|
Speeding up the fan -
The conventional wisdom is that highly stressed 911 race engines (e.g. turbo race engines) can benefit from additional cylinder and cylinder head cooling. A flat fan is a great option, but out of reach for some. There are some options to speed up a 245mm fan, like using the smaller fan pulley on a later fan (84mm?), and using one of the larger crank pulleys, also off a later 911. I've heard of a couple reports of this working well.
So, would it help to speed up the fan even more? For a turbo engine running high boost, for example. A larger crank pulley could be adapted/made to do this. Is there a limit to how fast the 245mm fans can spin and deliver air efficiently? Scott |
I would imagine cooling is proportional to flow in one way or another.
I would start by considering the 3 basic fan laws: 1. Flow is proportional to speed 2. Pressure is proportional to speed^2 3. Parasitic loss is proportional to speed^3 Noise will also be influenced and will significantly increase as this is a feature of blade tip speed. As blade tip speed approaches 500m/s there may be significant heating effects and the blade tips could soften. I haven't made any calculations so I am not sure where the standard fan lives but we never ran the signal generating teeth of our Phase Shift Torquemeters at speeds greater than 350 m/s without providing cooling. Obviously you also need to avoid transonic effects as you approach Mach 1. The other issue may be fatigue life due to increased centripetal forces and the resulting stresses and as we have no idea of the 'blade' off speed it is difficult to estimate when this would become an issue. We used to manufacture fan test rigs and also 'blade off' test chambers. It may be a better idea to produce a fan with an improved volume capability at the standard speed. It would be interesting to know the relative flow/speed behaviour of a 964/994 fan to see if this could be a solution. Interesting project though :) and I would like to see some of the basic numbers. |
Can't remember where I read it, but I distinctly remember that the fans on our cars become less efficient if spun too fast.
|
Some good information previously posted on air flow rates and the various 911 model fans.
http://forums.pelicanparts.com/9427978-post3.html Good luck...I'll be subscribed and interested to track your progress. |
Thanks for the link to the other thread. There were some more threads as well. One thing that was mentioned is air speed through the fan, which I suppose would become problematic if it were to approach sonic speeds. One of the threads included someone's calculations of air speed for each of the fan/pulley combinations over the years, and the highest air speed (5 blade fan) was a little over 460 mph at 6100 RPM. If you used that pulley combination (presumably with an 11 blade fan) in a race engine turning 8000 RPM, this would be over 600 mph air speed at the fan. If these numbers are correct, there is a little more room to increase fan speed, but not a ton, before approaching sonic speed (~760 mph at sea level).
In addition to using the largest crank pulley and the smaller (smallest?) fan pulley, with a later 245 mm fan, one idea to increase the pulley drive ratio a little would be to use the maximum number of shims behind the fan pulley. This would sink the fan belt deeper into the fan pulley groove and effectively reduce the fan pulley diameter a little. Maybe even put some extra shims or a spacer back there and use a smaller fan belt if needed. Not sure if that last bit is possible, but something to try. One thing that came to mind as well is a larger crank pulley may not physically fit. I believe the fan housing requires some clearancing for a 131mm pulley, and at some point the engine mount cradle would require clearancing if you use a large enough pulley. Or try to figure out how to make a flat fan :). I don't have any illusion of engineering a bevel gear drive like the factory did, but one interesting option is to use pulleys to turn the fan belt drive direction 90 degrees and drive a flat fan that way. There is a VW type 4 flat fan conversion using a 911 fan that uses this method. I'd be concerned about the stability of the fan belt, but supposedly it works for the type 4. Pic below, and some more info. in this thread, including one guy who claimed he made one himself => Horizontal T4 Cooling Fan Shroud Info - Shoptalkforums.com http://forums.pelicanparts.com/uploa...1486098431.jpg Scott |
Another couple pictures showing the fan belt. Note the twist in the belt in the second pic due to the pulley arrangement.
http://forums.pelicanparts.com/uploa...1486098884.jpg http://forums.pelicanparts.com/uploa...1486098884.jpg |
Quote:
At the RPMs we are running our race engines at, if anything, we want to slow them down. That's exactly what Porsche did! Why? Because all you are doing is burning up horsepower spinning fans as fast as our race engines do.Why? The fans loses efficiency at provides virtually no more cooling benefit beyond a certain speed. Porsche designed the fans to be most efficient at the fan speeds they would see in a street car. That's why 3.6 liters engines have smaller crank pulleys available. Too slow the fans down, not speed them up. |
Normally aspirated air-cooled 911 engines don't need flat fans.....come on.
|
A better way to increase airflow would be to remove the internal oil cooler and compensate that with a larger external cooler, like the 964/993.
Removing the HE and using that airflow to the cylinder/heads would be good to. |
Quote:
I'm referring to turbocharged aircooled motors. I see my first post was not clear about that, so I've edited it. |
Quote:
I'm not sure I would trust a 90 degree fan belt like above on a high stressed turbo charged race machine or trust some home brewed flat fan shroud to distribute the air evenly. The last the entire reason for using a flat fan, except for the cool look! |
Why does a flat fan improve cooling as for a give size and a given speed flow will be very similar unless the restrictions in a vertical cause a significant pressure drop.
|
Quote:
|
Corvairs used belt driven flat fans. The turbos could be built for 350hp with IC. Bottom end wouldn't take it for long though unless big $$$.
http://forums.pelicanparts.com/uploa...1486138811.jpg |
Why does a flat fan improve cooling as for a give size and a given speed flow will be very similar unless the restrictions in a vertical cause a significant pressure drop.
has to do with air distribution. On a vert fan cyls 3,6 get more airflow than 1,4 so 1,4 run hot. |
Quote:
I've wondered the same. I suppose the factory must have been onto something in making the flat fan though. I'd love to see some back to back tests of upright vs. flat fan and cylinder head temps. And getting the airflow right in a homebuilt version of a flat fan would probably be hit or miss. Perhaps not the best idea. I'd guess that air distribution might be more even over the cylinders and heads with a flat fan. I may have heard some discussion about less air getting to some cylinders, but I am not sure about that. Perhaps others can comment on that? Edit - looks like boosted79 already did, above ... |
After some more looking, it seems flat fans have been discussed a number of times before. Here's one thread => http://forums.pelicanparts.com/porsche-911-technical-forum/520886-flat-fans-revisited.html
Maybe I should have contained my thoughts to how much a vertical fan can reasonably be sped up to provide additional cooling to a 911 engine that requires it (like a turbo). Anyone have experience to share using a 127 or 131 mm crank pulley + small fan pulley (80-84 mm, not clear) on a 245mm upright fan, at 8000 RPM or higher? Any measured difference in cylinder head temp or oil temp? Any issues with belt reliability? Anyone try spinning the fan even faster than that? |
Does any one know the effective diameter of the pulleys?
I think it is convention to take the inner diameter and add the extra size for the neutral axis of the belt which is at around 1/3 of its depth. This would mean the following: 65-73 Crank Pulley 98mm Fan Pulley 69mm Ratio 1.4:1 At engine speed of 6000rpm the fan would be 8400rpm 73- onwards Crank Pulley 108mm Fan Pulley 59mm Ratio 1.83 At engine speed of 6000rpm the fan would be 10800rpm If you mix them up the following occurs: 65-73 Crank Pulley and 73 onward fan pulley the ratio is 1.66 so the fan is faster than with its original fan pulley but slower than the later set up. 73 onwards crank pulley and 65-75 fan pulley the ratio is 1.58 which is slower than with the original pulley. It seems that you can't speed up the later fan without making a custom fan pulley and this may be an issue for the belt as reducing the diameter will cause more fatigue damage and could reduce belt life. If there is no room to increase the crank pulley then the solution would become complex. If you could add a bearing pedestal and run two belts it is possible that you could increase the ratio. I would imagine that a late fan will have a parasitic loss of between 3 and 5kW at full speed as it will be delivering around 5000 m^3/hr. It would be interesting to measure the pressure in the shroud and this would be a good guide to airflow. The higher the pressure the more likely that flow is uneven would be reasonable assumption. Measuring the temperature at the outlet of the engine cooler would also be interesting. |
A lot of this has been discussed here, http://forums.pelicanparts.com/porsche-911-technical-forum/942046-993-fan-carrera-2.html. Yea, I know.
But there are still some unanswered questions. How fast should, or do you need to, spin the fan? How does that speed affect the alternator bearings? So; '65-'74 1.3 to 1 @ 7K RPM = 9,100 fan RPM '75-'79 1.8 to 1 @ 7K RPM = 12,600 fan RPM Turbo 1.7 to 1 @ 7K RPM = 11,800 fan RPM Mine 1.6 to 1 @ 7K RPM = 11,000 fan RPM (130 crank, 90 fan dia.) http://forums.pelicanparts.com/uploa...1486321468.jpg Here is my set up. 130mm SC crank pulley, 90mm '71 fan. http://forums.pelicanparts.com/uploa...1486321639.jpg |
Over the years I've posted a lot about fan speed.
Basically, fan speed is a compromise. The fan and ring constitute a ducted fan and as such is susceptible to strange harmonics. The higher the RPMs the louder the fan noise. As the fan tips approach supersonic the less effective it becomes. If your engine spends a larger percent of it's time at high RPMs you want to reduce the fan speed. If you sit in traffic with the ac on you want to increase the fan speed. Back in the 2.5 GTU IMSA days we discovered that when we reduced crank pulley diameter we reduced parasitic loss (increased horse power) and cooling remained constant. We now produce that reduced diameter pulley for race/performance applications. We call it the Supertec Power Pulley. http://forums.pelicanparts.com/uploa...1486326009.JPG http://forums.pelicanparts.com/uploa...1486326009.JPG We also make an aluminum fan center to convert the late style (small diameter) pulley to the early larger size. This aluminum pulley reduces rotating mass as well the fan speed. http://forums.pelicanparts.com/uploa...1486326009.JPG |
Good info Henry. I am familiar with the "power pulley" concept, but was hesitant to use a '71 pulley and fan setup on my new 3,0 motor.
I will switch back to my '71 crank pulley once my motor is somewhat broken in. It will be interesting to see if there is any noticeable difference in temperature. BTW, my '71 crank pulley measures 116mm overall diameter. Not sure why the chart lists it as 110mm. |
The pulley diameters given on the table are the outer diameters.
As I tried to suggest using these values does not give the correct fan speeds. As regards bearings most alternator bearings are good for around 20 000 rpm on a continuous basis and it the angular acceleration that generally causes them to wear out but they should tolerate 2000 rad/s^2 without too much trouble. If it became an issue then hybrid ceramic bearings could be used and this would offer a significant improvement in speed and acceleration capacity. Blade tip speeds are generally less than 50% of Mach 1 but is quite common for fans to run at speeds of up to about 1.8 x times Mach I depending on blade stiffness. The performance of the fan unit could also be improved by reducing blade tip clearances which would also reduce noise. This thread now seems to have switched from increasing fan speed to reducing fan speed which will of course reduce the air flow and hence cooling. Clearly parasitic losses will reduce but so will the airflow over the cylinder heads and without some basic measurements its all a bit of a guess. By increasing speed by 10% losses will increase by 30% with a similar reduction if the fan is slowed by 10%. |
Thank you for the input on this.
I had heard about some people decreasing fan speed (and the factory ran the smaller 226mm fan in its normally aspirated race engines - same goal). Is it possible that this is desirable for normally aspirated engines, but turbos require additional cooling? I've heard at least one report of increasing fan speed on a small displacement, normally aspirated motor resulting in lower temps. I believe that was with the 130mm and 83mm pulley combination. What are people doing on turbo race engines? Scott |
The cooling system with the small fan was generally used without engine mounted oil coolers.
This would have reduced the pressure drop and could well have improved air flow around the engine. It would be interesting to use a pitot tube to measure the differences in pressure for various installations. I will look to see if we can create a simple method to mount one and we could make some comparisons. I have a small 906 fan and it would be interesting to see the differences. |
What about blocking off the heater tubes? Isn't air constantly being blown into the heat exchangers? I know guys running headers eliminate these. Any benefit to cooling?
|
Quote:
http://forums.pelicanparts.com/uploa...1486401812.jpg |
Sorry about the cross post, but maybe some of you saw this in the 911 Tech forum.
Interesting that the highest HP 935 used the vertical fan. Of course it had water cooled heads? http://forums.pelicanparts.com/uploa...1486403797.jpg |
The early pulleys are a bit crummy and seem to vary between 115 and 117mm. The timing marks are not too accurate and can vary by 2-3 degrees.
They are supposed to be 116 but are said to rotate by 1mm for every degree which would mean they should be more like 115mm. I would agree that the table has some errors but notionally they only provide the OD so the numbers are misleading. The bottom of the V is 91mm. I wonder if one of the issues is spark plug temperatures as this engine has twin cam, 4 valve per cylinder heads as well as water cooling. Looking at the pulley sizes it may be running quite slow. |
Quote:
The first water-cooled 911 racecars had only water cooled heads - the cylinders were still aircooled (for example the 935/78 engine shown in the pic above). So it still needed a fan to cool the cylinders. Since the heads were water cooled, Porsche must have figured a regular vertical fan was sufficient for cooling. I think they went full water cooling (including cylinders) at some point on the 956/962 cars. |
Here is some research I posted about 10 years ago.
Of course, I am not an engineer so my conclusions may not be scientific. What I have more than engineering to back up my claims are years of actual experience watching how 911 engines perform given differing perimeters. Quote:
|
All we need to do now is add some heads and oil temperatures to complete the picture.
|
Quote:
Cheers |
One reason the 935 shown had the upright fan might be packaging - look at all the intake/intercooler stuff which had to be fitted into the engine compartment. Maybe it was later 935s which got the flat fan? After Porsche had time to figure out how to put all those parts somewhere else?
|
Starkey's book on the turbo Porsches says that when Metzger went to the four valve water cooled heads, the fan had less cooling to do, so they reverted to the upright fan. The fan was smaller and driven at 1.34 times crank speed, and absorbed 4php at 8,200 rpm. The result was even cooler cylinders. But the alternator was separate off to one side on a bracket, with its own belt and pulley (like the AC on street cars?).
|
Quote:
Exactly my recollection |
|
Quote:
Didn't realize it had been done for the 911 motor too ... |
Looks Corvair-ish. I would think at high RPM it might need a belt tensioner like the Yenko Stinger had to prevent belt from flying off. But maybe the ratio is different.
|
Re. vert. fans. I would think the flow to 1 & 4 could be improved with better guide vanes attached to the shrouds to direct more air from the fan discharge directly to those two cyls. It would be interesting to have a cyl head temp sensor under each plug to read the variation. Then again I have to assume the factory tried all of that and came to the conclusion it didn't do much and just went with the flat fan.
|
Is this a bad idea? Something about trapping heat in the exchangers? Sorry, if it is off topic, but I was thinking it might channel more air where it is needed most....
http://forums.pelicanparts.com/uploa...1489806205.jpg |
| All times are GMT -8. The time now is 09:04 PM. |
Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2025, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0
Copyright 2025 Pelican Parts, LLC - Posts may be archived for display on the Pelican Parts Website