Ion sensing. Finally did something.

[dnb]

I have just made myself an ion sensing circuit that doesn't need a high voltage power supply.  It "steals" power from the coil during the dwell and stores it in a capacitor, and then puts a 200v bias on the plug during the power stroke.  I can then measure the current through the circuit and derive information about combustion.

So far, this has been tested on the bench with my spare VEMS, my ECU test box and a single coil & plug assembly removed from the TVR.

The circuit survived this test and produced the following waveform from the measurement pin:


You can clearly see the coil charge time and ringing during the spark event.  (The capacitor and sense resistors make a nice LCR circuit with the coil secondary...)  But since the test was to fire a spark plug in free air, there were no ions to measure, so no current flowed after the spark oscillations had settled.

The next test is obviously to try this on a running engine, which I should be doing very shortly.

Has anyone else experimented with this?  Specially the version without the high voltage supply...

[MWfire]

yes.
http://www.vems.hu/wiki/index.php?page=MembersPage%2FDamirMuha%2FIonsenzor

[multiplex]

interesting in seeing more results on the car

nice job!

[andreNL]

Interesting subject nice job ...

[billman]

hang on this please.....sounds very promising.

[dnb]

I have made a little progress on this in between being a bit too busy at work.  I now have a good circuit board and box containing 8 channels worth of ion sensors (fits on to the coil bracket on the engine) and have started to get the processor box designed.  This should sit nicely next to the VEMS, although I haven't considered interfacing at all yet.

I'll post some pictures very soon, I promise!

[billman]

Very nice!!!

Noob question: Does it connect t othe spark plug through the coil and lead or does it come between them?

http://delphi.com/manufacturers/auto/powertrain/gas/ignsys/ionized/

[z0tya]

I have made a little progress on this in between being a bit too busy at work.  I now have a good circuit board and box containing 8 channels worth of ion sensors (fits on to the coil bracket on the engine) and have started to get the processor box designed.  This should sit nicely next to the VEMS, although I haven't considered interfacing at all yet.

I'll post some pictures very soon, I promise!

We are waiting for..

[dnb]

Well, almost a year on now.  Work's been a bit quiet for a week or two (calm before the storm...) so I have taken the chance to look at ion sensing again.

I have got around to testing my ion sense box on the bench, and also to how I process and analyze the data.  Given the data rate is going to be of the order of 2Mbit/second, and near real time processing would be highly desirable it's a bit of a challenge.

I have sourced some tolerably fast 16bit ADCs, and have created software to translate a crank trigger source into a crank angle in real time, and then plot the ion sense data against crank angle.  The test software does one channel, but there's nothing to stop it doing all 8, other than me not having soldered all the channels together.  An interesting use of some old sound recording software - code reuse is apparently a good thing, my software guys tell me...



This shows my software doing a quick-and-dirty line plot of ion sense current vs crank angle (720 degrees and wasted spark) on a plug spark in free space.  Naturally enough the interface will be improved, but the real-time aspect has been proved. 

(There's nothing to stop me using this for knock sensor analysis either.)

[rob@vems.co.uk]

Awesome work.  As usual

[dnb]

Tried this on the car today.  The first problem is that I hadn't provided enough attenuation for the crank trigger, so it saturated the sound card input.  Will have a new attenuator network built by tomorrow...

I did, however get a nice ion sense trace out of the one connected cylinder.

[NOTORIOUS VR]

awesome!