No advance retard

[nightdiver]

I was on the dyno again and pushed some advance more with better fuel (104)
The spark plugs have signs of deto but nothing on the datalog
Next run i see the old j&s warning for deto up tp -8 degrees
The j&s has been setted to the lowest sensitivety all the way down
Two spark plugs melted...thank god nothing else
Remooved the j&s ,one more round very carrefully and still no retard from the Fast
Any thoughts on that
Thanks

 

[bobc455]

A few things.

First of all, I don't run much race fuel, but I've never heard of 104. I hope you aren't just adding a can of 104+ to a tank of 93 and thinking that you have 104 octane- you don't!!! not even close.

Also, I don't know much about the J&S thing. Are you using the same knock sensor & annunciator (I don't think you can "share" the knock sensor like that)? If you do have a seperate circuit, I know that the sensors have to be torqued properly to indicate knock, you might check the installation.

Is the FAST showing that you get knock, and not retarding? Or can't it see the knock at all?

If the J&S is a seperate system, why does it show 8 degrees of retard - can this device control the timing even when the engine is managed by FAST? And my understanding is that knock retard is that it is progressive, if it tries to retard and doesn't see the knock go away it will try to retard further, so even though you might only need 2 degrees of retard this device might indicate 8.

I know not of what I speak, but only add a few stupid thoughts for consideration (and for other more knowledgeable folks to learn more about exactly what was happening)

-Bob Cunningham
bobc@gnttype.org

 

[John at J&S]

Bob:

The J&S SafeGuard is an individual cylinder knock controller. It's a good inductive discharge ignition, or it can trigger an MSD, Crane, etc. It sounds like he has it connected between the FAST and his ignition unit.

I'm guessing that he has an early unit. They retard two, four, or six degrees per ping, depending on how big the ping was, up to a maximum of twenty degrees. Newer units have a 10°/20° mode switch. In the ten degree mode, it retards one, two, or three degrees per ping, ten degrees max.

Software knows that the knocking cylinder is the one that just fired. It doesn't know which particular cylinder that is, just that it's THAT one, and that it won't fire for two more revolutions. Software doesn't crank in the retard until THAT one comes up to fire again. The system does this as each cylinder goes by, building up a different retard amount for each cylinder.

The unit can only retard, but it's always trying to re-advance to the stock setting. In the twenty degree mode, it re-advances at the rate of two degrees every twenty revolutions.

An optional bargraph display shows the knock retard amount. In individual cylinder retard mode, the display is updated with each spark event.

We recently introduced a direct ignition version for multi coil applications. See it here:

http://www.jandssafeguard.com/test/safeguard.html

 

[nightdiver]

Well its race gas 101 with 104 octane booster that gives aprox.104+ octanes
The Fast is B2B with a plug & play harness

I share the knock sensor for the fast & the j&s .I dont think that using two modules will make the signal weaker

The j&s is the one that KB were selling 5-6 years ago with the small monitor for the knock and the afr
I kept it till now cause it works

You can set it to retard all cylinders or only the "hot" ones


The ignition is stock exhept the coil which is a 50000 volts unit

The fast didnt see the knock and yes the sensor is new and carefully installed

 

[bobc455]

John, Sounds like a very intriguiging (sp?) device. Thanks for the info.

I assume that nightdiver couldn't tell (on his model) how many cylinders were knocking, but presumably it was the ones with the bad plugs...

If the sensor is working but the FAST doesn't see it, I guess the first thing I'd look at is the wiring. And if that is done correctly, then ...?

Hopefully some other ideas will come in from more knowledgeable people

(I assume he's using the knock sensor for *either* the J&S or the fast, not both simultaneously, that's what it sounds like)

-Bob
bobc@gnttype.org

 

[John at J&S]

nightdiver:

Ideally, a knock detector would ignore engine noise that we can hear, while detecting knock that we can't hear.

Most knock detectors use analog circuits to develop an average of the "background engine noise". The signal is then compared to the noise, and a detection is made if the signal exceeds the noise level.

You can do a lot with analog circuits, but I couldn't get them to work well enough. I played with tuneable filters, but the engine noise is often at the same frequency.

The SafeGuard knock detector is in software, and devotes significant resources of an 'HC11 processor to the task. Wish I could tell you more.

Bob:

The bargraph display will show how many cylinders are being retarded, and how far. A ten step analog voltage is sent to the display, and is updated with each spark event. If three cylinders are being retarded by different amounts, say 2°, 4°, and 6°, you would see the first three LED's on. Actually, only one is on at a time, but the persistance of the retina shows all three at once.

The sensors are high output, low impedence, and have no problem driving two loads.

All GM sensors have an internal resistor to ground, which is matched with an equal resistor in the ESC module to 5 volts. The signal then rides above and below 2.5 volts.

The resistor value varies with different part numbers. They range from 3.9 k to 8.2 k to 100 k. Don't know if FAST uses the GM scheme to bias the signal at 2.5 volts.

 

[nightdiver]

John i know about the engine noise, thats why i had the sensitivity setted very low-my engine is old with trws wich means the pistons rocks-

The Fast at individual cyl correction is setted at 0* retard but this has nothing to do with it since its a B2B unit and not the sequential one OR NOT ???????

If its really like this ill have to check the wires
Thank God i have all the books,diagramms etc

I lost through the years the directions from J&S and now i dont remember where i had this setted ,all cylinders or individual......

 

[John at J&S]

J&S mode switches

nightdiver:

On your unit, mode switch 1 is hidden behind the front panel, and is used only for factory programming.

Switch 2 is the All/Sep switch. Up is retard all, down is retard individual.

Switches 3 and 4 are used together to select the number of cylinders. For six cylinder, make sure 3 is down and 4 is up.

You should be able to tap on the sensor to make the engine retard. The J&S ignores the sensor below 1250 RPM. There is also a TPS/MAP enable wire. If the wire is left disconnected, the unit will self arm. Otherwise, the TPS/MAP voltage must exceed 1.75 volts.

Ask Craig for the conditions to enable their knock detector. They might also include TPS or MAP trip points. Satisfy these, then tap on the sensor. The FAST system should then retard.

 

[bobc455]

Individual cylinder correction has nothing to do with knock retard (at least as far as the FAST goes). You are correct, the individual cylinder correction would only be used in a sequential unit.

I think you should be looking at the No_IC3 parameter (can someone confirm that for me)?

-Bob Cunningham
bobc@gnttype.org

 

[TurboTR]

You do still have the ESC module inline between the knock sensor and the FAST, right? If not it won't work. The FAST doesn't have an integrated sensor interface; it's a discrete input only.

IME 101 with an octane booster, or any unleaded, higher octane fuel just won't get the job done when pushing the engine. Seems like your results are already trying to tell you the same thing.

John- neat box! There was a TI app note floating around about doing knock detection in the freq domain with their DSP hardware, but I'm sure the required MIPS are way out of range for a 6811 ;-) But even then you still have to know details about the particular application.

TurboTR

 

[John at J&S]

TI app

TurboTR:

Frank Parker sent me a copy of the TI app note a few years ago. They are doing some of the same things we are doing. I think they missed the most important part, though.

You are right, of course, about the horsepower required to do FFT's. They are good if you know the precise knock frequency. An SAE paper (900488), written by GM engineers, says the knock frequency can plus or minus 400 Hz for a given engine, due to combustion temperature.

 

[TurboTR]

Too bad you can't talk about your approach! I'm sure it would be *great* tech reading! Something we are starved for on the 'net in general

TurboTR

 

[nightdiver]

Originally posted by TurboTR
You do still have the ESC module inline between the knock sensor and the FAST, right? If not it won't work. The FAST doesn't have an integrated sensor interface; it's a discrete input only.

IME 101 with an octane booster, or any unleaded, higher octane fuel just won't get the job done when pushing the engine. Seems like your results are already trying to tell you the same thing.

John- neat box! There was a TI app note floating around about doing knock detection in the freq domain with their DSP hardware, but I'm sure the required MIPS are way out of range for a 6811 ;-) But even then you still have to know details about the particular application.

TurboTR

AM I MISSING SOMETHING ??????
ESC MODULE MY ASS
Jesus , i thought that it was inside the box. I mean i ordered it with ESC and it came without it, i just didnt know that this is a seperate module
Can you imagine i didnt had the J&S safeguard ? my engine wouldnt be an engine anymore............

Now should i get the module or let the engine be with the J&S ?

Man i feel a little stupid that i missed a whole module , but i didnt know it was like this....simple things made difficult...........
Thanks for the help , i appreciate

 

[John at J&S]

ESC module

nightdiver:

Measure the knock sensor resistance. If it's 100 kohm, order sensor #10456031, and the matching ESC module (don't know the part number). This sensor is for a four inch bore, and has a 3.9 kohm resistor.

With this combination, both systems should be happy.

 

[Craig Smith]

Pin H1 on the ECU is the knock retard input. This signal should come from a module, not directly from the sensor itself. It works in the manner John described above; switch it to ground and the ECU considers the input activated. It will pull timing out at the rate specified in the ESC parameters.

If you have the J&S unit doing this, I definitely would NOT have the ECU attempting to do the same job. Two things trying to do the same thing at once is asking for trouble.

 

[John at J&S]

Craig:

Thanks for your input, but there shouldn't be a problem. If the SafeGuard hears the knock first, the ESC module may never hear it.

We heard from a Syclone guy that put one on and he said that's all he did, and he went faster. He said it got him into the 12's. 12.91 IIRC.

My theory was that the SafeGuard heard the knock first, and retarded only the knocking cylinders. The stock system never got the chance to retard all cylinders.