High-Speed Gear Reducer

I would be happy if I could get about a 5:1 reduction, although a steeper reduction would be more managable.

This is very interesting. 10Ft-lbs of torque is I think about 150 N.m. If so; at 60,000 RPM we are talking of nearly a megawatt of power. No easy task to reduce the speed to 3000 RPM. I can't help wondering where you ran into such a problem.

Power[W]= T*ω = T[Nm]*n[rpm]*π/30= 13.9*60000* π/30 = 8.7 E4 [W] = 87 [kW].

Or am I wrong?

I used this equation HP = T(Ft-Lb)*RPM/5252

745.6999 watts/HP

70990.63 watts ~71kW

Is the torque 13.9 N.m ?

By my calculation, I come up with about 71kW. This is not as high power of an application as it seems. I am trying to use the turning power of a small turbine to drive some mechanics. The torque load is fairly low, but finding gear reducers with the proper bearings is difficult.

In the industrial gear design you almost do not find the combination of high rpm and low torque as you need. I am very afraid that if you want to do something correct you will be obliged to have a specific design and such one will cost a LOT!

If it is small turboprop then you are also confronted with weight constrains and this makes the problem even more difficult.

Bearings are available for such speeds and even slim line so that it is possible to make it but at which cost!

Is it not possible to connect a (high speed ) fan to the input shaft. Reforming the mechanical rotation into an axial air flow en spreading this air flow over several slow turning turbines ? (am I telling someting stupid ?)

I am trying to remain compact. I have seen similar applications in RC aircraft, although those use are approximately 1% of the power output.

Were it not for the low power input, I'd be tempted to ask if you are you trying to design a gearbox for jet turbine powered helicopter?

Can you tell us more about the application?

L.J.

Why not find a bone yard for helicopters and use one of there gearboxes for a guideline if nothing else.

See my comments above. That is a very keen observation. This is not as high powered as a helicopter operation, but yes, I am trying to create what would be essentially a small turbo prop power train.

Please forgive my fading memory. I saw a gear reducer that had the input shaft turning an eccentric inside a bearing. The bearing was inside an external gear. The external gear was inside an internal gear, and as the ecentric orbited the external gear stepped one tooth at a time inside the internal gear. The reduction was a lot! I don't know how the thing didn't wobble or vibrate while in use, I only spun it on a table to look at it.

You are right that this kind of planetary gear offers a very high reduction but there are difficult problems related at high speed with the eccentric masses as well from the input shaft as from the internal gears themselves.

I shall send tomorrow the schematics so that you will see where the problems are.

For such speeds it compulsory to use axis symmetric systems with a much easier static and dynamic balance. Those systems have as well the advantage of spreading the load on several teeth so that the modules are smaller and the gears also thus lower inertia and lower weight.

this gears are called "cyclo drive" ....

As promised here is the principle of the gear :

Did you consider another way of achieving the transfer function? (Eg Electrical- through say high speed alternators ? or hydraulic ?)

Your idea is quite interesting and if hydraulics will not be optimal for different reasons the couple electric generator+motor can be a simple solution with the remark that its efficiency will be lower than the one of a well designed gear transmission (as example a 2 or 3 stage planetary drive optimised for efficiency). If there is a problem of weight then it should be considered that electric machines for same functional characteristics are heavier than a gear drive.

I have looked into electrical transmission, but it is equally difficult to find an electrical motor/generator which could sustain the RPM level. I would most likely still need some gear reduction i.e. 2-5:1.

The first stage reduction could be a friction drive with hydrodynamic lubrication to avoid metal to metal contact. Even so the wear would still be a problem.

Such drives work based on sliding and the sliding speed will be higher than in a teeth meshing thus the wear risk due to a film interruption is greater. from the other point of view the weight will be more important than for a gear drive.

Thank you everyone for your ideas. I appreciate all of your help.

10 ftlbs is about 13.56 Nm * 60,000 rpm / 9550 = 87 kW or roughly 116 hp, worst case. At 50,000 rpm input it is 71 kW or ~ 100 hp. Still quite high, especially for that speed.

The ratio is between 12 and 20 from your given data. This surely calls for a "Cyclo drive" reducer. (http://www.smcyclo.com/)

You can buy these commercially but these are not rated for such high speeds. The next thing is duration? How long or rather for how many hours should the reducer last? is it 200 hrs, 2000 hrs or perhaps 50,000 hrs plus?

This makes a big difference.

As far as I know, you would need a specially designed unit for this job, not easy. Besides the Cyclo you could also use a differential gear reducer that still needs two pair of gears, which makes it probably easier to use two planetary gears in series as suggested by nick name.

That is likely a practical route but as was stated the lubrication would require an oil jet at least at the high speed unit. It is not an easy task.

Check out the cyclo, it makes the most sense here.

Why 87kW?

Torque = 13.56 Nm.

Rotational speed of 60,000 prm = 1000 Hz (1/seconds).

Power = (Torque) * (Rotational speed) = 13.56*1000=13,56kW

Hi Tam,

I am not sure where you have that formula from. Are you and engineer? Where did you go to school?

In my books power equals torque times rpm divided by a factor.

Metric: Power in kW = Tq (Nm) * n (rpm) / 9550

Imperial: Power in HP = Tq (ftlbs) * n (rpm) / 5252

Let me know if you agree. Can you tell where the 9550 comes from or how it is calculated?

I apologize! Forgot to multiply by 2Pi

As I explained above a cyclo gear drive is not the solution.

Following sketch shows whic drive confuguration seems to be optimal, but as mentioned several times by different participants it is quite impossible to find t available of the shelf:

Hoe it will help.

Nick

Very interesting concepts you presented nick name.

Regarding the Cyclo I did not find your comments on it and your reasoning why it would not work. Would you care to elaborate? Thanks.

All cyclo-x drives are based on an eccentric entry shaft. The drive object of the thread has an entry rpm of about 50000! You can imagine how difficult it is to statically and dynamically to balance a shaft working at such speeds.

I, personally, think that the drive as a whole has to be axis symmetric and accept several satellites so that the load will be shared by at least 2 (better 3) meshing.

The torque is small but the speed is big and it is of advantage to reduce the gear diameter. This can be done with a smaller module (I use the European approach) which is obtained by load sharing. The smaller the diameter the smaller the sliding speed on the flanks and the risk to local temperature increase and loss of lubrication film.

There are several other factors but a thread is not the place to develop the full theory and practice of planetary gears.

"Load sharing" can only mean several teeth in contact at the same time. Either by two or three gears on one mating pair or several gears such as present in a planetary with perhaps 6 planet gears.

"(I use the European approach)" Do the Europeans have some gearing system we don't know about? How does it differ?

Regards

In Europe we use the "module" and the basic circle is defined as d=m*z where z is the number of teeth. In the USA the diametral pitch is used much more often than the module.

The standard tooth has a head height of 1*m (a shorter version has only 0.8m) and the clearance at root is 0.25m. Usually the flank angle is 20° but depending on the kind of application it can be changed. An other preferred angle is 14.5°. For involute used for shaft profiles the angle 30°. 20° is a compromise in order to optimise as well bending strength of root as the length of a double contact in meshing.

Some companies working in same field for a long time use a different flank angle to oblige customers to buy spare parts only by them since the tools for cutting gears are very expensive. For instance in small speeds and high bending an angle of 25° was used.

In the diametrical pitch system d=Z/DP.

25.4 / DP = Module or vice versa. Other than that there is not difference between the two. The module system would not give higher load capacity. Tooth height and flank angle are all the same. Even splines of half tooth height are used in both systems equally.

Your last point is interesting, however. But does it really pay on the long run? I for one would get upset and would not buy machinery with an oddball pressure angle.

Looks like you have worked in Germany or somewhere in Europe for some time?

Hi,

What sort of size, weight and price range are we talking about here?

What sort of environmental factors?

Due to the very high speeds friction and the resultant heat generated by it are definitely significant factors. Because of these factors 'air' bearings may be the way to go, but that does not address the actual speed reduction issue itself.

I can appreciate that you may want to keep the information as to what it's application is for proprietary reasons, but can you give us some sort of a (general) idea of what the application is, what the power source is and what your trying to utilize the greatly reduced RPMs for?

As our generic 'Nick Name' Guru has pointed out, one finds such speeds and requisite reductions in turbines that have can have a VERY high power to weight ration by virtue of the very high speed with relatively low torque, but by the time it is geared WAY down it has significant torque at more usable speeds.

Addressing the above approach of an indirect power transfer with no actual mechanical connection, what sort of inefficiency is tolerable.

I was trying to keep this unit fairly portable. It would be best if it was under 15kg. Efficiency is important. I am trying to create an efficient, constant speed, variable load power unit. If I lose too much energy in a gearbox, I may as well use a standard internal combustion engine.

Not sure if you found a solution to your gear reduction problem. I'm working with a turbine powered project as well, but the unit I obtained is already geared down from 50,300rpm to 12,000rpm. Total power of the turbine is 150hp. The turbine is off of the Tiernay 28V generator; check ebay for low hour units.

Hope this helps.

Thanks DBazos. I'll watch for these gearboxes. Do you know how many hours these gearboxes are rated for as a MTTF?

What type of gear reducers are they using with the turbine?

Even though this is just only a little over a 4 to 1 gear reduction, with the rotational and surface speeds being so high, this would still make the lubrication between the intermeshing and rubbing gear surfaces quite challenging.

Lay I mention that the initial request was for 50000-60000 to 2500 to 5000 which means reduction ratios of 10 to 24. The transmission you mention has a ratio of 50300/12000= 4.19 which is a COMMON ratio for planetary gears. So that it does not seem to be the expected solution the OP asked for.

I am seeing more options around the 10k output range. 5k is optimal, but I may have more ability to work at the 10k range than I originally thought. I am also looking at reducing my input speed. Testing closer to 30-40k range is possible.