ski_dwn_it

I recently put together a 406 with some serious compression. I knew I would experience pinging, especially with the knock tables zeroed out. I know, I know all about the damage I could sustain, but my friend is running the same exact setup and his engine is run nearly every weekend at the track on race gas or a mixture of toluene or xylene to up the octane and keep it from pinging. And it does work to keep the detonation down.

Here is my deep thought though. I have been running the car around on 94 octane, the highest I can get here. After ~50% throttle I hear the pinging, and naturally leave off. I have a video in the general section of the car with 10" pure slicks at ~60% throttle and its downright scarey. So there hasn't been any real reason to go WOT up to know.

Anyways I was on a trip down to my parents house this past weekend and had the Ease Software recording the entire 25 mile trip. My part throttle tune is dead nuts on. 127-130 BLMs across the boards.

A few times on very open roads I would increase the gas unit I heard the ping. The funny thing was, and I am not real familiar with the Ease software, as I just got it, and have used Craige's in the past on my 350 engine. But the knock counts are no present when I pysically hear the pinging. Rather they count up before I actually hear it. I thought at first my knock sensor was not picking it up for one reason or another, but I think I figured it out and wanted to run it past you guys to see what your thought were on the subject.

Since the knock sensor picks up a certain frequency and retards the timing, before you ever hear the detonation under a normal timing table. I don't think it picks up the knock once it becomes audible to the human hear. Am I correct in thinking this? Or is something worng with the knock sensor? I am 99% sure I am correct, but never really thought about it before.

Thanks I look forward to hearing your responses.

V8Astro Captain

iTrader: (1)

I know that you can experience knock without hearing it. When the knock reaches a certain intensity you will be able to hear it. That's the stuff that blows holes in pistons and pops head gaskets from what I understand.

ski_dwn_it

yeah I know you can experience it before you hear it.

My situation is I can hear it, but the knock sensor doesn't detect it, once I can hear it. The frequency once it becomes audible to the human ear must be out of the range of the sensor.

Also this is probably why the algorithm for the sensor is the way it is. It detects knock and pulls a little, then listens and if it hears it again it pulls a TON. That must be set up that way to ensure that the knock remains in a frequency the sensor is able to control.

I don't know any of this for certain, just piecing bit and pieces of knowledge together to make sense of it all.

JPrevost

I follow what you're saying and it makes sence to me. I just had an aweful time trying to find out why I wasn't picking up any knock counts when I was hearing the detonation. It was really bad, so bad that I felt bad driving the car. By the time it's audible it's back out of the knock sensors "range". This probably happens when you've got a LOT of timing coming in really fast and don't allow the ecm to pull out enough timing to get rid of it. So what happens is the motor is still pinging and can't stop, the ecm just can't do it since you're probably pegged at your max SA.
My solution was to just go back and start from new changing exactly what I knew was working and eventually I found a few problems. I'm sure if I had taken notes better or had the time to be more patient I would have figured it out in a matter of minutes, not weeks. Keep notes and be pretty conservative with the timing. You might want to allow the ecm to pull out more timing to see if it's able to quench it before it gets out of hand and audible.
I'm not sure if you were just being short with your description or not but I guess I'll just explain how it works for some people unsure. The knock sensor itself is just like a microphone. It's tuned to pick up a certain range of frequency associated strongly with the bore and wall thickness of your motor. The ESC (external on my TBI and 165, internal on the 7730) is what does the real work. It filters out noise and false knock to the best of it's ability and then send a high 5 volt signal to the ecm (again only the older ecu's) to tell it what's going on. The ecu then takes into account how many knock counts it's recieved over a delta time and does it's retarding from there. The constants in the ecm have thresholds that shouldn't ever be seen but are backup to keep things from going too goofy. These thresholds sometimes get met when you start to get creative. I know I don't know everything and I know I'm missing some info so if I'm wrong in the above please speak up (Grumpy, Bob, Saturn). I guess what we're getting at is it possible that the ecm isn't seeing detonation after it gets to be audible? Like I said, I wasn't getting any knock counts when I could hear it but I never got an SES light on from a malfunction in the ESC system.

Grim Reaper

The frequency range of the knock sensor is different than the human ear. By the time WE hear it, the knock sensor goes deaf. Been saying this for the last two years.

ski_dwn_it

Never heard you say it, but that is my thoughts exactly. It was just sorta weird to actually experience it and furher confirms your thoughts.

I have had run the car with a chip that has the timing tables activive and its REALLY sluggish and I couldn't hear the pinging.

Once I drop the race fuel in again I will enable the knock sensor and find out if there are any reports from the knock sensor.

Grim Reaper

Are you running with the knock sensor disabled? WOT or always?

steve8586iroc

Could it also be that the knock sensor is tuned for a 350 and not a 400 block, which would have to be tuned to a different frequency altogether. Not saying that the sensor you are using would not pick up any knock counts at all but maybe a possibility.

Steve

Grumpy

Assuming the sensor is good you really should be able to see it at least trigger some what.

If it's not triggering at all, I'd look at the wiring sensor etc..

The ecm will try to peg the non or PW knock limit at the onset of knock, but only *listens* at the limits of the ESC attack rate.

Try setting the attack rates, WAY up, and also the ESC Decay values, and see what you get.

Depending on cal, 0 out the PE timing adder and try running the timing all on the table.

You might be hammering the engine with TOO much fuel and timing, and while you hear bad noises of about the right frequency to be knock, the acoustics of the *NOISE* are so bad from the way your forcing the knock that it doesn't excite the sensor into responding.

The knock sensor is tuned for a freguency range.
You can generate NOISES of *about* the right frequency the sensor will ignore.

JoBy

I made a simple knock detector light.

First I made a simple test.
A 1k resistor in series with a large high intensity LED.
I connected one side to the 12V feed to the ESC module and the other to the output wire.



( click the pictures to enlarge )





And a short video of me tapping at the exhaust. You can see the LED flashing when a knock is detected.
http://jobyteknik.homeip.net/corvett...r/MOV08144.MPG

I made two LED's with resistors and cable. I installed them in the dual pillar pod. I am going to use one as knock indicator and the other as shift light.

ski_dwn_it

Joboy,

Very cool! I will have to make one of them.

Grumpy

Someone remind me to turn the volume down when surfin the 3rd Gen Page.
Dunno who jumped more, me or the cat..........

Nice lil indicator..

BLOWN85/TA

iTrader: (2)

I have a TSS-1 Yamaha 5.1 speaker system, which awakened my dog as well! Then he went running for the front door barking like hell. sick-em sick-em boy!!! Cool as hell indicator, right click save as, will replicate in the future.

JPrevost

I built this circuit. I just haven't had the time to hook up the wire. I'll do it eventually. Is it the same thing?

JoBy

The 555 is a timer chip that will keep the LED on for a longer time when a knock is detected. My simple LED and resistor will only be lit for as long as the knock pulse is active. That might be a very short time.

IroczInOz

JoBy, would this circuit work on a 730 ECM? As the people who have converted to MAP from MAF disconnect the ESC connector and jumper 2 terminals together. Leaving the connector disconnected.

From what I understand the knock sensor circuitry is on the memcal and do without the ESC on 730 ecms?

JoBy

I don't know much about the 730, but you can probably get the signal from somewhere on the calpack if the knock sensor circuitry is located there.

JP84Z430HP

Would it be possible on the 730 to send the "knock" status work to an unused output, then power an LED as any other ECM output?

For example, my car will not have the AIR solenoids to deal with, so maybe use one of those outputs to turn on the light, then you could have the ECM leave the light on longer for more severe knock, or do something along those lines.....

RBob

iTrader: (1)

Try pin B8. If I am correct that pin will go low whenever the internal knock filter detects knock.

RBob.

JP84Z430HP

I take you mean B8 on the memcal?

IroczInOz

Rbob good stuff will give B8 a go and see if it does the job!

I think Rbob means B8 on the ecm pinouts. So basically just splice into that wire

JP86SS

iTrader: (1)

If you try it please post if it works or not. That would make a nice addition to the dash to see if knock is detected when not logging.
That way you'd be sure your seeing what the ECM is detecting.
Jp

IroczInOz

I will get around to doing it but probably not for a week as I am real busy. I will come back though to let you guys know if it works or not. Maybe someone else would have done it by them.

Like you say this would make a nice addition!

IroczInOz

I tried connecting a test circuit up according to the above directions for 730 ECM users. Unfortunately I did not get any flashes on the LED when I tapped the engine with a spanner. Maybe the car has to be running ??

I tried pin B8 which is the Distributor Low Reference signal originally connected on the 165 ECM (black/red) wire.

I also tried connecting it to BB9 (white wire) which originally was EST reference signal on the 165 ECM. That had no effect either on the LED.

So maybe it is not possible to wire up a LED indicator for Knock as easily as on the 165 ECM or maybe there is something I am missing?

The way I had the LED wired up was one leg to 12v and the other LEG to the pinout on the ECM.

I had key turned to ON while performing the tests. So anyone have any other ideas?


PS. Well I just had another look at the ECM pinouts for the 730, and looks like maybe the Pinout to connect to is GF9 (Dark Blue)

I'll try again tomorrow.

Cadillac

Impressive. :hail:

IroczInOz

I tried pin Gf9 quickly today and no luck. So for 730 users we have no luck at the moment.

I was giving the engine a fair whack so I don't think it was a lack of strength.

92Z

iTrader: (1)

Back from the dead.

Anybody have any other ideas on this? Is there a wire or spot in the ECM (7730) that has or drops voltage when the computer is pulling timing?

JP86SS

iTrader: (1)

Just did a quick check and it could be done by (guess what?)
Writing some code.

If "WE" used a check on:
$003a #$01 HIGH KNOCK FLAG
and/or
$003a #$02 NO KNOCK FLAG

as indicator if knock is detected, then turn on a digital output.
(I didn't go that far yet, just needs to be looked at an open one)

Qualify it using $003C #$20 KNOCK ERROR BIT
This check that the knock sensor is working

Then find the bit to turn on and pass it into the 3FFC location and let execution do the subroutine to turn the light on if "no knock" is turned off.
Not fully into the outputs to find one, but this looks possible.
jp

vernw

iTrader: (2)

When RBob say the B8 signal on an $8D/730l "will go low" when ever knock is detected --- sorry for my ignorance here --- does that mean it will go to a good ground?

Meaning it you run 12v power to one side of an indicator light, and the the other side to B8, that the light will come on when knock is detected?

RBob

iTrader: (1)

Need to use a transistor in order to drive a lamp/LED. There is a 10K resistor in line with pin B8. Use something like a 2N3904 with pin B8 into the base. Emitter to ground and collector to a lamp or a LED with a 470 ohm in series (too 12v).

Sheet 4 of the schematics, lower right corner (ESCIO#).

RBob.

92Z

iTrader: (1)

I tried this on my 730 ECM and the light is always on. I used pin B8 (which has no wire going to it from the wiring harness), and it has 5 volts when the key is on. Any other ideas, as i think a simple knock detector for the 730 guys would be great. Are there any wires in the harness that go to the distributor that we could tap into to light up a LED when there is knock? Any help appreciated.

RBob

iTrader: (1)

I need to try this out. It may be that the light on is no knock and the light will go off with knock. If so, then another transistor to invert the signal will do the trick. Not sure when I'll get a chance to try it though.

Can't pick up anything from the distributor. All of the knock stuff after the sensor is in the ECM.

RBob.

92Z

iTrader: (1)

RBob,
Thanks for the reply. Could you explain how to hook up another transistor to invert the signal? If so, I may be able to try this in the next couple of days. If not, i guess i'li be digging up my Intro To Electronics Book (circa 1991). Any help appreciated

92Z

iTrader: (1)

RBob,

I got it working!! Thanks to you, the internet and the old Electronics book. My problem was that i had the LED in series between the resistor and the collector on the transistor. Once i tied together the input voltage, collector, and anode, along with placing the cathode, emitter, and ground together the LED is off when base voltage is present and on when not. I verified this with a quick drive using datamaster and the LED lights up when it knocks. Now, if i can only figure out how to drive a little 12v buzzer also. Any ideas? Do i need another transistor, or can i tie into the existing one? Thanks for you help!!

RBob

iTrader: (1)

That's great, and an easier method then another transistor. If you use a low power piezo buzzer could probably place it in parallel with the LED.

How does the LED look when it knocks? Just some quick flickers or maybe a little longer? If need be can use a 555 to lengthen the pulse.

RBob.

vernw

iTrader: (2)

Sorry, but would you gents mind re-stating that in English? Maybe including a line drawing (not a circuit diagram) or some pix so us numb skulls can figure out what you did and try to repeat it?

92Z

iTrader: (1)

RBob,

Yeah the LED just flickers. I tried the buzzer parallel with the LED, but it wasn't very loud (it's an 85dB buzzer). Probably because its only getting about 2 volts thanks to the resistor (I was testing with a 9 volt battery). I looks like it will need the timer to keep the LED on longer. I don't know anything about timers! Also, a way to get 12 volts to the buzzer when the LED is on. Time to read up some more. I wonder if that diagram that JPrevost posted will work with the signal input goin to the pin B8? It's got the LED and buzzer. More tomorrow hopefully. Thanks.

JP86SS

iTrader: (1)

Could you put a small capacitor on the circuit to keep the LED lit longer? I don't know how to size it correctly or put another resistor there to control the "on" time.

92Z

iTrader: (1)

I haven't had anymore time to mess with this. I'm considering buying the following: http://cgi.ebay.com/ebaymotors/1986-...QQcmdZViewItem I believe its made by Caspers Electronics. It requires power, ground, and knock signal. Don't know if the Grand national knock signal is a voltage reference that goes low when knock is detected or not. Stay tuned.

JPrevost

I have a feeling it is the circuit I posted up earlier. Total cost in parts was under $10.

92Z

iTrader: (1)

JPrevost,

Yea, they are overpriced. I've been searching turbo Buick sites, and some of these have been going (used) for $20. I don't know enough about electronics to look at the schematic and determine exactly how it works. I sure don't want to spend $50 and have it not work. Spending $20 is a little easier to swallow. I'll post up here when i find a cheap one and get to try it out.

Toehead

yes you can. You can have a led charged by the knock signal discharge trhough a large resistor. the capacitor will control a second transistor, which will keep the light on as long as sufficient voltage remains in the capacitor. Have the capacitor discharge through a potentiometer and you can adjust how long the light remains on

Grumpy

While it might be a copy of Casper's, Casper's Electronics is in IL, not FL. That and visiting his store, it doesn't appear to be Casper's stuff.

The GNs use a remote ESC module, ie C3 stuff.

Try here for more info..
http://www.gnttype.org/techarea/proj...jectspage.html

92Z

iTrader: (1)

Grumpy,

I thought it may be Caspers because of this pic on Caspers site: https://www.casperselectronics.com/M...Category_Code=

Maybe the eBay ad is just using Caspers image. If i purchased one i would want to be sure that it was a genuine Caspers.

Regarding this statement found in the Theory00.txt from your link:

Normally, the output line of the filter going to the ECM is held "high," or about 10-12 volts. When the filter detects knocking-type sounds, it "grounds" the signal line going to the ECM.

do you think, that since the 7730 ECM outputs 5v on the B8 pin with no knock present and grounds it when there is knock, that the detector would still work? Any thoughts greatly appreciated.

firebirdjosh

iTrader: (1)

$50 for a buzzer, transistor, resistor and maybe a cap or two? I'm in the wrong business! Thats roughly $.60 worth of parts.

JPrevost

I am too. I thought 100% mark-up was rather high. Now we can all see why Asian's are doing so well.... it probably costs them more to ship their electronics and cars over here than it does to make them!

RBob

iTrader: (1)

Interesting thoughts on pricing. Lets look at it from another view point, first, from Casper's web site:

"This module will 'chirp' during mild detonation, 'beep' during moderate detonation, and give a long tone indicating severe detonation. Cleverly designed to produce a tone only when the knock is enough to be concerned about."

I'd think that it just may be a little more then a simple 555, could be wrong but don't think so. Now, let's look at what it really takes to bring something to market:

Design time
Prototyping and testing of said design
PCB layout
PCB manufacturing
Parts purchasing
Possible flashing/burning of uproc
Board assembly
Testing of said board
Repair and rework of faulty boards
Potting
Documentation (writing of, pictures and printing)
Advertising
Web site support
Shipping
Storage cost
Cost of Opportunity
Boxes and shipping materials
Warranty coverage
Customer Support (answering emails, phone, letters, . . .)
Value of product to customer
Profit enough to stay in business to provide more products

This is the difference between a DIY project and a commercial product.

RBob.

RednGold86Z

iTrader: (1)

What would it take to put a second sensor in the driver's side, and MAYBE A MODULE (and indicator LED or...), AND MAYBE SOME FORM OF OR GATE OR WHATEVER (aside from the LED), to run into the ECU (also) to be abso-f'n-lutely sure no (conventional) knock is applied. No idea if it works, but someone (still in the US, and motivated) should try it out, and hopefully have some real equipment like an oscilloscope to see problems.

The "MAYBE A MODULE" statement only is a "maybe" if somehow the knock signal from two sensors sent through some OR gate or whatever would interfere with each other in some way. If it all works with two sensors, great, otherwise maybe two systems would be the slick setup for this getup.

RednGold86Z

iTrader: (1)

100% markup is a little on the low side, unless you're actually in the country that makes the product. Then it's "about" 100%.

I don't know what to feel about someone who takes a public idea, puts it in a box and puts their name on it. It's not like they're competing against the DIY community, since there's no serious fighting back (and DIY is free, in all respects). The only AFAIK thing that they should fear is the original documents that can be proven to date earlier, and a patent (and making false claims - like "our" idea is to blah blah).

Honestly, I don't feel the need to build everything that I put on my car, because it's a big time investment (usually), and I'm lazy about my own projects (but kick it in gear on customers' or friends' projects - thanks to dad for that feature - our cars always have known problems but just aren't important enough for our own time most of the time).

I also pity any engineer who has to do customer support. I've done it without any hostility, but reading some of the horror stories just makes it a dreadful job.

I'm in Chiner, and could probably get hundreds of these printed up in a jiff if ya'll were in favor and would rather pay shipping than RadioShack. The company that I work for makes connectors, switches, and cables, so could prolly crank proto's out no problem (but it would take me some time to get approval). If ANYONE is against it, I won't even look into it. Actually, I'm probably too busy to get'r done in a timely manner.

Zai Jian.