Open Source Electric Supercharger

[neutron256]

Originally Posted by AN

Thanks ... I was pretty serious about a T72 untill you told me about the coupling issues ...

How stable has the collet adapter on the 8mm shaft and high speeds ... does the motor.housing shake at all ?

It seems very solid, I don't notice any significant vibration at higher RPMs. There is very little clearance between the wheel and housing (as is needed) and it's not making any contact at any RPM I run it at with the 8mm shaft.

[AN]

Can you please share the link to the collet adaptor you purchased ?

Also, another point that came to mind was .. that any big compressor wheel that moves air at low rpm tends to surge easily (my newly acquire turbo knowledge :d) ... have you thought about that ? Hopefully this does not manifest itself in out application .. but if it does, we will need to limit peak compressor rpm at lower engine rpm ... even at full throttle ..

Thanks

[neutron256]

Originally Posted by AN

Can you please share the link to the collet adaptor you purchased ?

Also, another point that came to mind was .. that any big compressor wheel that moves air at low rpm tends to surge easily (my newly acquire turbo knowledge :d) ... have you thought about that ? Hopefully this does not manifest itself in out application .. but if it does, we will need to limit peak compressor rpm at lower engine rpm ... even at full throttle ..

Thanks

Here is the 8mm to 10mm prop adapter I'm currently using which also happens to be about perfect spacing with a 30mm motor shaft length.
http://www.espritmodel.com/prop-shaft-adapters.aspx

The controller I'm currently designing will use closed loop control with two potentiometers. The first will allow you to adjust the throttle position at which boost begins to activate, and the second will set the boost limit based on the stock MAP sensor in a closed loop fashion. The brushless speed controller I'm using has many configuration options including breaking which allows it to actively slow down the compressor very quickly.

We'll have to see if there are any surge issues, but I don't expect there will be. In relative terms the rotating mass of a large turbo is much higher then what we're dealing with here. The shaft and turbine/exhaust side of conventional turbos are all steel since they need to deal with continuous exposure to exhaust temperatures. Just a guess but I would say at least 3/4 of the total rotating mass of a turbo is from the exhaust side which is what is hard to slow down and can lead to surge. Anti-surge compressors are also an option. You'll find most larger turbos have anti-sure channels in compressor intake.

Lower mass combined with closed loop control, and active breaking should address any issues with surge.

[AN]

Originally Posted by neutron256

Just a guess but I would say at least 3/4 of the total rotating mass of a turbo is from the exhaust side which is what is hard to slow down and can lead to surge. Anti-surge compressors are also an option. You'll find most larger turbos have anti-sure channels in compressor intake.

Lower mass combined with closed loop control, and active breaking should address any issues with surge.

Thanks for the link ... I was actually no referring to surge due to throttle plate closure for shifting ... the other (more serious) kinda surge ... when the the compressor tries to build boost with little flow.

Per my calculations , a 2L engine @ 7500rpm and 3.7psi (1.25 PR ) would be flowing 21 lb/min of hair.

same 2L / 3500rpm/ @ 6psi (1.4 PR) will be flowing 1.5 lb/min.

Now if you look these co-ordinates up on say a T77 turbo, they lie above the surge line.

I would imagine that the surge line would be even lower for the seadoo, it being a much larger wheel .. this means that 5 or 6psi cannot be achieved at a low flow rate of 11 lb/min.

[neutron256]

Originally Posted by AN

Thanks for the link ... I was actually no referring to surge due to throttle plate closure for shifting ... the other (more serious) kinda surge ... when the the compressor tries to build boost with little flow.

Per my calculations , a 2L engine @ 7500rpm and 3.7psi (1.25 PR ) would be flowing 21 lb/min of hair.

same 2L / 3500rpm/ @ 6psi (1.4 PR) will be flowing 1.5 lb/min.

Now if you look these co-ordinates up on say a T77 turbo, they lie above the surge line.

I would imagine that the surge line would be even lower for the seadoo, it being a much larger wheel .. this means that 5 or 6psi cannot be achieved at a low flow rate of 11 lb/min.

First I'll say the pressures we're dealing with are relatively low compared to what would be seen in most turbo applications so surge isn't likely to be much of an issue. Second I'm not sure if the damage caused to a turbo by surging is really even a serious issue since you won't have that same fight between turbine and compressor.

The compressor will need to be limited to about 5psi max (assuming it's capable of more) otherwise it will exceed the range of the stock MAP sensor (in the BRZ/FRS) and cause a CEL.

Edit: If the compressor is capable of higher pressures this could be addressed through swapping out the MAP sensor and/or tuning but it makes it considerably more complex.

My controller will act much like a waste gate in a conventional turbo. You set the max boost, and if that is exceeded then then it will slow the compressor motor until boost pressure falls below the set max boost pressure. With active breaking configured on the ESC this occurs very quickly. If the rain lets up I'll go out today and do some engine off runs of the compressor to give you an idea of just how fast it reacts.

The control concept I'll be using is pretty simple. I'll tap into the stock MAP sensor and send that 0-5V signal as an input to my controller. The controller will have a potentiometer as a 0-5V input. The controller will compare the input from the MAP sensor to the potentiometer. If MAP voltage exceeds the potentiometer voltage then it will begin to slow incrementally until the MAP voltage falls below what the potentiometer is set for.

It's a fairly simple program. The activation setup I plan to use will need to be integrated too, but still the programming is fairly simple.

[neutron256]

Here is a quick video of the compressor going from to full power and then back to zero a couple of times with hard breaking. Keep in mind that the response is actually slower because I'm manually controlling the speed.

[ame]http://youtu.be/MJXVTcMFdk0[/ame]

[AN]

nice ... do you know what peak rpms were you hitting ... also, was there an outlet for the air ? I imagine the engine was not running ?

[neutron256]

Originally Posted by AN

nice ... do you know what peak rpms were you hitting ... also, was there an outlet for the air ? I imagine the engine was not running ?

Correct the engine was off, no outlet other then the fact that I haven't sealed the compressor housing so some air escapes that way.

I dont really have any way of measuring RPMs.

[AN]

Originally Posted by neutron256

I dont really have any way of measuring RPMs.

http://www.eagletreesystems.com/inde...&product_id=64


Also, some motor cooling solutions for 56mm motors on this page.
http://kershawdesigns.com/BrushlessM...aleSystems.htm

Also, this should be useful in optimizing the system and identifying bottlenecks ..
http://www.eagletreesystems.com/inde...&product_id=54

[neutron256]

Originally Posted by AN

http://www.eagletreesystems.com/inde...&product_id=64


Also, some motor cooling solutions for 56mm motors on this page.
http://kershawdesigns.com/BrushlessM...aleSystems.htm

Also, this should be useful in optimizing the system and identifying bottlenecks ..
http://www.eagletreesystems.com/inde...&product_id=54

Nice find, but honestly I don't think it's really that useful. It would be nice to know how fast the motor is turning, but it really comes down to how much boost it can produce at a given engine RPM. That's going to be determined by motor RPMs and torque in the end the only real way I see to test that is to hook it up and measure boost.

[AN]

I think knowing motor rpm could be useful in that you will know if the motor is being utilized fully or weather the batteries are no providing enough current.

Also, Why dont you upgrade your MAP sensor to a 2 bar sensor so that you can accurately measure boost and tune accordingly ?

For Miata all owners going FI get the upgraded MAP sensor.

[neutron256]

Originally Posted by AN

I think knowing motor rpm could be useful in that you will know if the motor is being utilized fully or weather the batteries are no providing enough current.

Also, Why dont you upgrade your MAP sensor to a 2 bar sensor so that you can accurately measure boost and tune accordingly ?

For Miata all owners going FI get the upgraded MAP sensor.

My personal goal for this isn't to try and max out the potential of the system. If the batteries I'm currently using end up being the limiting factor I'm okay with that, and the same with the motor. I'm actually looking at a slightly less powerful motor from SkyRC because I'm not very happy with the cluster of cables the AquaStar T20 has with it's dual configuration.

The same goes for swapping the MAP sensor. Unlike many others on this thread I'm not really interested in getting more then about 5psi of boost. The goal for me is more to see how close I can keep to that 5psi near redline, and how smooth/stable I can make the control system.

For me this is more of a learning exercise to see if I can make it work as well as keeping up some of my other skills that I don't get to practice in my professional life. It's not the destination as much as it is the journey.

[AN]

Sounds good ... thanks for working out the kinks for the rest of us... However , w/o an upgraded MAP sensor, how will you find out your boost per rpm ? will you be getting a boost gauge ?

[neutron256]

Originally Posted by AN

Sounds good ... thanks for working out the kinks for the rest of us... However , w/o an upgraded MAP sensor, how will you find out your boost per rpm ? will you be getting a boost gauge ?

OpenFlash Tablet can log boost, or at least what little boost the stock MAP sensor will read.