Discussion and questions related to the course EFI Tuning Fundamentals
Hi Guys, In the latest webinar (203) we were not told how to calculate the knock frequencies for rotaries, only piston info was provided. Can you please advise how this is calculated. Sorry I was not able to attend at the time and ask this live.
What is the best knock retard step size etc. eg when the ecu sees knock approaching.
What are the best knock window settings for a rotary?
Personally I don't tend to rely on knock control for a rotary engine. I know there are tuners out there who do use it but in my opinion it's of very limited use - the reason being that you need to make the engine knock in order to validate the knock control and particularly with stock or ceramic seals, almost any level of knock can result in damage. Personally that's not something I'm too interested in pursuing. For exactly the same reason, if you make the engine knock while tuning then you risk quickly damaging the engine so care and conservative timing is really the name of the game.
Now if you do want to pursue knock control after the above discussion then unfortunately I don't have a formula to calculate the knock frequency. The best I could suggest would be using taking an audio sample that includes knock and then run it through an audio processing program like Goldwave. Using this software you can perform an FFT (Fast Fourier Transform) analysis which analysis the amplitude of the sample in the frequency spectrum. What you're looking for is where in the frequency spectrum the knock is occurring. It sounds complex but it actually isn't and in the end you have a visual representation of the noise amplitude on a time graph where frequency is on the vertical axis. Knock (or high amplitude noise) shows up as white areas. Provided your sample shows both knock and no-knock operation at the same rpm and load it becomes very easy to see which frequencies are associated with the knock. In the perfect world the knock frequency will be well removed from normal background noise.
Thanks Andre, I've swapped the engines factory knock sensors to donut style wideband ones hence why I'm looking into this. I figure the oem had them for a reason.
To me bending an apex seal is just like melting a piston would be to a conventional reciprocating engine enthusiast and is all part of learning how far something can be pushed.
My aim is just to minimise any potential damage by having the ecu pull timing before the next spark event and alert me. I can then look at the logs and alter that portion of the spark map and make appropriate adjustments.
I'll look into Goldwave, thanks for the info.
Hi Steve, I’ve dyno tuned a lot of rotaries mostly with MicroTech ECUs, some Links, some Motec. I would reiterate what Andre said, you don’t want to experiment by making the engine knock as you will soon get tired of pulling engines!
In the early days of the NZ S6 class race cars when we were all trying to get the most out of the de-turboed engines I had the na 13B knocking without it failing. I was using a simple knock-ear device but it was quite loud audibly to the ear, however I was lucky it wasn’t damaged and would not recommend it. If your engine is Turboed you most likely won’t even hear it, they will knock and instantly damage the seals (usually in the rear rotor) so fast that you won’t even know it has happened. The ammount of damage will depend on the type of seals you have, the power will drop off then most likely the first clue you will get as to what has happened is when you are unable to restart it with the low compression.
However if you do still want to experiment I think the frequency is around 3.5kHz.
Thanks for the response. Yeah 3.5Khz is what I've got it set to currently. Yes, Turbo motor JC Cosmo. I don't intend to rely on knock control as Andre mentioned. It's more so if it does happen by chance in some condition even once the ecu can alert me and we can then take action to prevent any or further damage.