FSAE and CVT

[Jon Burford]

Hi all,
We have already embarked upon a slightly different concept car to normal but with a very similar chassis to house it in.
We have used for many years a half and half aluminium honeycomb/steel spaceframe rear car and we are confident in making it. It is in fact about to be cut out.
the team this year consists of a good number of very experienced FS people with people specialising in most areas.
Here's the catch, we are making a change from the CBR600 to a twin, a Yamaha 500c parallel twin to be exact. in it's native snow mobile it comes attached to a CVT transmission.
The CVT is what we really wanted. In my opinion a correctly tuned CVT should be much easier for an inexperienced driver to use, with much more consistent engine braking and (should) give better or smoother pull out of corners. It also has the reduced complexion of no clutch or shifter needed.
All this is at the expense of the time needed to set it all up (and the room is takes).

The new engine is here, and has run briefly on our dyno last week with the restrictor and me and an old hand at FS engine tuning (he has experience of CBR and singles) are confident that she's going to work.

The reason for my post is I'm wondering if anybody has any experience of CVT's, CVT tuning, CVT problems and any other General CVT advice.

We know how ambitious this is, but we are very eager to learn as much as we can this year, FSAE is a learning curve at the end of the day.
at the moment we are ahead of schedule and are further along than we have ever been with any other car, although I am sure we will fall on problems. Hopefully we can get through.

Any advice people can add would be great

[bob.paasch]

My suggestion would be to go to Aaen Perfomance and buy a copy of their Clutch Tuning Handbook. Every decent BajaSAE team has one.

[Jon Burford]

funnily enough we have an order just gone in for a copy. Thanks for your advice.
Congratulations by the way on a truly brilliant car!

[CB23]

You might want to surf the SCCA Formula 500 forums, where they talk about their snowmobile engined and CVT clutched race cars.

The main site is

http://formula500.org/forum.php (http://www.formula500.org)

A breakaway group formed their own forum, where much more "how to" info is posted:

http://f500.us/forum/index.php

Tuning the clutches is the main challenge, but with a bit of "intelligent trial and error" you can do so quite well. The Aaen book is a good reference, but it is oriented towards a particular style of snowmobile racing, and doesn't always apply to race cars. . .

[Jon Burford]

CB23,
Thanks for your recommendation,
I will check it all out.

[bob.paasch]

Originally posted by Jon Burford:
funnily enough we have an order just gone in for a copy. Thanks for your advice.
Congratulations by the way on a truly brilliant car!

Thank you. Our Baja team is not too shabby either. http://fsae.com/groupee_common/emoticons/icon_smile.gif


Tuning the clutches is the main challenge, but with a bit of "intelligent trial and error" you can do so quite well. The Aaen book is a good reference, but it is oriented towards a particular style of snowmobile racing, and doesn't always apply to race cars. . .

Yes, the Aaen book focuses on snowmobile drags, and doesn't much cover backshifting. As noted in the Aaen book, an absolute minimum requirement is a data acquisition system that can collect primary and secondary speeds to generate shift-out plots.

I'd recommend starting with acceleration tuning to get an idea of the effect of different weights and springs on the shape of the shift-out curve and on accel times. Finding the configuration that gives minimum accel times is relatively straightforward and repeatable. Once you have a good idea of what does what, you can start on backshifting. The optimum configuration for acceleration will most likely be poor on backshifting and vice versa. Finding a good compromise will take a lot of testing as it's a lot harder to come up with a repeatable test for backshifting performance.

In Baja we are not allowed to change parts, but we are allowed to change settings. We use different secondary preloads for acceleration/hillclimb and maneuverability/endurance.

[Jon Burford]

I am sorry, I know/knew very little about Baja. I am learning!

What methods of acquisition have you used? just speed sensors on the primary and secondary?

cheers for your help

[CB23]

What methods of acquisition have you used? just speed sensors on the primary and secondary?

Eh, can't be. You can easily get secondary clutch speed from vehicle speed, and primary speed from rpm. That does not tell you how much each clutch is engaged (open/closed).

Remember - or learn - that the primary clutch, on the motor, is rpm dependent, and the secondary, typically on a jackshaft, is torque dependent. These two clutches, working together, keep your motor at it's peak. . . let's say power/rpm, while gearing up or down to best accelerate the vehicle.

[bob.paasch]

Originally posted by Jon Burford:
I am sorry, I know/knew very little about Baja. I am learning!

What methods of acquisition have you used? just speed sensors on the primary and secondary?

cheers for your help

We use a small AIM DAQ unit to measure engine rpm and either wheel speed or secondary speed using a magnetic pickup. We have a 9:1 reduction between the secondary and the wheels so measuring the secondary speed gives better resolution without putting multiple magnets on a halfshaft.

Everyone in Baja has the same 10 (more like 9 actually) HP, so the competitive key is powertrain development. The top Baja teams put in about the same effort on CVT and transaxle development as most Formula teams do on engine development.

[Jon Burford]

Yeah I am sorry,
I don't know what I was thinking there.
I just typed it and didn't think.
I have designed a rig to setup the engine and CVT on our dyno, hoping to try and get some setup done before the car is ready.
Do you think it's a good idea?

[CB23]

I have designed a rig to setup the engine and CVT on our dyno, hoping to try and get some setup done before the car is ready.
Do you think it's a good idea?

Eh, F500's have been "rear wheel dyno-ed" a small number of times. But never twice by the same person. You learn virtually nothing from such a test. Even if you use a brake dyno with a smaller/lighter drum, and able to test accurately near the HP you have. Drivetrain efficiency sounds useful, but it's not. X hours of your time for a number in your report - a poor use of your time IMHO. You cannot, and should not try to, tune your clutches for maximum efficiency.

You use an engine dyno - a snowmobile dyno - to tune your engine and tell you what rpm to tune your primary clutch to. "Sixty-three percent efficiency" (number made up) tells you nothing. 8,325 rpm = max power tells you very much. The CVT is then tuned to keep the engine at this rpm, and you adjust it to give you the best <deep breath> upshifting/downshifting/backshifting.

[Jon Burford]

At the moment the engine is setup direclty connected to a dyno waiting to be tuned. it is running though.
when I know my max power point do you think it would be usefull to setup the engine connected to the cvt on the dyno to try and set up the clutches since this will be before the car is built. Or is it useless and waiting for the car is better?

[bob.paasch]

Originally posted by Jon Burford:
At the moment the engine is setup direclty connected to a dyno waiting to be tuned. it is running though.
when I know my max power point do you think it would be usefull to setup the engine connected to the cvt on the dyno to try and set up the clutches since this will be before the car is built. Or is it useless and waiting for the car is better?

One potential problem is the torque verses speed capacity of your dyno. For our CRF450, I know we do all of our dyno work in higher gears to keep the dyno speed up and torque down.

We made a test rig for our Baja dyno some years back to do CVT tuning on the dyno. The main goal was repeatable backshift testing. With the 4:1 reduction typical of most CVT's, even the little 10 HP Briggs exceeded the torque capacity of our dyno at the resultant low speed (3200/4= 800 rpm). We ran a 4:1 chain step-up after the CVT to get the speed back up, but the inertial of the system plus frictional losses resulted in a system that really didn't behave anything like the vehicle.

You might have better luck. If so, I'd like to hear about your system.

[Jon Burford]

Thats very interesting.
we have always dyno'd our cbr in third and fourth giving a dyno speed of around 3-4.5K
my idea was to run the secondary directly to the dyno, bypassing the reduction completely.
there is obviously a problem then with the tension of the cvt belt sinc our dyno's engine frame is rubber mounted relative to the dyno.
Im sure there would be a workaround though.
An interesting Sideline, A very clued up team mate of mine is embarking upon a CVT dissertation, so I would like to see what he comes up with.

Cheers for your help again