Official Baja SAE Forums Homepage
Forum Home Forum Home > General > Design Discussion
  New Posts New Posts RSS Feed - CVT Cooling and air flow
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

CVT Cooling and air flow

 Post Reply Post Reply
Author
Message
Rian View Drop Down
Bolt Sorter
Bolt Sorter
Avatar

Joined: May/25/2012
Location: South Africa
Status: Offline
Points: 4
Post Options Post Options   Thanks (0) Thanks(0)   Quote Rian Quote  Post ReplyReply Direct Link To This Post Topic: CVT Cooling and air flow
    Posted: May/25/2012 at 10:09am
I'm doing my final year research project on the analysis of the cooling and air flow inside the Polaris CVT cover. This is being done since our CVT cover heats up quite a lot. I haven't found any significant information on any of the forums about the cooling of the Polaris CVT. Aaen's Handbook did give me some info though.

I will be doing a CFD analysis to compare it with the experimental results I am going to obtain. I need to investigate if there is any significant air flow through the CVT cover to find out if it cools down the drive belt to a point where it improves the performance or the belt life.

I already have a few ideas on this, I just want to know if anyone has some recommendations or experience with the same type of research? And if you have experienced any overheating problems, please notify me as well, it might just support the purpose of my investigation!

Will appreciate anyone's feedback...
Rian Lauwrens
University of Pretoria TuksBaja
2009 - 2011 South Africa
2012 Oregon
Back to Top
sandres913 View Drop Down
Welding Master
Welding Master


Joined: Aug/17/2011
Status: Offline
Points: 216
Post Options Post Options   Thanks (0) Thanks(0)   Quote sandres913 Quote  Post ReplyReply Direct Link To This Post Posted: May/25/2012 at 11:42am
There are some teams that have had their belts scorched due to a variety of factors (just search for "scorched belts" on here and you'll get some hits back), some because of the cover, others because of excess belt tension. I think the cover would have the worst airflow, best of luck with your project
Shaun
Purdue University Calumet - PUC Motorsports
2010 - Washington #18
2011 - Peoria #30
2012 - Wisconsin #54
Back to Top
schooter View Drop Down
Organizer
Organizer
Avatar

Joined: Feb/22/2010
Status: Offline
Points: 217
Post Options Post Options   Thanks (0) Thanks(0)   Quote schooter Quote  Post ReplyReply Direct Link To This Post Posted: May/25/2012 at 12:12pm
There's some aftermarket cooling kits available for snowmobiles, particularly ski doos running a etec. From my understanding the cooling kits just improve airflow and maybe add a cold air duct, not sure. 

Personally I would focus more on reducing heat generation rather than improved cooling (though I do believe improved cooling deserves some attention). We've worked with Precision Auto Research http://www.precisionautoresearch.com/ who sold us belt speed and temperature sensors. The speed sensor lets us monitor belt slip and the temp sensor obviously measures temperature of the belt. Somewhere in my library I believe I have a published article discussing the causes and effects of heat generation in CVT's with rubber belts. I'll refer you to that if I find it.


Edited by schooter - May/25/2012 at 12:55pm
Chase
Data Scientist
Organizer | Broadcaster | Design Judge
Back to Top
collinskl1 View Drop Down
Baja Godfather
Baja Godfather
Avatar

Joined: Jan/21/2009
Location: Saginaw, MI
Status: Offline
Points: 1056
Post Options Post Options   Thanks (0) Thanks(0)   Quote collinskl1 Quote  Post ReplyReply Direct Link To This Post Posted: May/25/2012 at 1:39pm
I generally tried to make the cvt cover as sealed as possible.  Water, mud, and other debris can cause belts to slip and wear out faster.  Even with no venting, we never had problems with hot belts.
Kyle Collins
Lipscomb University Alumni
2x Project Manager

Nexteer Automotive
Product Engineer, Electronic Power Steering

... and the 8th simple machine: a bigger hammer.
Back to Top
T_Patn View Drop Down
Welding Master
Welding Master


Joined: Jun/12/2010
Location: Long Beach
Status: Offline
Points: 108
Post Options Post Options   Thanks (0) Thanks(0)   Quote T_Patn Quote  Post ReplyReply Direct Link To This Post Posted: May/25/2012 at 10:34pm
second on what kyle said
Tyler Patten
CSULB RACING
csulbracing.com
Back to Top
Rian View Drop Down
Bolt Sorter
Bolt Sorter
Avatar

Joined: May/25/2012
Location: South Africa
Status: Offline
Points: 4
Post Options Post Options   Thanks (0) Thanks(0)   Quote Rian Quote  Post ReplyReply Direct Link To This Post Posted: May/26/2012 at 4:48am
Thanks for everyone's input. Shaun, I searched for "scorched belts" and found some great posts! 

Chase, I am possibly going to try some aftermarket cooling kits as well, just to see if it has any effect. We have 3 years of research behind us on CVT tuning and performance and we have had some excellent improvements so I doubt that there may still be excessive belt slip that is causing the heat generation. Thanks for the link to Precision Auto Research, I was actually still wondering how I was going to measure the belt temperature. If you can find that article, I will really appreciate it since I am struggling a bit to find any research done in that area of the CVT.

Kyle, we also make sure our CVT cover is as sealed as possible. We actually had a very interesting incident in a past competition where the oil from the engine spilled into the CVT cover!! The car finished the race but with very poor performance of course. So we learnt from that the importance of sealing, but if you put a intake and outlet sufficiently far away from any mud or water sources I don't think that you will have any problems.

One of the main reasons for this investigation is that if you look on a Polaris Sportsman (where our P-90 CVT is being used) it has a very definite air intake at the back of the primary cvt (where the radial fan is) and an air outlet as well. Polaris definitely designed it that way for some important reason, so the fact that we don't use an intake at the moment has to have some effect. I guess everyone also knows that one of the major heat sources is the engine! Since it is right next to the CVT I am also going to do a test to see what effect some insulation between the cover and the engine may have.

I hope that my project will clarify most people's questions about heat buildup and cvt cooling in the long run.
Rian Lauwrens
University of Pretoria TuksBaja
2009 - 2011 South Africa
2012 Oregon
Back to Top
Purduebaja View Drop Down
Welding Master
Welding Master


Joined: Sep/19/2010
Status: Offline
Points: 137
Post Options Post Options   Thanks (1) Thanks(1)   Quote Purduebaja Quote  Post ReplyReply Direct Link To This Post Posted: May/27/2012 at 1:27am
There a lot of things I have learned about cvt efficiency and heat generation so I apologize if my email sounds choppy.

First off a little history. We have been using a CVT system since I joined the team 4 years back. Initially we used a Comet but since then we have been using the Gaged GX-8 and this year the GX-9. Also our design for a CVT cover has changed damatically throughout the years, switching between the stock polaris cover, vacuformed ABS covers with varying degrees of success and more recently thin Aluminum covers with plastic removable end plates. In addition we have tried vented and non vented designs and sometimes we have had problems and others we have not.

As for scorched belts we really haven't had issues with this even with a completely sealed cover except for at the very end of last years endurance race in Peoria. However we have noticed that the additional heat buildup from the sealed cover significantly reduced our belt life. We measure our belts in several spots with calipers and also run the drive-train at full speed with the rear wheels off the ground to determine this. As the belt wears, it no longer rides up the primary sheaves as high as a new belt with the sheaves closed completely.

To us the advantage of cooling the CVT is obvious, but if the cooling is done incorrectly it will either be ineffective or will allow foreign matter into the CVT. We located our intake on the backside around the input shaft as in the Polaris design and located the discharge near the top and away from moving parts. For the discharge we used a PVC pipe fitting and put air filter foam captured inside so that it could be easily cleaned and or replaced. Theoretically foreign matter could get in through the intake or discharge, but we used lots of small holes and the foam filter is designed for engine air intakes and prevents mud and debris from falling in. The only case where I see the venting being an issue is if we were running though huge puddles or were at a water event which we never go to.
 
The theory was that the inlets were near the fast moving primary sheave which is in close proximity to the backing plate and that it would act as a sort of impeller but with no fins. We had plans to machine a small centrifugal impeller with short fins out of delrin that would go on to the back of the CVT if there were any problems to increase the flow, however it was not needed. We placed the discharge away from moving parts so that the swirling would be less and dynamic pressure would be higher. Of course CFD would have given us a much better model but the theory was decently sound and it worked well. We have measured air flow with a thermocouple based air velocity sensor and determined we were getting a significant wattage of heat out of the housing. The cover was chosen to be thin aluminum as it was easier to work with and the high thermal conductivity of the aluminum may have increased the cooling. We were going to make the inside black to absorb more IR radiation but never got around to it.

As for sources of heat you have two main losses in a rubber v belt CVT known as torque loss and speed loss. Speed loss is usually a very small component provided the CVT is setup properly and is getting enough clamping force on the primary. In baja the RPM at which we achieve peak HP is near the governed RPM so a common error I have seen with teams is to not put enough weight on the primary so that your shift speed is technically higher than governed speed. In this case you never really get enough clamping force on the belt as it is alway in clutching mode and you will generate a ton of heat. This is the problem a lot of teams have on the sled pull event. There is a clear difference between when the CVT is grabbing and when it is slipping. We found this out when we originally got our gaged GX8 which doesn't really have enough room for brass weights at 3500 RPM unless you are very clever about how you machine them. You can also get speed loss during a steep hill climb where you are basically exceeding the torque carrying capacity of the CVT. There is still a small amount (less than 3%) of speed loss in all conditions due to belt stretch, rubber compliance, and seating and unseating of the belt. A good way to determine if whats going on with the CVT is to get the MyChron 3 660 from Aim Sports. Is designed specifically with CVTs in mind and has software that can be used to analyze shift speed, losses, etc.


Torque loss is from belt bending and unbending, wedge in and wedge out, sliding motion, and belt compression. Typically there is not much you can do about this other than messing with different belt compounds and sheave angles, however I could be wrong and I am sure someone has it figured out. Torque loss in a CVT can be anywhere between 15 and 50% of the energy you put into the CVT in the rubber V-belt type, and generates a lot of heat. 

Another main source of heat in the CVT is from the engine, not only from the cooling system blowing on it but also from the PTO on the engine. Bill from Gaged has been trying to work on a way to reduce heat coming into the CVT housing from the input shaft as it gets very hot and conducts right to the primary which acts as a heatsink into the belt and the surrounding air. If you can find a way to reduce this heat input it will help you out a lot with cooling in addition to the heat shielding around the engine.

One last thing we have looked at doing is attaching a peltier plate to the OHV cover or to the muffler and then using the power from that to run a small computer fan attached to our CVT discharge for some "free" cooling. We haven't done this yet as we need to confirm with the judges if this considered modifying the engine. However, I do think I remember seeing RIT do this on the OHV cover of one of their cars to power some electronics on the car.

Here's some scholarly articles on the efficiency of rubber v-belt CVTs



Hope this helps, and I will probably edit this post several times as I am sure there is stuff I have forgotten to add.

Thanks,

George 


Edited by Purduebaja - May/27/2012 at 2:00am
Back to Top
JeremyB View Drop Down
Organizer
Organizer
Avatar
Last one standing

Joined: Oct/13/2008
Location: Huntsville, AL
Status: Offline
Points: 135
Post Options Post Options   Thanks (0) Thanks(0)   Quote JeremyB Quote  Post ReplyReply Direct Link To This Post Posted: May/31/2012 at 11:27am
Originally posted by schooter schooter wrote:

Personally I would focus more on reducing heat generation rather than improved cooling (though I do believe improved cooling deserves some attention). We've worked with Precision Auto Research http://www.precisionautoresearch.com/ who sold us belt speed and temperature sensors. The speed sensor lets us monitor belt slip and the temp sensor obviously measures temperature of the belt. Somewhere in my library I believe I have a published article discussing the causes and effects of heat generation in CVT's with rubber belts. I'll refer you to that if I find it.
 
The sensor work well? Any caveats?
 
How do you figure belt slip without knowing the actual diameters of the primary/secondary sheaves?
Back to Top
schooter View Drop Down
Organizer
Organizer
Avatar

Joined: Feb/22/2010
Status: Offline
Points: 217
Post Options Post Options   Thanks (0) Thanks(0)   Quote schooter Quote  Post ReplyReply Direct Link To This Post Posted: May/31/2012 at 1:38pm
Great post George! Nice to know someone else has discovered and actually reads some legit CVT research. 

Rian, I was glancing through my collection of CVT articles trying to find the ones discussing heat generation. I was actually going to refer you to the same ones George did. Somewhere, I thought I read a article about CVT heat generation and cooling, just can't seem to remember where I read it. Might had been in a thesis or something. 



Originally posted by <span style=font-family: Arial, Helvetica, sans-serif; font-size: 15px; font-weight: bold; line-height: 21px; text-align: left; : rgb251, 251, 253; >JeremyB</span> JeremyB wrote:

 
The sensor work well? Any caveats?
 
How do you figure belt slip without knowing the actual diameters of the primary/secondary sheaves?

Yep, that's the sensor. I unfortunately haven't been able to personally use it much because the funding for a DAQ didn't come through until after I graduated. But I have stopped back in the shop on campus a few times to see that DAQ and all the sensors being used. The belt speed sensor seems to work well. To calculate the belt slip there is a big assumption made; no slip is occurring on the secondary (driven) clutch. That is a reasonable assumption since the primary (drive) clutch is engaging the belt and majority (not all) CVT's spend more time operating in a lower range where the belt is wrapped around the primary clutch sheaves less than the secondary. Additionally this is a better assumption when considering only up-shifting when the primary clutch is causing the belt to bend/slip outward on the sheaves. I believe a good scenario of when the assumption is best would be during a drag race. 

Without that assumption researchers either use lasers to directly measure the belt pitches or linear potentiometers to measure sheave displacement and correlate it to belt pitch. And well instrumentation like that is expensive so making the no slip on the secondary clutch assumption lets you use more affordable sensors.

With the assumption of no slip on the secondary clutch you can use two simple equations, one relating belt speed to belt pitch on the secondary clutch and another equation relating center distance, both belt pitches and belt length. So you have a system of two equations and two unkowns allowing you to calculate the true shift ratio. You can then compare the true shift ratio to the speed ratio of the two clutches to determine slip percentage.


Edited by schooter - May/31/2012 at 1:49pm
Chase
Data Scientist
Organizer | Broadcaster | Design Judge
Back to Top
Priyank View Drop Down
Welding Master
Welding Master
Avatar
I don't search

Joined: Jul/04/2011
Status: Offline
Points: 241
Post Options Post Options   Thanks (0) Thanks(0)   Quote Priyank Quote  Post ReplyReply Direct Link To This Post Posted: Jun/20/2012 at 2:53pm
AT baja this year, our belt literally burned into smoky flames, and it was a scary moment back then...
But then we realised it was mainly because the belt was ill-fitted...
We adjusted it accurately, and then it showed no problems...
Point is, ideally a belt doesn't budge, unless it isn't positioned correctly, in which case, the rubber simply burns off...!!
Back to Top
schooter View Drop Down
Organizer
Organizer
Avatar

Joined: Feb/22/2010
Status: Offline
Points: 217
Post Options Post Options   Thanks (0) Thanks(0)   Quote schooter Quote  Post ReplyReply Direct Link To This Post Posted: Jun/27/2012 at 5:01pm
Just wanted to mention something about belt slip. A lot of ATV manufactures have been using a Kymco CVT (AKA variator), its become fairly popular in recent years for ATV in the 300cc class and lower. Check the parts diagram below. At first this may seem like some cheap high production drive system and well it is. It's low cost and seems to have created a large impact on many recreational vehicle markets. However, it actually has what I believe to be a very good design. 

The helix on the secondary clutch is enclosed and built right into the sheaves so there won't be any issues with impact loading on the helix like there sometimes is with an open helix. The primary clutch has 6 or maybe as many as 8 roller weights and the moveable sheave has cooling fins. The real trick thing that should help reduce belt slip (though I have not read anything to confirm this) is the centrifugal clutch built into the secondary. This allows the belt to be always engaged even when idling. This way the belt is not being possibly burnt up when idling. But more importantly your engagement dynamics should be greatly improved. You could tune the primary clutch to have enough clamping force at the same speed the centrifugal clutch engages. This would gret rid of the whole process of dealing with the huge macro belt slip during engagement. So now the engagement dynamics would largely be defined by the centrifugal clutch which I would bet to be a lot cleaner.

Another thing to note is the aluminum CVT housing. Check the pic at the bottom. It's used as a heat sink. Adds weight but might be worth it with if you can gain the performance in a cooler operating CVT.





Edited by schooter - Jun/30/2012 at 10:54am
Chase
Data Scientist
Organizer | Broadcaster | Design Judge
Back to Top
Purduebaja View Drop Down
Welding Master
Welding Master


Joined: Sep/19/2010
Status: Offline
Points: 137
Post Options Post Options   Thanks (0) Thanks(0)   Quote Purduebaja Quote  Post ReplyReply Direct Link To This Post Posted: Jun/28/2012 at 11:15am
Hmm, time to get to the cad station to see if I can make this work on our gaged setup. I was already thinking about adding fins to our primary go suck in fresh air through the intake, but this idea of a centrifugal clutch really has me going. We have additional shaft length on the primary and we already have a centrifugal clutch for dyno use so I am thinking about putting the secondary on without a key with a clutch bolted to it and keyed on the shaft output. Only problem is there would be a lot of rotating mass on a 3/4 shaft so I would need a way to brace it.

Im guessing you could just take the little stall angle out of the primary ramps so that it engages at a much lower speed, and then just set engagement speed with the spring on the centrifugal clutch. Ill be updating here with my findings
Back to Top
schooter View Drop Down
Organizer
Organizer
Avatar

Joined: Feb/22/2010
Status: Offline
Points: 217
Post Options Post Options   Thanks (0) Thanks(0)   Quote schooter Quote  Post ReplyReply Direct Link To This Post Posted: Jun/28/2012 at 1:31pm
If you can afford it it looks like you could buy a complete Kymco system for a couple hundred bucks or so. Look up a online parts breakdown for a Arctic Cat 150 ATV and check the prices. Stuff is cheap compared to a Gaged Engineering system. 

With a centrifugal clutch added to a Gaged Engineering system I would start by simply removing the spring from the primary clutch. Check the part diagram in my last post, there is not a spring on the primary clutch. I should emphasize, with the centrifugal clutch built into the secondary the belt is always engaged when the engine is running.


Edited by schooter - Jun/28/2012 at 1:43pm
Chase
Data Scientist
Organizer | Broadcaster | Design Judge
Back to Top
rjwoods77 View Drop Down
Milling Master
Milling Master


Joined: Jul/05/2009
Status: Offline
Points: 65
Post Options Post Options   Thanks (0) Thanks(0)   Quote rjwoods77 Quote  Post ReplyReply Direct Link To This Post Posted: Jun/28/2012 at 5:08pm
The cooling fan on the lat model B&S vtwins like the one used on the UB formula car is make of plastic and is very light and about the same diameter as a cage secondary. About 1.5 inches thick and moves HUGE amounts of air. Food for thought and cheap part.
Back to Top
Rian View Drop Down
Bolt Sorter
Bolt Sorter
Avatar

Joined: May/25/2012
Location: South Africa
Status: Offline
Points: 4
Post Options Post Options   Thanks (0) Thanks(0)   Quote Rian Quote  Post ReplyReply Direct Link To This Post Posted: Jul/07/2012 at 7:32am
Originally posted by rjwoods77 rjwoods77 wrote:

The cooling fan on the lat model B&S vtwins like the one used on the UB formula car is make of plastic and is very light and about the same diameter as a cage secondary. About 1.5 inches thick and moves HUGE amounts of air. Food for thought and cheap part.

Where can I get more info on this cooling fan? 

Does anyone know where I can get a data/spec sheet on a Polaris v-belt (with partnr 3211077)? I need some info like belt material, elastic modulus, max operating temperature, etc. I don't think Polaris will provide me with this info!
Rian Lauwrens
University of Pretoria TuksBaja
2009 - 2011 South Africa
2012 Oregon
Back to Top
rjwoods77 View Drop Down
Milling Master
Milling Master


Joined: Jul/05/2009
Status: Offline
Points: 65
Post Options Post Options   Thanks (0) Thanks(0)   Quote rjwoods77 Quote  Post ReplyReply Direct Link To This Post Posted: Jul/08/2012 at 8:58pm
They come on basically any industrial v twin engine maker. They always used to be cast into the flywheel but since around 2000 maybe they all went to plastic squirrel cage blower fans. Go to a B&S dealer and look up the parts diagram on a lower hp V twin as they will have a smaller fan I believe. They look like this...

http://compare.ebay.com/like/180922209441?var=lv&ltyp=AllFixedPriceItemTypes&var=sbar
Back to Top
otto View Drop Down
Milling Master
Milling Master


Joined: Oct/06/2011
Location: OR & Colorado
Status: Offline
Points: 100
Post Options Post Options   Thanks (0) Thanks(0)   Quote otto Quote  Post ReplyReply Direct Link To This Post Posted: Jul/18/2012 at 10:35pm
Originally posted by schooter schooter wrote:

Just wanted to mention something about belt slip. A lot of ATV manufactures have been using a Kymco CVT (AKA variator), its become fairly popular in recent years for ATV in the 300cc class and lower. Check the parts diagram below. At first this may seem like some cheap high production drive system and well it is. It's low cost and seems to have created a large impact on many recreational vehicle markets. However, it actually has what I believe to be a very good design. 



for those who like to see it working


edit:
much more detail



Edited by otto - Jul/18/2012 at 10:59pm
Cornell Baja 2008-2012, suspension stuff
Oregon State 2012+, saw chain & some more baja
Back to Top
Vladmir123 View Drop Down
Welding Master
Welding Master
Avatar

Joined: Aug/07/2012
Status: Offline
Points: 115
Post Options Post Options   Thanks (0) Thanks(0)   Quote Vladmir123 Quote  Post ReplyReply Direct Link To This Post Posted: Aug/18/2012 at 5:58am
this year itself at Wisconsin, we faced some problems with our belt. We were using a stock Polaris cover (completely sealed) and after a while at the track during the endurance, our CVT was performing quite badly and wasnt giving us the kinda road feedback and torque values we expected. (though that could have been partly coz of the excess tension in the belt .. im not sure yet). but after opening the cover, the CVT was running way too hot to even touch and the belt seemed worn out. we went to the extent of drilling a few holes into the top of the CVT just outta desperation. Im not sure if scorching was the reason. but it'd be nice if some one could help me out here

Edited by Vladmir123 - Aug/18/2012 at 6:01am
Back to Top
zglazer View Drop Down
Baja Godfather
Baja Godfather


Joined: May/27/2012
Location: Metro Detroit
Status: Offline
Points: 593
Post Options Post Options   Thanks (0) Thanks(0)   Quote zglazer Quote  Post ReplyReply Direct Link To This Post Posted: Aug/19/2012 at 10:26pm
Originally posted by Vladmir123 Vladmir123 wrote:

this year itself at Wisconsin, we faced some problems with our belt. We were using a stock Polaris cover (completely sealed) and after a while at the track during the endurance, our CVT was performing quite badly and wasnt giving us the kinda road feedback and torque values we expected. (though that could have been partly coz of the excess tension in the belt .. im not sure yet). but after opening the cover, the CVT was running way too hot to even touch and the belt seemed worn out. we went to the extent of drilling a few holes into the top of the CVT just outta desperation. Im not sure if scorching was the reason. but it'd be nice if some one could help me out here
If you can't tell if the belt is damaged from a visual inspection, an obvious test would be to run your vehicle with that belt and with a new belt and compare the performance of the two. If the new belt works significantly better then that tells you something with the way your CVT is set up is causing your belts to wear out prematurely.
Zack Glazer
McGill Baja Racing Alumnus 2009-2013 http://baja.mcgilleus.ca/
2012-2013 Drivetrain Designer
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down



This page was generated in 0.047 seconds.