Plasma Ignition

We are working with plasma ignition for turbo-methanol bike. It will give 50-60 us duration spark and average 40 A current  to spark plug.  Basically it is small CDI-box controlled tig-welder. Spark duration is 3-4 crankshaft degrees at 12000 rpm. If spark duration is 1-2 ms as with magnetos spark duration is 72-144 degrees at 12000 rpm. One thing what I  don't uderstand is why nitro-engines need so long duration (and small current) spark to ignite? What is the advance of long duration spark? If ignition point will vary a lot it has an effect to the power. Theoretically short duration and high power (current at spark plug cap) spark is better.

Video
http://www.youtube.com/watch?v=DSoA2RKUv9U

190 mJ MSD + Nology Hotwires


190 mJ MSD + Nology Hotwires +plasma with 20 uF capasitor:



Edited by shmot - 12 October 2009 at 7:25am

Does not look like 12,000 RPM to me.  Must be metric RPM.  Looks more like MSD running multi-spark versus single spark at very low RPM.    I don't understand the reference to plasma. 

Post some more details.

Originally posted by funny1 looks too high tech to me I would just grab hold of them and see which one shocked me the most---the fastest

Similar system:
http://www.aquapulser.com/plasma/radiant_plasma_spark/plasma_products.html

I don't know if that is plasma or what is it, but 10 - 50 kV is needed to create spark and with our measurements it needs 40 V between spark plug cap (0.043") to keep spark alive at 40 A of current. With 0.32" cap it was 65 V. Idea is that usual high voltage CDI-ignition will create spark and after arc-threshold voltage is dropped low voltage (320 V DC) system will supply most of the current to spark plug cap.  In this system 12V to 320V DC-DC  power supply will charge 10 uF capasitor (could be around 5 - 60 uF depending of system) and charged energy will be discharged to spark plug through 3-5 ohm resistor and series of 15 kV diodes. Igbt-transistor will control capasitor and igbt is controlled by using cdi-box coil signal.

That video is just first test with low frequency done by hand. We did test it at around 5000 rpm frequency by using frequency-generator. Next thing is to do pressure test like Mr Geek has been doing and study how pressure would effect to threshold voltage. I don't know if this ever going to work in practice, but theoretically it is better way to feed energy to spark plug. As with tig-welder high voltage is used to create arc and low voltage is used to control current.  In our system current 12 V consumption from battery will be around 40 A for 8 plugs at 12000 rpm (50 us spark, 46 A average current and sequential ignition).


Edited by shmot - 15 October 2009 at 5:58am

The plug your calling out is sealed, so your plan is not to inject a gas that you will ionize to get plasma.   You can see from the patent and the picture of the plug how this is done.  

Hey, you removed your detailed post.  I was going to read it tonight.   

I think we are clear on what your trying to do now. 

It's been a while since I looked at the actual waveforms.   Using the stock MC-4/blaster/wires, etc. what do you measure for the voltage across the plus and current?   Any way for you to post them?

LOL.  Yea, that post. 

When you looked at the signals did you write down what you measured?    I would be interested in the data, even if it's in text format. 


I'll stay off the soap box when it comes to buying test equipment.  Most of my equipment was rescued from the junk pile.   My room looks like something out of a bad 70's science fiction movie.  

The picture of your setup helped me understand.  

 

I read your previous posts again.   Looks like the bench setup is similar to what you described.  A few minor differences, like the 4 parallel resistors in the picture rather than the 3 you described.  Current sense resistor is not shown.   Modified power supply is really just the supply with rectifier and filter cap.   So your just using the AC peak, or 1.414 X 230 AC - drop of rectifier gives you your 320V.   Looks like you then limit the current with another set  of resistors before going into the IGBT.  

 

Why make the IGBT drive the big filter cap?  Maybe it's not.  Hard to tell.   Appears all open loop at this time (so no current control).   What's the plan for the one you will use with the engine tests?  Stay with open loop?

 

Did you take any pictures of the setup when you measured the current?  I would like to see this as well if you don't mind. 

 

Originally posted by geek The picture of your setup helped me understand.     I read your previous posts again.   Looks like the bench setup is similar to what you described.  A few minor differences, like the 4 parallel resistors in the picture rather than the 3 you described.  Current sense resistor is not shown.   Modified power supply is really just the supply with rectifier and filter cap.   So your just using the AC peak, or 1.414 X 230 AC - drop of rectifier gives you your 320V.   Looks like you then limit the current with another set  of resistors before going into the IGBT.

Very good  It was 3 ohm - 5 ohm resistor is suitable for output. With 5 ohm resistor peak output current is around 50 A and with 3 ohm peak current is around 100 A. There are four 20 ohm and 50 W resistors connected to pareller to get 5 ohm.  15 kV, 550 mA continuous and 50 A peak current microwave diodes (NTE517) will give a limit for maximum current. Four diodes in series will give 60 kV protection for system. I will probably use 8 diodes (2x4 diodes in pareller) for each plug and there are 8 plugs in the Vortex II head. Diodes cost $8 each and for final version $500 goes to diodes Between power supply and capasitors (2x20 uF capasitors connected to series = 10 uF)  there is 50 ohm resistance (2x100 ohm pareller) to reduce current peaks from power supply.

Now we could talk:


One thing what is needed to know is coil polarization. This circuit will work only with positive spark voltage. In twin output coil one is positive and other is negative.  I am going to use one coil to each plug and sequential ignition. Two of these MSD boxes (4 output together) are needed to create sequential igition and two coils of each cylinder connected to parallel at primary side (cdi-output). That would give 50 000 V from each coil and sequential ignition will split in half of the current needed from the battery (otherwise it would be 80 A ).


Edited by shmot - 16 October 2009 at 6:22am