Miles' Retro-Direct Gearbox

Miles

100 TW
Joined
Mar 16, 2007
Messages
11,031
Location
London UK
This uses 2 synchro belts, running from pulleys on each end of the motor
shaft, with overrunning clutches between each pulley and the
shaft (back to back orientation). One or other of the drive paths will be activated, dependant on the
rotational direction of the motor. One belt drives a large pulley on the output shaft directly,
the other goes to a countershaft from where a small spur gear drives a large
spur gear on the output shaft - compounding the ratio and also reversing the
direction of rotation. So, while the direction of the motor can be changed, the
output shaft will always turns in the same direction. This gives you a 2
speed gear change by flipping the motor direction. Two more overrunning
clutches on the output shaft pulley and spur gear (orientated in the same
direction, this time), completely isolate the two drive paths.

The main attraction of this design is that the gear change is done electrically
and is therefore easy to automate - based on motor speed, for instance. One of the
drawbacks is that it can't be easily used for a regen. set-up.



Development Version 1

Top gear 3.4:1 Ratio difference 2:1

5mm synchro; 1.5 Mod spur gears; 8mm synchro output.


.
 

Attachments

  • Retro-Direct 3B 003.jpg
    Retro-Direct 3B 003.jpg
    52.8 KB · Views: 4,702
Very cool indeed. :)

I'm discussing with Richard different ways to implement phse wire switching using a standard HV110 as a controller. The motor needs to stop before reversing direction, which can be done using the brake option setting on the controller, but we still might need some sort of electronic interlock that keeps the throttle cut until it detects the motor is stopped. I was thinking with a 3-position switch you could simply have the center position cut the throttle signal, which would engage the brake. As I recall this will stop the motor very quickly, so it might be fine to have a very short amount of time the switch would have to be in the middle position.

-- Gary
 
Thanks Gary,

How do the boat controllers do it? Do they go straight from one direction to the other, using a two-way throttle?

What do you think would be the most generally useful ratio difference between the 2 gears?
 
The boat and car controllers are programmed so that a two-direction throttle can be used. Basically, the mid-point of the throttle is zero, but is actually at the 50% point in the range of movement (0 to 100%). Anyway, they handle everything in software because they program knows the speed and direction of the motor all the time. What we're trying to do is do this externally to the controller.

Also, I agree that 2x is probably perfect.

-- Gary
 
I suppose the market size is too small for Castle Creations to consider adding reversing capability to the HV110. :(

It would only take a few lines of code borrowed from one of their other controllers.
 
Industrial three phase motors can be reversed instantaneously by switching two of the phase wires, what happens if this is tried with an RC motor?

I assume the controller would get confused and shut down. Could this work with a sensored motor?

Even if it did work, it would likely be too much of a shock on the components.
 
PaulM said:
Industrial three phase motors can be reversed instantaneously by switching two of the phase wires, what happens if this is tried with an RC motor?

I assume the controller would get confused and shut down. Could this work with a sensored motor?

Even if it did work, it would likely be too much of a shock on the components.

RC brushless motors work exactly the same. If the motor spins the wrong way, you simply reverse two of the phase wires. If you stop the motor before switching the two leads, the controller doesn't know or care that the motor will now go the other direction.

-- Gary
 
Is 2:1 really the number? Gary says his 1.77:1 range with the cyclone is too much range with 1st gear being too low, but I keep coming up with 3:1 for my needs. Is it just because I'm focused on long 10-15% grades and running my motor at only partial throttle on the flats? I realize my 2kw without a sweat motor is different, and I want to run at 30mph with a tail wind at partial throttle equivalent to about 30V and 500W, and 15mph up hill at full throttle on 60V consuming only 1500W. It runs in the mid 80's in terms of efficiency from 600-900rpms at 60V according to the test report. Is that much different that other motors, or is it that mine has a speed wind designed as a 70kph hub motor and I'm not gearing for WOT on the flats? If I can get away with just using a Nexus3, which is pretty close to 2:1, it sure would simplify my build and testing plan.

John
 
GGoodrum said:
PaulM said:
Industrial three phase motors can be reversed instantaneously by switching two of the phase wires, what happens if this is tried with an RC motor?

I assume the controller would get confused and shut down. Could this work with a sensored motor?

Even if it did work, it would likely be too much of a shock on the components.

RC brushless motors work exactly the same. If the motor spins the wrong way, you simply reverse two of the phase wires. If you stop the motor before switching the two leads, the controller doesn't know or care that the motor will now go the other direction.

-- Gary

Yes, but what if you do it before the motor is stopped?
 
PaulM said:
Yes, but what if you do it before the motor is stopped?

I think the controller might be pretty unhappy, but it would be interesting to see what actually happens with a sensorless controller.

With a sensored controller/motor, if you hit reverse while it's running, you'll get a super large current spike as the back EMF of the motor is now opposite the battery power coming in. If the rotor has lots of inertia, it would almost certainly blow the FETs. I tried this using a Crystalyte analog controller and my small bench test motor. The motor pretty much reverses instantly and tries to jump off the bench.

It should be possible to brake the motor to make it stop quickly and use some kind of sensing to detect when the motor has actually stopped before allowing the phase wires to switch. It would be nice if it wasn't too complicated.
 
With my experience with RC controllers, if you interrupt the connection to them, even just 1 lead, they shut down and need to see zero throttle input to reset themselves before they will run again. Some controllers wont run again until you unplug them for a bit. But I'm sure this stuff can all be worked around. I love the idea of having D/W switching on the fly. Very very cool.
 
All RC controllers used in airplanes have a brake function. All you need to do to activate this is bring the throttle to zero. In the "hard brake" mode, a CC controller will stop the motor almost instantly. What I was thinking is you'd have a 3-position switch, one position for one direction, and one position for the reverse direction and then the center position being zero throttle. As long as you switch through the zero position on your way to the reverse direction, that should be enough to kill the speed long enough to go in the opposite direstion. You can also program the controller to ramp up the acceleration, so it won't matter if you don't let go of the throttle.

-- Gary
 
Here's a compacted version with a 1.5x difference in gear ratios.

I'll start on a detailed design of this one, I think.
.
 

Attachments

  • Retro-Dirext 3B V2.jpg
    Retro-Dirext 3B V2.jpg
    42.6 KB · Views: 2,508
The main disadvantage I see is the spur gears. Open they are noisy and prone to wear, however, it is difficult to enclose them while minimizing weight and keeping things sealed.

What if you just used two belts, but wrapped one around the motor pulley backwards while using an idler ahead of the motor? I don't know if timing belts exist with teeth on both sides, if not maybe a serpentine belt could be used. Another thought is to run the timing belt inside out; if the driven pulley is large enough, there would likely be enough friction with just the flat face. Belt life may be reduced though. I think the most compact setup would be with both belts on the same side of the motor. Use an extended motor shaft with another bearing supporting the end. The motor would be inline with the sprocket or pulley for your final drive.
 
PaulM said:
What if you just used two belts, but wrapped one around the motor pulley backwards while using an idler ahead of the motor? I don't know if timing belts exist with teeth on both sides, if not maybe a serpentine belt could be used. Another thought is to run the timing belt inside out; if the driven pulley is large enough, there would likely be enough friction with just the flat face. Belt life may be reduced though.

I've already investigated that avenue:
http://www.endless-sphere.com/forums/viewtopic.php?f=28&t=7679&start=0

Less efficient. Less compact.

Doubled sided synchro belts are expensive and only made in a few sizes - most are too long.

Could be done, as second stage, with chains.
 
PaulM said:
I think the most compact setup would be with both belts on the same side of the motor. Use an extended motor shaft with another bearing supporting the end. The motor would be inline with the sprocket or pulley for your final drive.

I did consider that. The overall width would be greater that way, but it might have advantages in some cases.

If you use a chain for the output, its sprocket can be anywhere on the output shaft, of course.
 
PaulM said:
The main disadvantage I see is the spur gears. Open they are noisy and prone to wear, however, it is difficult to enclose them while minimizing weight and keeping things sealed.

Agreed.

The gears would only be operating for short periods, though. The redundant drive path is completely isolated.

Nylon/composite would be quietest - they would probably only need shielding rather than enclosing. PTFE coated, hard-anodised aluminium could also be run dry and would wear well, but would be noisier. I'll try to find a good way to totally enclose the gears, though - the intermediate and output shafts have a more or less fixed relationship...
 
i think that everything I've ever owned that used plastic gears in some part of it, has always had the plastic gears be the weak-link in the design. Nylon often looks like a fantastic gear material on paper, but often seems to get soft and then strip under real life conditions.

For wide gears, just a slight bit of helix can make them run nearly silently. Even nylon hums and sings when it's straight cut.

In fact, for large scale RC planes, they sell adapters that bolt to the front of an RC motor and step the speed down 2:1, 3:1, 4:1 (maybe 5:1???). Some use a slick compact planetary gear setup with helical cut gears.

I didn't search very hard, but i found these smaller version units. They make some giant ones for huge planes.

http://www.icare-rc.com/gearboxes.htm

I know I've seen some that used helical cut gears to run quietly. Something along these lines could make for a good first stage to enter your gearbox.

You can also get planetary reduction boxes that are very quiet for large servo motors. They are compact and handle hundreds of ft-lbs of torque. Most powerful drills have them, and some very slick drills even have built-in 2 and 3spd gear boxes with bronze helical cut gears all cut into a very compact little package. They often use bushing rather than ball bearings, which would not last in an E-bike, but I bet you could source some gears, shafts, shift collars, and bits from them to help you in your project.

Something like this could help you get pieces:

http://www.ien.com/innovation-watch/cordless-drill-3/11904
 
In the larger RC helicopters, there are always a large platic gear driven by a steel motor pinion. The best ones of these I've seen are the ones used by Thunder Tiger on the Raptor 50 and Raptor 90 models. They use a hard plastic that impregnated with glass. They are much tougher than the soft nylon gears most other models use, and I have had a hard time stripping/breaking these, except in crashes. I tried using these in my very first ebike setup, using a Hacker A60-18L, shown below, but several teeth broke/stripped, mainly I think because the motor slipped under the big load. I had this mounted on my Townie, driving the large chainring with a 7T sprocket on the output shaft.

eBike-Hacker-02.jpg



Anyway, I think as a secondary drive, using a composite/metal gear combo will work just fine, and also be nice and quiet.

-- Gary
 
Back
Top