Orr89RocZ

iTrader: (20)

Since there are 2 part numbers for knock sensors for the 2 different motors available in these cars, 305 and 350 with each having a different bore size, I believe using a 350 knock sensor on a 4.125" 400 bore will be ineffective. I have read each sensor is specifically tuned to the harmonics associated with a specific bore size. I have read a 350 sensor on a 305 will be less sensitive.

My question is how far off can it be? And is there anything I can do to the attack rates and such in the bin that can be used to make it more accurate?

Is there an alternative sensor for the larger bores that can work?


I'm using code $59 on my 401" sbc and I think the motor could handle more timing but the sensor is picking up signals and pulling timing so I backed off til it no longer pulls timing. Motor is happy but I think its conservative.

At cruise with just 9 to 1 compression, fairly cold spark plug but tight gap and my AFR 195 heads I cant seem to get more than 34-35 deg timing in the motor without it pulling a deg or two. My 383 with same heads and similar plug ran 35-36 with no hiccups and could have ran more.

On boost, I have noticed other guys with similar setups being able to run 2-3 deg or more timing than my setup.

I was just curious to see if any other guys using 4.125" bores in 377-427+ inch motors are seeing any problems with timing being overly conservative or not when using 350 knock sensors.

Apeiron

Supposedly the frequency of the knock is related to the bore size, so for "optimal" detection you should use something for a 4.125" block. A recent trend from manufacturers been to use non-tuned sensors with flat frequency response, and then do the frequency detection in a DSP.

Orr89RocZ

iTrader: (20)

Do you know of any factory 400's using knock sensors? I cant really think of any in the 85-92 range. Not sure if anyone makes one for a large bore block that has an output that will be read by our ecms/esc

Apeiron

They stopped making 400s before the beginning of the computer age, unfortunately.

Orr89RocZ

iTrader: (20)

Yeah I was looking around and it seemed to stop in 1980 or so... Looks like I may be out of luck and just use what I have. I dont mind if its on the conservative side for this motor

RBob

iTrader: (1)

No direct answer to what you are seeing/hearing. Although, I can give some insight.

GM uses a broad-band approach to knock detection. This is due to the knock characteristics changing over the engines life time.

GM has also cost reduced the knock filters, which is the real brains of the operation. This cost reduction means more false knock reports (from my experience).

Which is why I believe/know(?) the older 4-pole filters are better then the newer 2-pole filters.

Another interesting aspect of knock detection is automotive versus marine. The GM marine knock systems use a filter with a higher center frequency. This is 454 auto, versus 454 marine. As 5 KHz versus 7 KHz.

As an aside, one way to lower the sensor sensitivity is to mount it on a 45* street elbow. Brass is preferred, but iron also works. And don't over tighten it. A single layer of Teflon tape and 11 ft/lbs of torque is good.

Interesting that you mentioned what others are running for SA. This can really be a can of worms. The calibration latency table may not match the ignition module. Which means that the commanded versus the at-crank timing may not match.

Search terms: liguidh8, EBL, latency. This can get ugly. But when correct, works out.

There can also be octane differences, compression ratio, chamber type, and so on that affects the required SA. Not sure how well you know the others, but maybe a little methanol is in use.

And, maybe they aren't running an ESC system. This has been reported/seen here on TGO. Car goes [more] better without the spark retard of an ESC system.

Back to detected knock and the resultant retard. It may be that the detected knock is true. It may be that the chambers are right on the edge of detonation. In which the ESC system is doing what it is supposed to do. Pull some timing.

But this isn't the fastest way down the track. Back in the 70's Jenkins ran right on the edge of detonation the whole way down the track. He found (*) that this was the quickest way to go.

(*) The Chevrolet Racing Engine, Bill Jenkins, 1976, page 16, para 4.

Most likely, the engine was in detonation. And if there was an ESC system, it would have picked it up and pulled timing.

What this really comes down to is the tune. Does the owner want a safe tune. One that will withstand some bad gas from time to time. Or does the owner want to eke out another tenth. Knowing that the engine may let go, or at a minimum be torn down each off-season.

No problem with either approach.

RBob.

P.S. the "what is a good tune" thread with your response is in the Tuning Guide Book sticky, Chapter 1.

gta324

I have 4.125 bore AFR heads, running code $59 and I can have alot of timing compared to you.... I can post latest bin on code59 page just need to download it from my car first.

Orr89RocZ

iTrader: (20)

This is interesting to me because I have my sensor mounted in the bottom side of the Dart block, which is just like L98 roller blocks. Same port is available for knock sensors. However I found it very hard to get the sensor into the block. At first, I could only get maybe 2 full threads covered before it got very tight... just didnt seem right to me, and I feared it may leak oil since it wasnt threaded in very far. I tightened it abit more to get a few more threads but I fear its well over 11 ft-lbs torque spec. Would this be a problem?

If so, will it cause an error to the conservative side or be less sensitive and miss detonation signals?

I think you sent me your file a long while back when I was having problems getting this $59 up and running. I know a few guys running 8D and $59 setups on 400's running more timing at cruise than me with AFR heads. I understand lots can go into that including plugs, gas type, quench height, air fuel ratio, piston shape, cam events, etc but I thought I was alittle conservative. Just curious to see what other guys are experiencing. I guess i dont mind being alittle on the conservative side but I'd like to beable to squeeze out as much power as I can from this setup.
Being a boosted setup, its harder to determine whats optimal timing and just how much you can get away with. You may only have one shot to find out

Awesome...that was just a basic response. Could have gotten alot more detailed but kept it general. I'm sure you could add alot to it/clean it up as well.

The_Punisher454

Sounds like you may need to tap the threads a little deeper

JP86SS

iTrader: (1)

You won't leak oil from the KS port, it is in the water jacket.
That's why the teflon tape can be troublesome by adding resistance to the ground path of the sensor.
I am hoping to be running my 4.155 bore next summer so I'll be having the same issue.
As a reference, is it better to change the resistance of the KS or modify the filter?

Orr89RocZ

iTrader: (20)

Hmm i thought it was in the crank case.. oh well I just used liquid PTFE type thread sealer.

RBob

iTrader: (1)

An over-tightened knock sensor will be more sensitive. Which can cause the filter to pull timing when there isn't an issue.

A little more on how the filters work. They do have a center frequency where the sensitivity is the highest. With the sensitivity sloping off at both higher & lower frequency.

Too much noise from the sensor (a microphone) can make the filter deaf. In this case it won't report any knock.

The filter acclimates to the current magnitude and frequency of engine noise. Then if there is a sudden change in the frequency of the noise, the filter reports knock.

The magnitude doesn't need to change, just a sudden change in frequency is enough to trigger a report of knock. This is why items such as an exhaust-to-frame rattle when the engine torques over causes false knock.

It isn't so much hitting the center frequency as much as a sudden change in frequency.

RBob.

RBob

iTrader: (1)

The 454 BBC is a 4.25" bore. That would be closer then a 4" bore 350. I don't know if there are any 100K ohm knock sensors for that engine. Although there should be as there are filters available for them. Note that the early 454 TBI engines didn't run an ESC system.

I know that the 3.9K sensor for the PCMs are available.

The reason for the 3.9K sensors is diagnostics. The ECM supplies +5 volts via a 4K resistor in the ECM/PCM. The sensor voltage is read via the ADC and should be centered around 2.5 volts. If not, then there is a problem.

RBob.

Orr89RocZ

iTrader: (20)

With the combination of overbore size and possibly over tightened sensor, I could be way over sensitive? Guess there is no easy solution for this.

RBob

iTrader: (1)

I would work on the knock sensor mounting. Use a 45* street elbow to mount it. And possibly tap the hole a little deeper for better thread engagement. As JP86SS noted that hole should be in the bottom of the water jacket.

Although a Dart block may be different, or if hardblok was used.

I usually don't worry too much about the actual sensor. There is more going on in the filter. From what I've found that piece if the key.

Reading the plugs will also help in getting the tune/knock understood.

RBob.

JP86SS

iTrader: (1)

I've never looked into it and am curious now from the disscusion.
Does the KS output voltage drop or rise with indication of knock?
Or is there a constant variation going on that the knock filter "normalizes" and then reacts to the magnitude change creating the rise/fall of voltage?
That's kind of what I'm getting per your description.
How would I be able to simulate that best on a bench?

RBob

iTrader: (1)

The sensor is a microphone. It is an audio output with a DC bias of 2.5 volts. The knock center frequency is in the 5.9 KHz range.

Can use a 555 that free runs in the above mentioned range. Use a .1 uf from the output pin to a 3.9K resistor to ground. The feed to the knock filter is from the junction of the .1 uf and the 3.9K. Use a NO push button switch to either power up the 555, or place it between the 555 output and the .1uf.

RBob.

???

iTrader: (5)

latency is handled in 58/59 different than 8D. so comparing timeing numbers might not be 100% right.


I think we noticed it more going from dizzy to coil packs cause one lacked a adjustment for the coil packs, but still.

jwscab

I have a knock sensor connected to a test bench, you can simply rap it with a wrench or other piece of metal, and watch the 'knock' signal out of the memcal drop low during an event. an impulse has wideband frequency content, and should be enough to trigger the circuit.

I believe in most applications where the 454 use a knock sensor, it was the same part as the 350 application.

I hopefully have attached a document that shows the center frequencies and gain setting for esc processing circuitry, pay notice to lines 22, and 23, for both a 5.7 and 7.4 engine, they are more or less the same, the tuned frequency and gains and slight offset difference.