DIY underdrive pulley?

[serialk11r]

I thought I would just stick this post here since I was already speculating about my future car (which is now my present car ).

I got the Spyder, this example came with slick tires, coilovers, and roll bar. Amazingly, even with the high rolling resistance slick tires, I got 40mpg going home doing around 55mph at night, and that includes burning up a bunch of fuel idling to stare at the OBDII scanner. The engine idles at a sort of high 770ish rpm, but the fuel flow rate is indicated as 0.17 gph which is quite good.

The scanner indicated 45mpg, but the scanner thinks the car is going a little further distance than it really is, and uncalibrated seems to read a little high, so I'm calling it 40. I have a sort of hard time believing how high that number is (since it would indicate I got maybe like 45mpg without the idling), but I guess it's possible with an efficiency minded driver such as myself. I did engine-on pulse and glide it a bit, but the car doesn't coast well at all thanks to the big sticky tires, and I actually redlined it a few times for fun so definitely not shooting for max mpg. Even if I discount it another 10%, 36 would be amazing for slick tires.

I'm a little disappointed and also happy at the same time. It seems a healthy engine with high compression ratio and low lift cam profile plus a tiny car does quite well to start with, but beating that number is going to be difficult. I thought this was going to take some work because the 5 speed transmission is so short.

Unfortunately having a job means I will probably not get to removing the serpentine belt and measuring stuff for a while. I believe there is some room in there for the bigger water pump pulley, I haven't given up yet.

[serialk11r]

Well unfortunately my engine developed rod knock (I think I know why, it's a long story), so it's going to come out. That means I finally get to measure the pulleys and underdrive the accessories.

After I got my car a few months ago, I immediately noticed that the factory Toyota ECU likes to keep it idling above 800rpm with lights off and 900rpm with the lights turned on, which is pretty crazy. That means I really don't need to be spinning the accessories this fast.

I noticed Summit Racing has a crapload of alternator pulleys that are very cheap, I think I might as well get one if only for the bling factor, assuming I find one with the right bore. Spinning an alternator 10% slower would free a few dozen watts of bearing/fan losses, and somewhat reduce iron losses.

The water pump pulley will be interesting. The stock pulley looks to have slightly under 10mm clearance to the tensioner, and is somewhere around 100-110mm in diameter. If I use up 6mm of that clearance to reduce the speed of the pump by ~11%, reducing its power consumption by 30%, that would be pretty good.

[serialk11r]

Today I decided I would try a bit harder to get a measurement on the pulley, and spent some time with my hand stuck in the tiny gap next to the motor to try and get a measurement. I can't be sure if I managed to get a reading of the actual diameter, but I got a 112.5mm on my calipers, which is including the belt wrapping around one side, so the pulley is at least 108.5mm in diameter.

Eyeballing it, there seemed to be a bit over 1cm of clearance between the tensioner and water pump, so adding 5-6mm to the radius should not be a problem. This should reduce the speed by 9%, decreasing pump power consumption by 30%, which should save me something around 100-200W while cruising depending on speed, and around 2hp at high rpm.

Can anyone make me a 119mm OD, 109 ish ID, 21mm wide plastic ring with a 3d printer? Nylon preferred, but frankly anything is probably fine. The pulley is smooth, it doesn't need grooves. Happy to give you money for your trouble of course.

[serialk11r]

I got a measurement on a spare engine, 109mm exactly. So at the rev limiter of 8200rpm, the water pump will hit 10000rpm! Holy crap.

That means the one aftermarket pulley available must be 116mm, because 7mm off that would be exactly 6% reduction as they claim. I could probably push it to 124mm without clearance issues, but that might be tight, so I'll stick with 121mm, which is 2.5mm greater radius than the Monkeywrench Racing pulley.

At 121mm, my water pump's power consumption will be reduced approximately 27%.

I am thinking of having the ID be 108.9mm, sticking it into hot tap water and heating it to 140F would make it expand to 109mm, and then it would cool for a slightly tighter fit.

I'm debating what material to make this out of. I feel like this isn't actually that important, but the marginal cost of using something better is pretty small. PETG is really cheap and I can find someone in San Francisco to do it for me, nylon and alloy 910 would be more difficult to find, and glass filled nylon would be awesome but quite expensive. I'm also unsure what to secure the plastic ring to the pulley with. I feel like normal superglue should be fine, but the strength is not that great. Two part epoxy seems like it would make a giant mess. Hmm...

Any input?

[serialk11r]

Well, I got some parts, but my car is still not fixed yet so I can't test them just yet...


Black is PETG, white is alloy 910, which the guy I found on 3dhubs gave me as a discounted sample. The PETG is dirt cheap and more than strong enough, so I am not bothering with nylon or 910 anymore.

This will be 11% underdrive (so 89% of original rpm). I will need to make sure the belt actually fits, I may have gotten too ambitious and the tensioner might hit the pulley like this.

[serialk11r]

Got around to measuring the pieces, and realized I did something extremely stupid. I forgot to take into account the fact that the parts will shrink upon cooling. Between that and what I am guessing is the print head not laying the material completely precisely, the piece is nearly 2% too small, which I cannot compensate for by boiling it in water. I did some sanding, which didn't really help.

I did learn that Alloy 910 is extremely soft after you put it in warm water, which worried me a little. I think I'm sticking to PETG as it is stiffer and much cheaper.

I am going to try a 110.5mm print which in theory should shrink to 109.1mm, but if it is anything like my experience this time, it should end up more like 108.7mm, which would expand to 109mm after I boil it for a bit.

At this point I am getting close to 70% of the cost of the MWR aluminum pulley...T_T The quest for that last 0.3hp is a lot more difficult than I thought, but at least I am getting some experience with 3d printed stuff, which should be useful for future car endeavors.

Anyone have a Celica GT-S or Matrix/Corolla XRS they want one of these for? Help me recoup my costs lol.

[serialk11r]

Okay this has been dead for a while since my car was out of commission but the 3d printing guy made me another one that's slightly bigger which I'm picking up tomorrow.

In theory I need around 15mm more serpentine belt, which turns into 7.5mm less tensioner travel. That shouldn't be a problem right? Or should I go up a half inch in belt size? I think my AC pulley is slightly smaller than stock because it's the one off the MR2 not the Celica GT-S so it might make up for it.