Ion Sense

[sepp2gl]

Hi John,
you wrote:
"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."

Why do you try to detect knock @1600rpm idle-load?
kr, Gerhard

[JohnP]

Gerhard:

Thank you for that question, but all I have for the answer is excuses.

The 1990 Chevy Suburban with 454 engine failed to pass smog. It's unlicensed and could not be driven on the road. All I could do was test in the parking lot, while the owner's son rev'ed it in gear, holding it against the brake.

It's owned by a machine shop that makes a few small parts for me. It had been sitting since 2012 with a dead fuel pump. The last time it was licensed and actually driven on the street was '09.

I volunteered to get it running again if I could use it as a test vehicle. I replaced the rusted out tank, fuel pump, starter, and battery, got it running. They put a new set of tires on it and planned to use it again as a tow vehicle for their race car trailers. They did some careful testing and pronounced it road worthy, so I drove it to a smog test station but it failed the test.

They bought a new catalytic converter but "haven't got around to installing it". The two brothers that own the shop are now in their '80's with health problems. I stopped bugging them about it two years ago. They have a welder on staff that could install the converter, but the truck sits.

I lost my focus and my ion project has also been sitting. I will pay to have the converter installed and get it street legal again, and continue my development with renewed vigor, etc.

Nothing happens until you put it in motion, right? Again, thank you for the question, Gerhard!

[sepp2gl]

Hi John,
sh.. happens. I understand your trouble.
But: what is the relation to my question?
kr, Gerhard

[JohnP]

Gerhard:

As I said above, 1600 RPM was all we could get out of it while holding it against the brake. It's not street legal so I couldn't drive it.

In the video that shows my analog scope, you can hear my assistant rev the engine to a point, then it goes no further, as it is limited by the torque converter.

I'm going to see about resurrecting the truck later today.

[JohnP]

Stopped by the machine shop this morning. I connected my Lokithor jumper box and cranked it about five seconds and it started, after sitting three years.

I didn’t replace the fuel sender when I replaced the fuel pump and tank, as funds were limited at the time. Going to pay to have it done now, then have the new cat installed.

[mpgmike]

I normally check on this Forum about 2-3X per week, as it's not that active. However, I find myself checking in every day to see if there are any updates on this particular thread. Just know (JohnP) I'm your captivated audience. I'm hanging on your every word. Thank you again for sharing so freely your research -- hours and hours of testing and developing -- sharing what I would consider proprietary information -- freely. I don't know if I will ever develop anything, even at a personal level, but I am surely enjoying the journey. I call "shotgun" for this trip!

[JohnP]

Yes, I’ve considered it proprietary, but I’m 78 and may never release a marketable product. So why not get a little recognition while I can?

For the good of Mankind (LOL).

It would take a lot of effort, but the fact it can be done is enough.

And I am still hiding a little, for the good of JohnP, as it were, just in case.

You’ve seen Plan A. But there is always a Plan B.

Now I need to get back to my patent application…

[JohnP]

Gerhard, I was reading through the whole thread then came across your first post, where you said:

"There might be technical solutions to work around it, by adding the electronics to the spark-plug rather that to the coil, and it might be a fun part for a passionated hobbyist.
Having said that, I would not declare it to be technically impossible. But it might be rocket-science for a hobbyist."

Yes, indeed, that was my approach. One ion sensor per cylinder, connected directly to the spark plug wire. As I was developing this, I came across a post on another forum by an EFI engineer, stating it was impossible to do ion sensing on a distributor. I had to refrain from correcting him. I found several patents from the early '90's by Mitsubishi showing an ion sense distributor, I'm surprised he hadn't seen them. It was the Mitsubishi patents that inspired me.

[JohnP]

The red MSD distributor caps used to be available with brass spark plug terminals. I was really bummed for a couple days when that option was deleted, as I was planning on soldering the ion signal wires to them.

Then I came up with the idea to use circlips to grip the stainless steel terminals. Made from beryllium copper, it was very easy to solder the ion signal wires to them. A much better solution, available from Inventor’s Supply (McMaster-Carr). I got the Teflon tubing there, too.

[JohnP]

I ordered more high voltage diodes. They would not sell to me unless I stated they would not be used in manned aircraft.

So for piston driven aircraft engines, that means ion sensing has to be done using a coil with isolated secondary winding, in series, at the low side.

[mpgmike]

Somewhere in my electronics inventory I have 30kv diodes. They're weird! They don't act like a diode at lower voltages, but do block the higher voltages just fine. Frankly, I don't like them.

[JohnP]

Mike: I'm using a 9s12c32 processor. The less expensive version, without the CAN Bus.

I developed the thresholding scheme for the knock processor on a Rockwell AIM 65 "trainer", in 1988. It used a 6502 processor. It's much more limited, but very similar to the 68HC11.

I switched to the 68HC11 when I designed the SafeGuard, as they had a secured version (M68SEC811E2FN) to protect my code. They were always on allocation. I had to order them six months ahead, then guess how many I would need.

The 9s12 series is very similar to the HC11. Most of the code ports directly over, but it's surface mount. It took me a long time to switch from through hole to surface mount. For me, change is hard.

The secured HC11 went out of production in '04, but I was able to buy left over parts from Asia until '07. By then, they were triple the price, so I went surface mount.

Most of the circuits in the Ion Processor box are just peak detectors, buffers, comparators, and amplifiers, etc. The 9s12 looks at either the acoustic knock pulse or the ion knock pulse (depending on a user switch), and decides whether or not to retard. Most of the code is borrowed from the SafeGuard.

[JohnP]

The ion processor board can directly retard timing of a distributor ignition, in response to either an acoustic sensor or the ion sensors.

To control timing of COP ignition, I would have to add either my Interceptor, for smart coils, or the Vampire, for dumb coils.

I could make "daughter card" versions of each, to install inside the processor box, for an all in one package. Or, the user could add the existing versions externally. Either way, the wiring is the same.

But, to save all that hassle, the ion processor board can supply, to an external ECU, the high pass output of the ion signal, or the envelope of it, for the ECU to control knock directly.

I think this would be the preferred way, as the ECU is reading load, etc, and can adjust fuel and spark in response to knock input from the ion system.