Ion Sense

[JohnP]

Please note, my intent is to provide knock control, and the detection of misfire and pre-ignition, for engines that are too noisy to use acoustic knock sensors.

Primarily, I think these would be racing engines with high clearance forged pistons. Drag racing, probably, as they don't want the engine to come to full operating temperature, where the pistons expand to stock like clearances.

I think another application could be offshore power boats, where they are under sustained load for long periods, where pre-ignition could melt them down in a few seconds. At present, there are no pre-ignition detection systems available, outside of automotive labs.

[JohnP]

I have been reading and gathering patents on ion systems since 2000, when Garfield Willis was active on the old DIY Fuel Injection forum, which later became mega squirt.

Willis was a retired engineer who read and could understand patents by Bosch on wide band O2 sensors, and the Mecel ion patents. He stirred up a lot of interest on the forum in his projects, which he called "EGOR" for the wide band O2, and "Ion Eyes" for the ion sensor.

The forum was supposed to be open source, but he went private with his findings and started marketing the first wide band O2 meter. Then he died.

I understand your longing for PPP control where everything is nirvana, and it's free because it's on the internet, just SOMEone, ANYone, program the damn thing. Here are the waveforms, just DO IT.

But real life gets in the way. I should have been farther along by now, but I had to stop work on it for three years as my wife Shirley lost her battle with cancer, then I lost another year when I was hit by a car while walking and was out of commission for a year.

But as you see, I'm back at it. I'm 78, may not reach my goal, but for now it keeps me out of the pool rooms. Oh, I still go.

[JohnP]

I think this is the original run where I was recording the waveforms on my SONY DAT machine, seen in an earlier post.

I was recording my scope with my phone. It's a lovely little scope, a Leader LBO-315, I paid almost $2k in 1989. A few days after this, I decided to spring for a bigger, brighter scope.

[JohnP]

Mike, forgive me for going off, I didn't mean you in particular. I was on a riff and just went for it.

But you see the difference between a crude system and the lab. There is no way I could develop a control signal from "the second peak" or the "third peak". I don't see them. The signals are not stable enough. I'd say it's doing good just to detect knock.

[JohnP]

This was my first prototype ion processor box, in 2018. Shirley and I actually ran it on a car in 2016. It "had issues". Proto #2 was the one on the 1990 Suburban, generating the waveforms shown earlier, in 2022.

[JohnP]

This was my bench before I designed the ignition tester board. I just kept adding proto boards until it was filled. I finally took a couple months off to design the all in one board.

So, when is PPP coming? I don't know but it was a lot of work just to get to this point. So I kind of roll my eyes when I see that question on ion threads.

Plex, DIY Autotune, or another worm crawling around in the woodwork, like me?

[JohnP]

Ed Van Dyne developed the Smartfire CD ignition with ion sensing in the early 2000's, and eventually sold it to Woodward controls. I don't know if it's still in their catalog.

The first time I saw it, asking price was $22k and that didn't include the knock sensing software. I spoke to someone who had tested it on the dyno and his comment was it didn't work that well. So whatever I come up with has to work better than that.

I'm not a user software guy, so it won't have all the data analysis stuff. But it should, I admit that. My goal is "a simple box" that shows misfire and pre-igntion, and controls knock.

[mpgmike]

Nothing like this is showing up on their site, even using their search feature and "SmartFire". I guess they gave up. The above document is dated 2005.

[JohnP]

In 1996, I developed a display with eight bargraphs, to show the amount of knock retard in each cylinder. It connected to the SafeGuard and was synchronized to the engine with a timing light pickup on #1 spark plug wire.

I was getting ready to start showing it around when I got a call from a guy that said he builds engines for NASCAR. He asked about the SafeGuard and I told him about the new display. He said could you build me one with eight knobs, so I can manually retard each cylinder on the dyno.

I remember I was on the couch having a bite to eat. Swear to GOD, I started sketching the block diagram on my paper towel. I said, yeah, I can do that. He said he needed it before Daytona. It was sometime before Thanksgiving, and I think Daytona is in Feb?

I told him I would try, then got to work. I was a couple months late, so he was no longer interested. I did, however, sell one to a guy that worked with Bill Elliot.

The system was later displayed at the 2000 SuperFlow engine conference, where someone from Richard Childress racing called and asked to borrow it for a week. I was hoping that SuperFlow would test it, but that never happened. Childress racing tested the manual control feature, saying with restrictor plate motors they don't get knock.

At the time, they had ten dynos running, so I was picturing each race team ordering one unit per dyno. I called to talk to them after they returned it, they said it did everything they wanted but they were going to use something from MSD. Or Geico, I forget. I should have charged them a rental fee, but I wasn't business savy.

We showed it to Kenny Duttweiler, who used it to tune a turbo engine for Hot Rod magazine for their "Power Tour" series. So were were in two issues of HotRod that summer. In return, we offered him the unit with the eight manual retard controls, he said no I want the little one, and he named it Son of Knock Finder. He got the only one I built.

I sold a grand total of six of the larger units, world wide. I was only asking $2500. My last sale was in 2001.

[JohnP]

World Headquarters is a spare bedroom in an 1100 sq. ft. condo in Garden Grove, CA.

[JohnP]

I found links to SmartFire, a few years ago, on archive.org:

https://web.archive.org/web/20040525182024/http://adrenalineresearch.com/racecar.pdf

https://web.archive.org/web/20040114023033/http://www.adrenalineresearch.com/price.htm

[JohnP]

MIke, I re-read the price sheet and just noticed they offered software for PPP:

"Ignition Timing Optimization Software, Includes Knock and Peak Pressure Control", very cool, but no prices are given for the various software packages.

So at one time, PPP was available. I have trouble selling my stuff, they had trouble selling their stuff. Marketing is the key.

Ed Van Dyne, while at Adrenaline Research, was trying to market it to Formula 1, NASCAR, OEM car companies, etc. He eventually found a buyer, in Woodward, but they couldn't do anything with it. Unless they have incorporated something similar into a later product line.

[JohnP]

Ah, finally, I have the noise blank pulse working. I'll make another video tonight to verify we have a clean knock signal.

Last week I thought I had pre-ignition, but it turned out to be ignition noise getting into the knock signal.

[sepp2gl]

Hi John,
I also read the paper.
Yes the shape of the ionization current is shown in an idealized, typical way. It is NOT based on any specific measurement.
This is how scientific work is documented and communicated.
Nevertheless, an ionization current signal contains those three phases
1. spark current, it doesn't tell you anything about combustion metrics.
2. chemical ionization caused by the flame; this can be used to detect misfire. It is a mandatory part of the signal.
3. thermal ionization caused by pressure increase; this part contains engine-load, combustion-phasing and knock information.
The ignition setup is not shown in the paper, because it doesn't matter for the results shown there-in.
If there was a second spark-plug used, the author wouldn't have shown the first "spark current"-phase.
Definitely THERE MUST BE A 3-PHASE ION-SIGNAL.
In low-load or idle conditions the 3rd phase can be very low, so you might not see it in the measurements.

Coming to your videos: they are not accessible from here.

Concerning your measurements: I simply do not know, what you show in the screenshots.
You need to explain, what signals you are showing, so that other people can understand your findings.

Please don't take it as criticism, but as proposals for improvement
T2US, Gerhard

[JohnP]

Hi Gerhard, great to hear from you.

You didn't get to see anything? How can I send the videos to you?

Thanks for the tip off about the first phase, being ignition. Normally they also show a stylized ignition pulse, too.

I used a file sharing website to send a movie I have on DVD to a friend. Great movie, by the way, The Mask of Demitrios. Peter Lorre and Sidney Greenstreet.

I would need an email address to send the file.

[JohnP]

Gerhard, are you getting a notice the video isn't available in your region?

[JohnP]

Youtube studio shows only one or two views of my videos.

No one is seeing them or just not interested?

Weird because I know there were over two thousand page views of the thread since I posted them.

[JohnP]

Gerhard, are you asking about the screenshots showing knock detections from my ion sensing distributor cap?

The system is installed on a 1990 Chevy Suburban with 454 engine. Part of the system is an MSD Digital 6 ignition. The truck is not licensed to drive on the street so all we could do was load the engine against the brake, max RPM was about 2000 under load.

At this low RPM, the MSD makes two or sometimes three sparks so there is a lot of ignition noise to deal with.

I recorded the signals from my ion processor box on a SONY Walkman DAT TCD-D7. One channel had the raw ion signal, the other had ignition pulses.

I made a proto board interface so I could play the tape back into the processor box on my bench for viewing at my leisure. I used my iphone to make a video of the output of the processor box, then scrolled through the video and found some nice clean knock detections and took screenshots.

[JohnP]

I love doing layouts. This is the ion processor V2.

[JohnP]

Gerhard, I sent a PM with links to downloadable files on icloud, containing a few videos.

Comments on the screenshots:

1) The large mass of cables (25') is for testing the distributor based ion system on the dyno, which has not been done yet.

2) Shows the rear view of the ion processor box. The connector on the left has two input pins for a knock sensor, eight input pins to receive the ion signals, in firing order, and three output pins to power the ion sensing distributor cap (12v, Gnd, and -100v).

The connector on the right has inputs for a "spark fired" signal from a trigger module attached to the coil, VBatt, Gnd, ouput to trigger MSD White wire, input from points, digital input, pseudo magnetic pickup output, magnetic pickup input, Gnd, TDC input, alarm output.

The black thing in the center is a fiber optic input from a trigger on #1 spark plug wire. In addition, there is an audio jack to listen to the acoustic knock sensor, a volume control, a switch to listen to raw knock or band limited, a switch to select manual or auto AGC on the ion amplifiers, a manual AGC level control, a knob to adjust the knock window plus or minus 5 degrees, and some mode switches to select the number of cylinders (I forget what else the switches do). There is a USB output, from an onboard OEM 8 channel fast data logger, to record the ion signals.

3) The front view of the ion processor shows two rotary knobs, one knob controls a sync output for your scope, to sync on any cylinder.

The other rotary knob selects which signals are to be viewed, such as ion signal, ion knock pulse, acoustic knock signal, acoustic knock pulse, and a signal from a low pass filter of the ion signal.

The black rectangle is a Gabatronics OLED oscilloscope.

The knobs and switches allow you to vary the gain of the amplifiers and set the operating modes, such as knock retard by acoustic sensor or by ion sensor, disable knock retard.

The LED array shows knock, misfire, and pre-ignition for each cylinder. A knock retard bargraph shows the amount of retard.

4) Vampire test shows my ignition test board connected to a first gen Vampire, for test. The ignition tester has inductors and associated IGBTs to provide flyback signals to trigger the Vampire.

The Vampire is a "piggy back" ignition timing controller to retard timing when knock is detected, for those that add forced induction or nitrous oxide injection to their car. When it detects knock, it grounds that coil before the spark can occur, then releases the ground later, allowing the retarded spark to fire.

5) Shows the ignition tester connected to the ion processor box, providing simulated ion signals for testing operation.

6) Shows my test bench before I designed the ignition tester board, in mid 2018. It just kept growing with plug boards, etc, as I was designing the ion system.

[mpgmike]

JohnP, I really wish this Forum had "Like" and "Thanks" buttons. Just know at least 1/2 of your posts would have been Liked/Thanked. I really appreciate you sharing your experience with those of us who have spent countless hours poring over the lab reports, but with no hands-on. You have given me insight and hopefully brought me closer to getting a handle on this Ion Sense -- REAL WORLD.

Thank You!

[JohnP]

Mike: You have no idea what your attaboy does for me.

I work alone, never get to brag to anyone. Most of what I do I try to hide until it's time.

[JohnP]

Mike, speaking of countless hours, did you check out that layout? And there's a schematic that came before that, and design, and experimenting before that. Several man years.

The distributor cap was what started it all.

[JohnP]

Of course, no one will want the distributor cap version, so most of that was just a hobby effort and learning experience. They will want coil on plug.

The IGN-1A coils have an isolated coil secondary, perfect for ion sensing. But I always look for something different. A harder way, because there is less competition that way.

[sepp2gl]

Hi John,
thanks for your videos on icloud.
What are the engine operating conditions (engine speed, engine-load)?
Looking at the spark voltage, i t looks like idle. Am I correct?

Which signal trace is the ionization signal?
Thank, Gerhard

[JohnP]

Gerhard: My PM center says the PM is still in my outbox. That's weird. Usually that means your contact hasn't opened the PM yet.

The top trace is the ion signal, middle trace is the high pass filter, bottom trace is the envelope of the high pass filter, or the knock pulse.

It looks like the period is about 9.5 msec, so 15/0.0095 is about 1600 RPM.

I think stock timing at idle is 5° BTDC, I had it at least 10° beyond that, to ensure we would see knock. The factory knock sensor was still connected, and it was controlling timing.

[JohnP]

Gerhard, I sent another video link to your PM's.