- Gates Toothbelt Drive Issues -
Myths, Rumors, Claims, Counterclaims, and Nonsense
There are rumors, claims, counterclaims and just plain false information floating around regarding the use of the Gates Polychain™ Toothbelt as a main rotor drive. I think it is important to debunk the myths and rumors and put some numbers to the claims. The issues covered here are:
- Belt Stretch Under Load,
- Belts Breaking in Service,
- Belt Preload,
- "Mah Harley don't need no Preload" ,
- Change in Preload with Temperature,
- The Crowned ProDrive Sprocket, and
- The Color of Your Belt.
ProDrive's website claims that the Gates belt stretches 1/4" under the drive load on the secondary shaft. Here is why that claim is absurd.
We all seem to have agreed that, in order to transmit 95 HP (230 lb-ft of torque at 2168 RPM) to the main rotor, the driving tension load (that is, differential tension, not including preload) in the ProDrive belt is about 1620 pounds. (BTW, for the same power, the driving tension in Al Behuncik's system is 1281 pounds.)
Gates Engineering has provided EPI with the belt modulus which they use to calculate the stretch of a PolyChain belt in response to a given load. Using the Gates methodology and that modulus, we calculated the stretch produced by (a) 1619 pounds of tension applied to (b) a 62mm PolyChain GT-2 belt over a tensile span of (c) 19.75 inches. That stretch is 0.066 inches (just a hair over 1/16 of an inch).
That stretch value (0.066") agrees exactly with the value which Gates Engineering calculates for the ProDrive system with 95 HP applied at 2168 RPM.
It is a puzzlement how ProDrive comes up with a belt stretch value which is four times the REAL value. Anyone who has actually examined the system in operation system would realize that 1/4" of stretch is just not reasonable.
Perhaps it is based on some fallacious observation taken from the misleading static display they had at Homer's-2004. That display consisted of two ProDrive sprockets (34 and 144 tooth) loosely attached to the side of a trailer, with a lever for applying torque to the driving sprocket. Of course, there was no preload on the belt. So when you applied 230 or so lb-ft of torque, the loose side of the belt sagged, mostly due to the movement of the large sprocket, which was loosely-bolted to the trailer. There was clearly no residual tension in the slack side of the belt, AND LOOK ! , ProDrive proclaimed, "the belt does not skip teeth".
The misleading part of this snake-oil demo is that if you were to try and transmit 230 lb.-ft. of torque through this drive dynamically (like at 2168 RPM, for example), with no preload on the system, you would certainly have tooth-skipping and horrible overloads to the shafts and bearings (not to mention to the belt).
And, of course, there was no mechanism in their demo to show the dramatic increase in SHAFT load caused by an under-tensioned belt.
Which brings us to point 2:
Recently, there have been two cases of the ProDrive belt fracturing across a single tooth root, as shown in this picture (Figure 1) of Mike Morgan's belt, which broke after slightly more than 3 hours of operation. Ardie Greenameyer's broken belt looked very similar to this one.
EPI sent Mike Morgan's broken belt to Gates Engineering for analysis. Their opinion (a consensus of several knowledgeable Gates engineers) is that the belt almost certainly fractured because of being operated in an undertensioned drive system. They said there was clear evidence of the belt having skipped teeth, which severely overstresses the belt cords AND the shafts AND the bearings.
A copy of that report from Gates is shown in Figure 2 below. (It has been edited to remove names and phone numbers. REMEMBER, Gates doesn't want their stuff on aircraft.)
Recently ProDrive claimed they were sending Ardie Greenameyer's broken belt (the one being fixed at Homer's in July, 2004) to Gates for analysis.
We really don't understand why ProDrive would send that belt to Gates for analysis, since they insist is not a Gates product.
(BTW, Gates engineering has told us unequivocally that one ProDrive belt sample we sent them was indeed a Gates product, and they wanted to know why the Gates logo had been removed.)
Our contact in Gates engineering is one of the people responsible for the engineering manual. In July of 2004, he told us that Gates is in the process of updating their manual to reflect some of the new engineering data they have gathered from field experience with the PolyChain toothbelt. One of the important changes is the fact that the maximum recommended tension ratio is being increased from 8:1 to 10:1 (That means the minimum allowable preload is LESS than before. (If you don't understand tension ratio, check it out here.) Throughout the EPI website, the calculations for loads imposed by the PolyChain toothbelt drive have been updated to reflect this new information.
It is interesting that ProDrive publicly states that they do not adhere to the recommendations of the Gates Rubber Company, the manufacturer of the Toothbelt used in their main rotor drive system. (Of course, at other times they deny that Gates is the belt manufacturer, so go figure.)
Since they openly admit to ignoring Gates' recommendations, one can't help but wonder if ProDrive has done more engineering and research on the properties, performance, ratings, installation and life of these belts than the designer and manufacturer of the Polychain™ product line. By their claims, it appears to us that ProDrive is suggesting that they know more about the Polychain belt system than Gates does.
During the discussion with Gates engineers about the broken belt (caused by under-tensioning, they all agreed), the subject of ProDrive's installation instructions (which specify a 5/16" mid-span deflection from a 10 pound load) came up. One of the engineers was very emphatic regarding the amount of under-tension the ProDrive installation spec represents. (Be sure to read Subject 5: Preload Change with Temperature)
This is a direct copy of the information in the Gates Manual with respect to the self-tensioning which results from insufficient preload in a PolyChain drive system:
Gates PolyChain GT2 Design Manual, Page 110:
"All synchronous belt drives exhibit a self-tightening property when transmitting a load. Lab testing has shown this property is similar for all tooth profiles. The designer should be aware that self-tensioning can result in increased bearing and shaft loads, as well as reduced belt life. These effects can be AVOIDED by following proper tensioning procedures."
"Belt overtensioning can impose unnecessarily high bearing and shaft loads. However, belt undertensioning results in self-tensioning. When a belt is too loose for the design load, the belt teeth try to climb the sides of the sprocket grooves, which dramatically increases the stresses on the belt, the belt teeth and the shafts and bearings, leading to rapid tooth wear and reduced component life."
The Gates analysis program (as well as the EPI program, which we wrote using the calculations provided in the Gates Engineering Manual) show that the belt PRELOAD indicated by a 5/16 deflection at a 10-pound SIDE LOAD is sufficient for no more than 12 HP at 2168 RPM! (Be sure you understand the difference between PRELOAD and SIDE LOAD.)
In order to apply the correct PRELOAD for 95 HP, the SIDE LOAD for 5/16" deflection on a used belt (more than about 10 hours on it) should be 53 to 60 pounds! (These numbers are from a Gates Engineering analysis of the ProDrive system.) Note that to get the right PRELOAD for 120 HP to main rotor, the side load for a used belt should be 72 to 78 pounds.
The distinction between new and used belts is necessary because in the first few hours of operation of a new belt, the belt takes a small permanent stretch, which has the effect of reducing the preload. So the preload values for a new belt are higher than for a used belt, in order that the correct preload remains after the new belt stretches.
That argument has been used several times to support the ProDrive claims that somehow their drive is immune from the need for preload. Let's look at the differences between "Mah Harley" and the main rotor drive.
First, the power level being transmitted by the Harley belt drive is extremely low about 95% of the time. How low, you ask?
Well, it's fairly easy to figure out. One of my friends has a belt-drive Harley, and he has taken careful data which shows he gets 46 miles per gallon at a steady 70 MPH (freeway). That is a fuel consumption of 1.52 gallons per hour, or about 8.9 pounds per hour. If we use a conservative estimate of 0.48 for the BSFC of the old 45° V-twin, then the power the engine produces to cruise at 70 is 18.5 HP (8.9 pph / 0.48 = 18.5 ).
The outside diameter of the rear wheel of his bike is about 26 inches, so at 70 MPH, the rear wheel is turning 775 RPM. Assume that only 1/2 HP is lost between the crankshaft and the rear wheel, so that 18 HP is going on the road. 18 HP at 775 RPM requires a torque of 122 lb-ft. The belt sprocket on the rear wheel is about 10" in diameter, so the belt driving force to transmit 122 lb-ft is about 293 pounds. (That's about 5-1/2 times less driving tension than the main rotor belt requires for a 95 HP cruise.)
Consequently, the preload Harley specifies for their belt is quite a bit less than what is needed on the main rotor drive. In fact, at a 10:1 tension ratio, 293 pounds of driving tension requires only a 32-pound preload.
Second, the preload Harley specifies (recommended by Gates, since they developed the belt specially for HD) is the value based on the cruise load, since that is the predominant loading the machinery sees.
Third, the tooth and sprocket profiles Harley uses were developed by Gates for the specific purpose of being able to tolerate the occasional burst of wide-open throttle without needing to have the preload set for the high-power operation. It's not the same tooth profile as you have on your helicopter.
ProDrive insists that temperature change does not significantly alter the loading in the system. Part of the convoluted reasoning behind that claim is based on their incorrect grasp of belt stretch (explained above).
If you think we're making this up, go check the belt tension on YOUR system when the ship has sat overnight and is at ambient temperature, and then go hover the ship for a half hour and check the tension again. You will find that the belt has tightened substantially by warming up. (Since ProDrive claims that does not happen, they must be MEASURING different hardware than they are SELLING.)
Our measurements (taken at Homers, 2003) indicate that the ProDrive belt system components settle out at about 50 degrees F over ambient after a half-hour flight (in Ohio).
Our calculations (using moduli and methodology provided by Gates Engineering) show that, for a 50°F rise, the ProDrive system gains almost 400 pounds of preload, bringing the system preload CLOSER to that required for the 90-100 main rotor cruise HP (but still undertensioned).
That increase comes from 3 sources:
- 150 lbs. from shrinkage of the belt,
- 220 pounds from growth of the 144-tooth aluminum sprocket, and
- 21 pounds from growth of the 34-tooth iron sprocket.
The total rate of tension increase is about 7.8 pounds per degree F.
It is our opinion that the only reason there haven't been more (known) belt failures is because of the substantial INCREASE in preload which is occurs as a result of the thermal tensioning described here, which occurs as the drive warms up to operating temperature. In fact, if the ProDrive system was adjusted to the correct 10:1 preload at ambient temperature, it would be VERY tight at operating temperature (but not in excess of Gates specs).
We speculate that one of the reasons for the high fretting and failure rate in ProDrive-equipped secondaries is the fact that, when tensioned according to ProDrive's instructions, the belt is operating in an undertensioned condition, and the self-tensioning thereby generated puts an even greater load on the shaft than we showed here.
In fact, that points out a major problem with the entire ProDrive system: no matter how carefully you shim the shaft, YOU CAN NEVER KNOW WHAT THE REAL PRELOAD IN THE SYSTEM IS AT ANY POINT IN TIME, because (a) it changes dramatically with temperature, and (b) the structure of the helicopter deflects with the changes in load applied to it, which alters the preload.
(BTW, Al Behuncik's belt drive fixes that problem quite nicely, without the need for claims of "magic sprocket material".)
As most everyone probably knows by now, ProDrive, in order to get a patent on their belt drive, had to come up with some claimed "unique feature". After a few failed attempts, they hit the jackpot with their alleged 0.005" "crowning" of the big sprocket.
The Gates engineers we spoke with were expressed grave concern about the existence of the 0.005-inch crown on the teeth of the 144-tooth ProDrive sprocket. Gates says that any crown on the sprockets is absolutely wrong, because of the uneven fatigue loading it applies to the center fibers in the belt, potentially leading to more rapid belt failure.
An attorney who researched the Spurling patent (6,139,455), in response to a threatened patent-infringement lawsuit against users of the Behuncik drive, wrote that the original Spurling patent application was rejected twice by the US Patent Office, and was only issued AFTER the claim for the tracking crown was added. According to that attorney, the "crown" claim established the only unique, and therefore patentable, idea in the application.
Gates says that in the Polychain belt system, tracking is provided by the inner construction of the belt fibers, which are specifically designed to make the belt track on a flat sprocket face, in order to avoid the asymmetric loading produced by a crowned face. The Gates system has anti-excursion fences on the sprockets to retain the belt on the sprockets in the case of severe misalignment.
In view of the large amount of "prior-art" with respect to crowned sprockets / pulleys, and Gates statement about the tracking properties of the belt, one cannot help but wonder if the claims in the Spurling patent might be incorrect.
It is also interesting that at least one builder we know has implemented a Polychain belt drive system using components which he says he purchased off the shelf and modified to fit the application. These purchased components are NOT crowned, yet his belt system seems to track perfectly.
NOTE: Gates did mention that, insofar as their Polychain toothform is protected by a patent still in force, that ProDrive might very well be in violation of a Gates patent by producing their 144-tooth sprocket. Gates says they are looking into that issue further.
Those of you who have an older ProDrive system, in which the inside surface of the belt is blue in color instead of the current grey color, should be aware that the blue belts are an older-generation Polychain which Gates says have about 30% LESS CAPACITY than the current (GT-2) grey belts.
We think it is a curious coincidence that, even though ProDrive claims their belts ARE NOT Gates PolyChain, that the ProDrive belts even have the same color scheme as the Gates belts from the same time-periods!