TheSamba.com :: View topic

Thu Apr 05, 2018 4:44 pm

Woo Hoo! Lots of good stuff, thanks all!

Just printed off that whole tread on the JS system for some light bedtime reading material... :D

Thu Apr 05, 2018 9:03 pm

Just for interest this is what I was thinking of trying out. I will still do it this weekend just to see what happens.

Fri Apr 06, 2018 3:58 pm

I ran the J&S for about ten years. Handy setup.

Fri Apr 06, 2018 6:33 pm

oprn wrote: Just for interest this is what I was thinking of trying out. I will still do it this weekend just to see what happens. If I see what I think I see, that would be a vibration sensor. I don't think it will detect detonation in any meaningful way. It could detect bad roller bearings and gears with the right software.
And 25mm/sec is probably way less than an ACVW produces at idle. I might be wrong - you will find out.

Fri Apr 06, 2018 6:54 pm

That's pretty funny. More like $89 for a replacement sensor FOR the kit. The kits are like $700, including wiring, computer, and sensor.

Bugsy61 wrote: 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.

Sat Apr 07, 2018 5:55 am

Boolean wrote: oprn wrote: Just for interest this is what I was thinking of trying out. I will still do it this weekend just to see what happens. If I see what I think I see, that would be a vibration sensor. I don't think it will detect detonation in any meaningful way. It could detect bad roller bearings and gears with the right software.
And 25mm/sec is probably way less than an ACVW produces at idle. I might be wrong - you will find out.
That is correct and you may be 100% right!

Any of this stuff is not cheap. This sensor I believe sells out of our shop for north of $800! The indicators pictured are likely $250 or so. This came off a new pump package and the customer didn't want the hassle of any "false alarm" call outs! "Get rid of it!" he says! Then when it costs his company $20,000 for a new pump barrel...

Oh well! Not my money!

Sat Apr 07, 2018 8:06 am

That J&S set-up is a neat bit of kit. A cursory reading of the literature on the website shows it using a Bosch knock-sensor, and retarding the timing by means of the coil-trigger wire, if I'm looking at the diagrams correctly.

It's interesting because it doesn't need an engine management package (MS or otherwise) to do it's thing.

A couple of questions for those who have run one:

Where do you mount the sensor(s)?

How is the control at discerning knock in the midst of a lot of flat-4 clatter?

Sat Apr 07, 2018 6:02 pm

Preliminary testing just in!

All readings at idle.
Boolean gets the prize! This is the reading I got on the heads, intakes, exhaust studs, and distributor base. Full saturation!

This one from the front #1 main end on the case.

These two on the generator stand, rear first then the side.

The last was on the rear transaxle mount.(solid mount)

This unit is definitely too sensitive. I do believe that it could still do what I want though. What I want to try next it to not mount it in solid metal but something that would dampen the vibrations some. How about a delrin or Teflon block? Screw this thing into it and then bolt it to the engine? Hardwood might be too much?

Sat Apr 07, 2018 6:34 pm

You need to have a sensor that is tuned to the specific freq range of knock, just listening for noise or trying to filter the noise and pass the knock will be way harder.

Sat Apr 07, 2018 7:43 pm

If I was trying to incorporate this sensor into an engine control system then yes that would be vital. All the ambient noise would need filtering out. What I am looking for here is just a visual indicator of a knock so it may not be that critical.

At this stage I am not yet looking at any specific frequency, just at amptitude. Once I get that down to below 20% or so then I will have a better idea if frequency will need filtering.

Maybe it won't work but it's only costing me a bit of time to try it!

Sun Apr 08, 2018 2:55 am

Isolating with a piece of plastic might do it. I would suggest that the sensor should be mounted as close to the crank centerline as possible to avoid shaking it. I.e not at the head. The cylinder base would be a great place, and close to detonation shocks.
Interesting experiment - preparing popcorn now.

Sun Apr 08, 2018 4:56 am

Part of my problem right now is that I am not bolting it in place, just physically holding it hard against the surface. Back off the pressure and the thing goes nuts from rattling against whatever surface I am trying to read. Maybe that I would get better readings if I took the time to make a bracket for every location.

I'm just thinking about this frequency thing this morning. If I was looking for a frequency range here is how it would look:

Lowest frequency, Just off idle and a knock on one cylinder only 1200 rpm / 2 = 600 Hz.

Highest frequency, rev limited to 5500 rpm with all 4 cylinders Knocking / 2 = 11KHz.

That is a very broad frequency range. All frequencies generated by our engines would be in this range with the possible exception on a crankshaft end play knock. On thinking about this I do not think frequency alone would be a valid way to isolate this issue as all normal engine sounds will be in this range.
what we are looking for then is a sound in this range but at a larger amplitude. Is that not what our ears pick up when we hear pinging?

Sun Apr 08, 2018 5:10 am

I just went back and watched that YouTube video referred to. It is a bit hard to see on the line graph as I don't have access to his lap top to stretch that thing out and count the time lines but it appears that the amplitude is what makes the biggest change.
Certainly if you watch the bar graph at the bottom that is the case but the bar spacing could be just sampling speed.

At any rate the sensor I am playing with is not a frequency sensor at all. It is not capable of putting out a signal based on frequency.

Sun Apr 08, 2018 5:21 am

58 Plastic Tub wrote:
A couple of questions for those who have run one:

Where do you mount the sensor(s)?

How is the control at discerning knock in the midst of a lot of flat-4 clatter?
That right there is the quest I am on!

Sun Apr 08, 2018 9:09 am

I have no idea if its even possible as I know less about this then anyone here but would there be any way to dampen or filter the signal electrically?

Sun Apr 08, 2018 12:10 pm

Top engine bellhousing might be good location, as it's at the far end from valve clatter. Go to the junk yard and fill your pockets with knock sensors. :wink:
Although the knock sensor is a great idea, they have not proven to be more sensitive than the trained monkey using his hear to listen. Maybe it will just be an engine loudness gauge.....but that may be enough?
If the engine seems loud.... back off the timing until you feel it lose some power and/or smooth out, and then go back up a few degrees. Get an MSD retard box with the dash knob.

Sun Apr 08, 2018 8:16 pm

modok wrote: Top engine bellhousing might be good location, as it's at the far end from valve clatter. Go to the junk yard and fill your pockets with knock sensors. :wink:
Although the knock sensor is a great idea, they have not proven to be more sensitive than the trained monkey using his hear to listen. Maybe it will just be an engine loudness gauge.....but that may be enough?
If the engine seems loud.... back off the timing until you feel it lose some power and/or smooth out, and then go back up a few degrees. Get an MSD retard box with the dash knob.

The problem is, the average monkey can hear about 15x better than this old man. A lifetime in mechanical rooms, and an adolescence marked by Van Halen at ear-bleed volumes have left me with a limited ability to hear or discern preigniton.

As I'm running a pretty aggressive CR, knowing for sure would be nice.

Sun Apr 08, 2018 8:17 pm

Ebel wrote: I have no idea if its even possible as I know less about this then anyone here but would there be any way to dampen or filter the signal electrically?
Yes it is totally possible but I wouldn't listen to my mother and stay in school so it is beyond my skill level!

Sun Apr 08, 2018 8:21 pm

58 Plastic Tub wrote:
The problem is, the average monkey can hear about 15x better than this old man. A lifetime in mechanical rooms, and an adolescence marked by Van Halen at ear-bleed volumes have left me with a limited ability to hear or discern preigniton.

As I'm running a pretty aggressive CR, knowing for sure would be nice.
This man gets me! Wait a minute... are we brothers.... ?

Sun Apr 08, 2018 8:34 pm

modok wrote: Go to the junk yard and fill your pockets with knock sensors. :wink:
Yes but a knock sensor on it's own is only half of the picture, without a way to read it, it will be just more junk to clutter up my life. That is why at the beginning of the thread I asked if anyone knew what communication protocol knock sensors use. Then I would know what I would need to unpack the information it is sending. So far I don't have an answer to that.
modok wrote: Maybe it will just be an engine loudness gauge.....but that may be enough?
Yes! and I hope so.
modok wrote: If the engine seems loud.... back off the timing until you feel it lose some power and/or smooth out, and then go back up a few degrees. Get an MSD retard box with the dash knob.
Hmmm... perhaps there is something there...

Or I could let the wife drive and reach over and twist the distributor from the back seat... same thing. Maybe I could here it ping better from there too...