The VeeCU

[WesRagle]

Hi Guys,

I've finished up the VeeCU user interface. As a reminder, I want to make the system as simple as possible. What I ended up with is a panel LED for warnings, a four position rotary switch for cylinder select, a toggle switch for up and down, a store push button to make changes permanent, and a mixture pot. There is more to it than that but I'll talk about that another time.

I took some time to just fool around. I have mentioned this before, but in my correspondence with Dynatek technical support I was told that the recommended dwell time for a three Ohm coil is 8 to 10 mSec and the recommended dwell time for the 5 Ohm coil is 10 to 12 mSec. I had the dwell set to 12 mSec thinking that I wanted the hottest spark I could get out of a 5 Ohm coil. But my test engine, and my Onex engine, (and I think most VWs with electronic secondary ignition) have three Ohm coils. So, I decided to reduce the dwell time to 8 mSec to see how much difference it would make.

I was surprised. I guess I shouldn't have been, but I was. At 4000 RPM and 14 Volts, the current per coil dropped to 1 Amp. That compared to 1.7 Amps at 3400 RPM at 12 mSec dwell. After thinking about it for a minute I realized that the majority of the 4 mSec reduction was time spent just cooking the coil while the core was already saturated (had stored all the energy it could). And, I still had a 2 mSec spark in air. I hope to test that with an air/fuel mix soon :)

Here is what the spark looks like on the primary side of the coil with 8 mSec dwell.



If you looked at data sheet you saw that the IGBT clamps at 450 Volt. Here is a pic of the 450 volt leading edge spike you see on the pic above. It lasts for right at 5 micro seconds. That's apparently how long it takes the voltage on the secondary to build up enough to ionize the air in the spark plug gaps and for current to start flowing.



So, I'm happy to see a near 50% reduction in power needed to drive the coils with only a slight reduction in spark duration. The heat generated by the IGBT will be reduced by the same percentage. I could even hold my finger on the IGBTs at 14v and 4000 RPM. Very cool 8-) (pun intended).

Wes

[gammaxy]

After thinking about it for a minute I realized that the majority of the 4 mSec reduction was time spent just cooking the coil while the core was already saturated (had stored all the energy it could)

You're making awesome progress!

I'm probably being pedantic, but to me a saturated core implies magnetic saturation. From the measurements of current I made, I don't believe this is close to happening. What's actually happening is the coil's resistance is limiting the current that we can apply through the coil at a particular voltage. If we could increase the voltage higher, we would get more current and be able to store more energy.

Without getting super fancy, we're probably just going to be limited by the ~12v battery, so maybe the distinction is moot. I do wish I could reduce the dwell time on the stock ignition to reduce wasted electrical energy and keep the coils cooler.

I don't think that extra 4ms is completely wasted. I calculate the coil has charged to 67% of its maximum energy at 8ms and 86% at 12ms. Definitely diminishing returns as dwell time increases, though. One thing to maybe keep in mind is that RL time constant is in reference to the voltage or current, but energy stored in the coil for sparking is proportional to current squared, so it takes a little longer than the 4.5ms time constant I measured would imply.

By the way, I'm sure I missed it somewhere, but what "trigger" or crankshaft position sensor will you be using? The standard secondary triggers?

[WesRagle]

Hi Chris,

Thanks.

I'm sure you are right. Me, I really haven't spent much time considering the magnetics of the situation. And I don't have a current probe :-( I have spent a lot of time staring at that pic you posted a long time ago.



I blow it up full screen, get out a ruler, and try to figure out how far up the RL charging curve a particular dwell takes me. If in that pic you consider the time the magnet has the current interrupted as about two mSec, then right in the middle, between the two graduations of the first full charging event, would be about 8 mSec. dwell. That puts me at just over 3 amps of the 3.5 amp potential. All the area under the curve to the right of that point represents how much energy is saved by cutting the dwell short.

During a spark event, the squared relationship between energy and current shows up as a reduction of both the duration and the current of the spark.

Here's how it's all shaking out:

1) Below 750 RPM no attempt is made to control dwell. So, all of the spark energy available is used during cranking and starting.

2) At Idle, by limiting dwell to 8 mSec, ignition current draw will be reduced by a factor of 10, compared to no dwell control, limiting ignition current to about 1/2 amp. total.

3) At higher RPMs, the ignition current will still be below 2 amps total keeping both the IGBTs and the coils in a more comfortable temperature range.

By the way, I'm sure I missed it somewhere, but what "trigger" or crankshaft position sensor will you be using? The standard secondary triggers?

Yes, I'm trying to design a system that makes maximum reuse of the existing components. The secondary ignition will be used as the crank trigger and set to 3 Deg. after TDC to simulate an impulse mag during "cranking". Above 350 RPMs the VeeCU takes over ignition timing and immediately advances the timing to 10 Deg. BTDC to push the engine through starting to idle .

It's kinda funny how much effort has gone into the ignition portion of this project. The only reason I decided to control ignition is so I would have enough current left over at idle to drive the fuel pump and injectors. As it turns out, fuel injection is the easy part :-)

Wes

[gammaxy]

I'm glad the graph is helpful. Unfortunately, I managed to lose the raw data or I'd send it to you.

And I don't have a current probe :-(

For $10, I measured the voltage drop across this shunt resistor and used ohm's law to convert to current: https://www.amazon.com/gp/product/B016EBRVGY The measured voltage is low, so some care with the oscilloscope's bandwidth settings are helpful to get good measurement SNR.

One detail I'm curious about from the graph is the rapid current increase to 2A before it smooths out into the typical exponential RL curve. When I recently manually triggered a 5ohm coil by simply connecting a 12v power supply to it, I did not measure the effect.

[WesRagle]

One detail I'm curious about from the graph is the rapid current increase to 2A before it smooths out into the typical exponential RL curve. When I recently manually triggered a 5ohm coil by simply connecting a 12v power supply to it, I did not measure the effect.

I don't know. It might be the effect of the distributed inter-winding capacitance along the length of the inductor. If you remember enough AC circuit theory you might be able to calculate the capacitance by looking at the ringing frequency at the leading edge of the spark. Too long ago for me :-)

Wes

[WesRagle]

Hi Guys,

I was out polishing the VeeCU software getting ready for an actual engine run. I looked at the spark plugs and saw something I didn't expect. One of the insulators of each pair of plugs was black and the other wasn't.



These plugs started out as brand new iridium plugs. Keep in mind that current flows from center electrode to side electrode in one plug and from side electrode to center electrode in the other. In both cases the plug with the dark insulator was on the bottom tower of the coil. I don't know which direction the current flows for that tower. I guess it would depend on which terminal 12 Volts was connected to on the coil. Also, the plug with the black insulator seemed to have a hotter spark, that is bluer.

Maybe someone who's schooled in reading plugs can tell me what the difference means?.?.

Wes

[BRS]

As soon as I saw the picture, before reading the text, I knew what was going on. I've been doing some TIG welding in building a LegalEagleXL. The electrode (tungstin) is always negative when welding in DC mode as the electrons flow from negative to positive. The majority of the heat follows the flow of those electrons.

So it seems one end of the coil is neg biased while the other is positive biased. The dark one would be +Electrode/-Ground. Surprise Surprise.

I'm guessing then your coil has a center tap connected to ground with one spark plug on one end and the other on the other and that they fire at the same time? A coil diagram would be interesting here.

[WesRagle]

Hey Brock,

That's interesting. And yes, the plugs are in series with a single secondary. That is, electrons are "sucked through one plug and pushed through the other. I assume all dual fire/wasted spark coils are like that.

Engine block -> center electrode -> coil -> center electrode -> to the engine block.

So I guess an easy way to remember where the heat is, it's where the electrons land :-)

I did take a closer look at the sparks, from straight above, with a magnifying glass, and both sparks look equally "hot". It's just that the blue end of the spark is at the center electrode of the dark plug and at the side electrode of the other plug.

Live and Learn,

Wes

[BRS]

I assume all dual fire/wasted spark coils are like that.

Wes

Well, not all. Revmaster has a small coil for each spark-plug although it too uses a wasted spark, only it's the primary power wire that is common.