before spark can be upto 8..10A and can fry if left under power continuosly. We should open an IgnitionPage/TransformerPage, and put up these stuff...

Feel free.

others: make adjustments, corrections, links..

There are transformers for older "non-electronic" systems, they usually tolerate leaving them on 14V all day long (max current: 1.. 2A). I must mention that the voltage protection to limit voltage on the transformer is needed in that case too, do not simply play with it with your battery + wire: For the table-tests, I recommend both a secondary-side sparkgap (I used slightly bent header-pins) and 400..600V diodes on the primary side: 1 is OK if you figure out polarity, but 2 connected in series with reverse polarity works too. Without any protection, the coil can fry its own insulation, zzzzz bad. Your coils are probably in between. If they are intended for current-limiting setups, recommended current before spark maybe 2..4A. Note that the square value (I**2) of the current is important (both for heating and for spark energy. => between 2 and 4 A the factor is 2*2=4, between 1 and 10A the factor is 100!) If you connect a 0.1 .. 0.2 Ohm Rcurrmeasure 2W resistor in series with primary side, and start from small 800 usec dwell, and scope the voltage on the Rcurrmeasure, you can find out how it's doing.

Determining config.txt values needed once you know desired dwell.

For an example let us presume you find that 4.5ms(4500 micro seconds) is good for you:\nÿ1ÿ

With factory modules you have less parameters to set, more documents to read, more to pay, and harder to debug. Just my opinion. Those modules are there for your mom :-)

http://www.transpo.de/catalog/E_specs/BM300.htm

Answering to a question:

I don't see if adding IGBTs to an EDIS module (EDISIgnition) has much benefits, since EDIS already has its output drivers, and with EDIS you can't have the above advantages.

You need some InputTrigger to do this (you need some anyway :-) [See how efi332 is doing this ]

A few links dealing with the subject of measuring coil saturation time (dwell) and obtaining the right dwell settings:

http://www.tonyfoale.com/Articles/Ignition/Ignition.htm Most modern coils will go up in flames if you try this. Remember that semi modern coils have a resistor in series to allow them to be used with an old ignition system. IRL very few coils want more then 3mS dwell today.

http://www.asashop.org/autoinc/feb97/DIS.htm