TPS in function of eBoost

R

Riccio

New member
Hi guys I am so sorry for the hacker attack you suffered. I sincerely hope that the forum will recover as it is truly one of a kind.
Thanks to your advice and various tests we were able to use an impeller with a timing belt. Everything is working properly and with little overpressure we have achieved a 30% increase in torque.
The issue now is interfacing the superhcarger with the position of the accelerator pedal. Whit a TPS of 100% we have full boost and work in the supercharger map without having pumping or chocking.
In contrast, we cannot find a viable strategy with TPS between 0 and 80%.
Our goal is to have a linear and predictable torque delivery that doesn't scare the driver.
Since the throttle valve is not fully open without having a blow-off or recirculation valve what can happen sending the compressor to the rpm where I have the overpressure I require?
Would a recirculation valve solve the problem of having pumping due to the choke induced by the throttle valve?
Since we control the VESC via CAN, what strategy could we implement to maximize torque delivery under these conditions?
 
I have not yet come to it, but I was thinking of tapping into the OBDII bus and digest process additionally RPM and MAP (or load if available). The response to TPS alone is flawed since on modern EFI systems with drive by wire TPS is an "outcome" of what the ECU(!, and they are different, especially if you have driving modes) thinks is appropriate on Pedal Position (request) at a certain load.

Best is you log some timeseries of (pedalpos, TPS, RPM, MAP/NMAP, (loadrequest)) and build your tables with the right X,Y dimensions and pwm values in the cells accordingly.
A small raspberry (Odroid in my case) with OBDII adapter and PWM out will do the job. Not sure if a Arduino (even if it is a fast Teensy) cann do all at the same time ... possibly yes but I am more comfortable with a Linux environment on a Odroid ...

oh ... and after all this you could even "close the loop" with NMAP/ boost for dynamic boost and dynamic upper limits.
 
Last edited:
Actually... I wrote all that and I may have misinterpreted your question (that's how excited I am about your setup). What you're trying to do is control how the electric supercharger comes on, right? In that case, cmoalem is right on. :)

Please post pics/vids!
 
Last edited:
Semes simplest to use an Arduino (5v) to read the TPS from the sensor it's self , since the Ardunio can control the VESC too
 
Thank you guys very much for your answers. Our application is very special since we are designing a racing car. So we have a CAN protocol based management system with an aftermaket ECU. The engine is a 700cc derived from a motorcycle. It is currently all mounted on an eddy current dynamometer and managed by laboratory software. We could be able send CAN messages at will to the VESC that controls the compressor electric motor.
In the end, we too came to the conclusion of @cmoalem to maximize the output. The "dream" is to achieve torque control through the accelerator pedal. To understand it, the same feeling that the permanent magnet electric motor gives in current control.

We can acquire in real time all the parameters regarding the pressure in and out of the supercharger and the one after the throttle valve. Then we could make various strategies to manage the VESC based on those values. The problem is that we want a system that is first based on a simple rule and then later corrected with MAP values. This are not stable and result in delays in delivery.
 
This is wild I was just tossing this topic around in my head. There are plenty of teensy based ecus out there that control the entire car. (Speedyefi, dropbear, openecu) you could tap the tps line as that will have a voltage output or we could use an addon to the pedal which would allow for a more universal setup. The benefits I see is that we would have universal controller vs specific Arduino code for each make and model. I would also want this pedal to allow control over a nozzle to provide more fuel to make tuning easier as not all basemaps or ecus are open.
 
Riccio said:
Thank you guys very much for your answers. Our application is very special since we are designing a racing car. So we have a CAN protocol based management system with an aftermaket ECU. The engine is a 700cc derived from a motorcycle. It is currently all mounted on an eddy current dynamometer and managed by laboratory software. We could be able send CAN messages at will to the VESC that controls the compressor electric motor.
In the end, we too came to the conclusion of @cmoalem to maximize the output. The "dream" is to achieve torque control through the accelerator pedal. To understand it, the same feeling that the permanent magnet electric motor gives in current control.

We can acquire in real time all the parameters regarding the pressure in and out of the supercharger and the one after the throttle valve. Then we could make various strategies to manage the VESC based on those values. The problem is that we want a system that is first based on a simple rule and then later corrected with MAP values. This are not stable and result in delays in delivery.
Click to expand...
What if you tried a Current controlled setup on the VESC for the Supercharger?
The VESC would run the Motor/Impeller at whatever Duty Cycle (Voltage/RPM) is needed to hit a certain Wattage (Inducer Shaft Power)
This is an interesting thought experiment because the Affinity Laws basically state that any Pump or Fan is a Proportional Load. So all you would need to know is the Supercharger Motor Power required to either Pressurize or Move a certain amount of air with your specific setup.
 
You must log in or register to reply here.
Top