TheSamba.com :: View topic

Tue Apr 03, 2018 5:03 pm

Does anyone know how knock sensors work? Are they simple on/off switches or do they give a proportional signal based on acceleration? If they do give a proportional signal is it in volts, amps or in a digital format?

Wed Apr 04, 2018 8:48 am

I can only speak for ford knock sensors, but they put out a frequency when the proper knock frequency is detected.
Years ago, I had an instructor show us a knock sensor reading by tapping an 11/16" wrench on an intake manifold. The sensor wouldn't recognize the tone of any other wrench. We were watching the knock sensor reading and ignition timing. When he'd tap it the sensor reading would go up and the timing would go down. That was on a gm, but I haven't touched one of them in a long time.

Wed Apr 04, 2018 10:43 am

It is often said that a VW aircooled engine rattles so much it confuses knock sensors.

There is also a predetonation detection technique that uses the spark plug to measure the ionization in the combustion chamber caused by flame that should not be there. This uses a coilpack where the ground side of the HT coil secondary is connected to a sense wire.
It applies DC to the plug and measures the small current, while having protectiin against the spark voltage.

Wed Apr 04, 2018 11:31 am

It is an on/off sensor, not proportional. Sends out a square wave digital signal. Aircooled engines are far too noisy for them to work. I used to test them with a small ballpein hammer on the side of the block.

brad

Wed Apr 04, 2018 8:58 pm

Thanks for all the replies!

It sounds like perhaps there are more than one type. The reason for the interest is I am about to put a fresh engine in my Manx clone and with carb noise, exhaust noise, wind noise and being a wee bit hard of hearing I am not sure I will be able to hear a ping. The point of course is to maximize timing without running it on the ragged edge of self destruction. Just a simple dash mounted read out would work for me until I establish the best timing values.

I have in hand an eccelerometer type vibration transmitter used on industrial electric motor powered pumps and natural gas compressors. It has a span of 0 to 1 inch/sec and puts out a milliamp signal. I also have a couple of local indiators that are compatable. My concern was as mentioned - that general engine noise and road irregularities will mask what I want to see.

I was just curious as to what comunication protocall the automotive knock sensors used and if they would support a local read out. Also if they had any sensitivity adjustment to dampen out ambient noise, the one I have doesn't to my knowlege.

Perhaps this is getting a little deaper than most hobbyists go...

Wed Apr 04, 2018 9:13 pm

Most of them "listen" for the "ring"of knock and transmit that to the controller. As said earlier, knock is a very narrow freq range. This one, http://www.viatrack.ca/ can be setup with as a stand-alone sensor/indicator and can be incorporated in MS if you go that way and is somewhat adjustable to help reduce the effects of noise. But the noise has to be in that narrow freq range to be a bother.

Wed Apr 04, 2018 9:23 pm

I can hear my engine ping plain as day when on boost even with the wind noise.
You don't have a lot of adjustment for timing with a stock engine so just run it the standard 30 degrees at 3000 rpm or whatever the specs are for the particular distributor you plan to run.
It's not like you are going to gain 5 hp from 1-2 degrees timing adjustment.

You would be better off a little on the conservative side than adjusting too close to the detonation point. It will give a little safety for the hotter air of summer or when the engine gets warmer pulling hills.

Wed Apr 04, 2018 9:26 pm

Thanks for that! I will look it over.

Wed Apr 04, 2018 9:31 pm

clonebug wrote: I can hear my engine ping plain as day when on boost even with the wind noise.
You don't have a lot of adjustment for timing with a stock engine so just run it the standard 30 degrees at 3000 rpm or whatever the specs are for the particular distributor you plan to run.
It's not like you are going to gain 5 hp from 1-2 degrees timing adjustment.

You would be better off a little on the conservative side than adjusting too close to the detonation point. It will give a little safety for the hotter air of summer or when the engine gets warmer pulling hills.
I am not so concerned about HP at this point. I want to see what I can do to maximize MPG.

If I get the itch for more giddy up I will install a hair dryer! You can take that to the bank!:D

Wed Apr 04, 2018 11:05 pm

There are a couple of types. The ones slalombuggy was talking about...are specifically the piezo quartz type....early. Most work on a similar function but they are tuned differently.

They make resonance and non-resonance types. They use metal combined with piezo quartz crystal. If you know anything about piezo.....when its subjected to a vibration in the frequency its tuned....and its tuned by size, shape, mounting and metal material its paired with.....it generate an electrical signal in a specific frequency range.

After the 80s .....wherein a simple single sensor would send a signal that the "middle years" injection and ignition systems would reapond to by simply pulling spark advance back...and later changing fuel curve as well.....later systems used 2 or more sensors mounted in different locations and tuned to different frequencies......along with software.....to do more delicate tuning along the same lines.

As already mentioned.....thouroughly modern cars may do away with the knock sensor....and use ion sensing through the spark plug. They do the same thing.....but do jot even look or listen for knock.

They sense the condition that causes knock...which is minute second to second changes in heat and resistance to arc propagation because of pressure.
Its why modern cars that use this system....which include virtually every car VW/Audi has made for at least 10 years.....are very particular....about spatk plug gap, spark plug resistance and spark plug ground.....and is why they use iridium slow wearing plugs and NO anti-seize...ever.

A few people over the years have successfully used BASIC knock sensors along with programmable injection on air cooled engines. A few discussions years back on the STF about this.

The tuned around the vibration problem.....by damping the excessive vibration that reaches the sensor....by mounting it on an oil filled and baffled box tuat mounted to the engine.

The sensors as far as even the early 90s had their own issues. Because the piezo quartz inside was brittle and vinrayion sensitive....they had an EXACT torque spec and angle when mounting them to the engine block. Over torque of even a few foot lbs.....would set up internal cracks which made the sensor junk. I have actually,done that on my own car twice.
A good waste of a $60 sensor in 10 seconds. Ray

Thu Apr 05, 2018 12:26 am

Dang, Ray, last rainy season I had a knock sensor fault on my 2.5 MK5 and cleared it by removing and reinstalling the same knock sensor with a specified torque setting. I’m glad I followed that torque spec!

Robbie

Thu Apr 05, 2018 4:38 am

Thanks Ray, I am aware that placement is important, at least it is on these units we use in industry. They need to be in a 90* plane to the shaft you are monitoring.

I think I will wire this unit up and give it a try on the wife's Subaru and see what I learn.

We use ionization for flame detection in burners but it sounds like a total nightmare to try and use for the hobbyist! It's bad enough in industry where there is no timing of the signal to factor in! We are constantly cleaning and gapping flame rods and trying to get better grounds!

Thu Apr 05, 2018 8:03 am

asiab3 wrote: Dang, Ray, last rainy season I had a knock sensor fault on my 2.5 MK5 and cleared it by removing and reinstalling the same knock sensor with a specified torque setting. I’m glad I followed that torque spec!

Robbie

This experience....right there! Its hard to say exactly what the REAL problem was with your knock sensor.
It could have been some moisture in the threads from factory assembly that over time produced some corrosion changing the ground....if that sensor even uses ground condition....which it may not.

It could have been heat cycling and bolt stretch over time....causing a very small amount of loosening of the sensor. As noted....how tight or loose the sensor is can change how it resonates and what data it sends. I am not even sure what type of sensor the 2.5L uses. I have a 2012 Golf with 2.5L.

Just loosening and tightening can make a difference.

Yep...They are very sensitive little sensors. The one on my Mk. 2-ish rabbit (1990 cabrio)....drove me insane getting it right slightly too tight or loose and the car ran like crap.
Ray

Thu Apr 05, 2018 8:07 am

You might find this helpful, especially for visualizing what the knock looks like on an aftermarket ECU.

https://www.youtube.com/watch?v=jUUPcNPGm3k

Thu Apr 05, 2018 8:09 am

oprn wrote: Thanks Ray, I am aware that placement is important, at least it is on these units we use in industry. They need to be in a 90* plane to the shaft you are monitoring.

I think I will wire this unit up and give it a try on the wife's Subaru and see what I learn.

We use ionization for flame detection in burners but it sounds like a total nightmare to try and use for the hobbyist! It's bad enough in industry where there is no timing of the signal to factor in! We are constantly cleaning and gapping flame rods and trying to get better grounds!

Theoretically yes.....tbey are supposed to be mounted 90° to the crank plane.....but in reality...I have found that many are not.

Some sensors may not be at a 90° angle inside of the housing they are made. Others.....are turned or moved to tune for the best signal coming from the knock...which gets filtered through all of the various angles and curves of the engine.....which can alter the signal....and may work better when not just aligned in any particular plane.

Ray

Thu Apr 05, 2018 9:59 am

J & S Safegaurd has been selling knock sensor kits for about 25 year. Do a Google search for them. They were popular with the Corvair crowd, so they may work on our noisy motors. The kit is only $75.

Thu Apr 05, 2018 12:29 pm

Where is the J&S Safeguard available for $75. I thought it was much more than that.

Yes, search for a post by Jake Raby who installed a J&S and upped the timing past the knocking point and then let the J&S pull timing out when running down the track. It enabled him to get full ignition advance for what his engine needed under those conditions and he knocked a bunch of time off of his ET

Thu Apr 05, 2018 12:53 pm

j-dub wrote: You might find this helpful, especially for visualizing what the knock looks like on an aftermarket ECU.

https://www.youtube.com/watch?v=jUUPcNPGm3k

That was a nice video! Thanks!

Its also hard to know that whether or not it looks similar to the software of a factory ECU.

The question and answers underneath were great with that video as well.

Ray

Thu Apr 05, 2018 12:58 pm

Sorry, $89 for a single knock sensor retard unit. Looks like you might have to also buy a knock sensor and cable for another $100. I would go get a JY sensor at a pick and pull. The original poster is hard of hearing and can't hear an engine ping. Looks like it has an LED when activated. I suspect this basic unit is all he will need to get his timing dialed in, then use it as a safegaurd unit. It is on their website.

Thu Apr 05, 2018 1:13 pm

Jake Raby info

http://www.shoptalkforums.com/viewtopic.php?t=3375

The prices I see for the controller are a lot higher
http://www.jandssafeguard.com/orders.html