Phaser transmission (Is it as easy and good as it looks?)

[John in CR]

I had to start a separate thread about this item I saw in Hal's thread. It's a really simple and apparently smooth working multi-speed gearbox meant for the cranks, but it seems ideal for a DIY transmission for our non-hub drives. http://www.nosmokemtb.it/UK_phaser.html

Other than nailing down the shifting mechanism to make it something trouble free, it's brilliantly simple and uses cheap off the shelf parts. Add in that it's more efficient than a derailleur with easily a 4:1 gearing range, and it just doesn't get any better. Well maybe it does, since imagine how small it would be using #25 chain, which is better at higher speeds anyway, and I think stronger than regular bike chain, but then you'd have to fabricate the 2 cassettes.

I came to the realization today that geared hub motors can never stand up to high powered demands I have, since their critical parts have no sufficient thermal pathway for heat conduction out of the motor. Add in their limited surface area and the issue worsens. I have 2 mountains to cross and one of them will require a minimum of 1000W (probably more like 1500-1800W) continuous for an hour at relatively slow speed, so a non-hub drive is now mandatory for me. Yes it's heavy and yes it will take up a lot of space, but my 12kg hubbie is coming out of the wheel, and I'm attaching at worst a jury rigged version of the Phaser, even if I have to stop and manually change gears for my coast to coast ride. I look forward to HAL9000's attempt at one, and I'll share my own more feeble attempts, both successes and failures.

Maybe as a group we can come up with a simple and bullet proof DIY shifter for this thing. Maybe the inventor's shifting rig is simple, but I can't tell so, I'd like to discuss that further.

John

[cargo]

.

[John in CR]

Cargo,

Wow, I'm missed that there's a video, and from the pics I couldn't really see the shifting mechanism. After seeing it in action, I'm definitely going for it, because it seems even easier than I thought. Just a couple of long thin rollers to function like a front derailler on the pulling side, and a lightly sprung idler on the slack side. No worries for me about gearing down the motor, since I'm going to use a hub motor. My primary concern is how it will handle the 600-900rpm range, where my speed hub is most efficient. Since the highest gear will be at least a 2:1 "overdrive", I will have to gear down at the wheel, but to me that's a good thing as it will reduce the torque load in the "transmission".

Maybe Matt can come up with something similar with belts and a pair of multi-sized pulleys inverted to each other. Just have a long pull belt tensioner that you release to disengage like a clutch. Then the shifter moves the belt to the new gear, and you release the "clutch" and continue in the new gear. Something like that must already exist. The silence of a belt drive would be nice, although that chain drive was pretty quiet after it settled in each gear. You'd probably need a 2 wheel tensioner to have enough slack when disengaged to move the belt sideways easily.

John

[TylerDurden]

For ebike needs, only about 3 ratios would be needed. That should be even smaller & easier to build.

Bubba's experiment could be adapted...

viewtopic.php?f=6&t=5317&st=0&sk=t&sd=a&start=75

[John in CR]

After seeing the video, it seems like maybe the more gears in the stack the smoother it will run. I kinda like the idea of 7 equally spaced ratios to dial in an optimum gear ratio in terms of efficiency. It would seem beneficial with lesser grades or strong winds. I had a hard time finding 5 speed cassettes the other day, and no 3's, and the basic Shimano stuff dirt cheap down here anyway.

The guys who live in flat areas don't know how easy they have it with their ebikes. They don't have to fool with any of this stuff. When I was at the beach my bike was whispering to me, "Become a fisherman and move here...I'll take you as fast and far as you want to go. We can go for rides on the beach every day when the tide it out."

John

[HAL9000v2.0]

I have table tested that system with 2 shimano 7speed clusters. The problem is that ideal deluier trajectory for lower cluster is concave and for upper is convex so shifter must go linear beetween them. Also the teeth number must encrease linear what you can find only on some expencive street racing bikes. I think that ideal sprockets will be with two tooth increase one cluster with odd teeth number and one with even. also I noticed that it shifts much better with narrow type of chain (9speed) then with 7 speed so the sprockets must have spacing a little bit smaller than 9 speed cluster. I tried to see phaser teeth number but it is hard to tell, also it seams that it is linear cluster.
My conclusion is that is not easy nor cheap to build it. ( That is good enough reasont to do it...)
From stock parts you can use: left crank, nuts for cranks and two small sprockets for shifter...

I like his sistem with two long rollers on other side. My idea was to make two PA plates wich guides chain on both sides but this is much simpler...

[John in CR]

I have table tested that system with 2 shimano 7speed clusters. The problem is that ideal deluier trajectory for lower cluster is concave and for upper is convex so shifter must go linear beetween them. Also the teeth number must encrease linear what you can find only on some expencive street racing bikes. I think that ideal sprockets will be with two tooth increase one cluster with odd teeth number and one with even. also I noticed that it shifts much better with narrow type of chain (9speed) then with 7 speed so the sprockets must have spacing a little bit smaller than 9 speed cluster. I tried to see phaser teeth number but it is hard to tell, also it seams that it is linear cluster.
My conclusion is that is not easy nor cheap to build it. ( That is good enough reasont to do it...)
From stock parts you can use: left crank, nuts for cranks and two small sprockets for shifter...

I like his sistem with two long rollers on other side. My idea was to make two PA plates wich guides chain on both sides but this is much simpler...

I seem to remember something from Sheldon Brown that related to a harmonic balance in the chain length and number of teeth on the gears, that would apply here. Even numbers in the chain length is what I remember. We'd definitely want the combined teeth count of each speed to be identical across the range.

John

[GGoodrum]

I really don't any more than 3 gears are required, for any of the RC motor-based setups we've been discussing here. I think you'd end up shifting as fast as you could, to get through all these gears, and it really isn't required. For your setup, John, where you are talking about using a hub motor, in a non-hub motor setup, I really think you just need a two-speed transmission, something that would let you double the torque for a given speed.

-- Gary

[HAL9000v2.0]

John in CR: That is what I am talking about...see picture...

GGoodrum: I agree that no more than 2,3 speed is what we need but I think that Phaser type gearbox can't be done with 3 gears because shifter angle will be to steap, or it can be done but with very close ratio and that is no point.

[Miles]

I still think 2 sprocket/chain pairs with a dog-clutch change, is the way to go....

Or shift one sprocket across splined shaft.

[John in CR]

I still think 2 sprocket/chain pairs with a dog-clutch change, is the way to go....

Or shift one sprocket across splined shaft.

I wish I understood what that would look like. It would probably be easier if I had ever seen inside a Sturmey Archer 3 speed. I'm just trying to figure out things I can build, so if it's easy, then I'm game. A 3 speed internally geared hub would be the ideal, if they were sized to handle our loads and had bigger top to bottom ratios.

Maybe I should just bite the cost and relatively small efficiency bullet and get a Nuvinci. Don't they have close to a 3:1 range? Knowing my luck I'd get one that slips, or is that just some rumor started when one guy got some slippage with a max'd out Etek? How about some actual efficiency numbers instead of a comparison to a dirty derailleur?

John

[natem]

If you want 3 speeds then planetary is the way to go. It's a very strong, very compact design.

See: http://science.howstuffworks.com/gear7.htm

That's the sort of thing that is used in those 3-speed hubs and is used it automobile automatic transmissions.

I would think that it's possible to find small 3-speed transmissions for industrial or small engine uses...

[John in CR]

John in CR: That is what I am talking about...see picture...

HAL,

I don't understand the purpose of odd numbers of teeth on one and even number on the other. Also, looking at the closeups of his cassettes, I don't see any perfect symmetry in the the teeth increase, but I do see a nice straight line for the change in diameter of the sprockets. I guess that's what to look for in selecting cassettes. Are you saying that only comes with the expensive stuff?

John

[HAL9000v2.0]

It is better to have even/odd cimbination because it is quieter and less wear of chain. When you have odd/odd os even/even conbination the little rollers on chain hit sprocket tooth simultaniusly and amplify the noise.

[paultrafalgar]

...
Maybe I should just bite the cost and relatively small efficiency bullet and get a Nuvinci. Don't they have close to a 3:1 range? Knowing my luck I'd get one that slips, or is that just some rumor started when one guy got some slippage with a max'd out Etek? How about some actual efficiency numbers instead of a comparison to a dirty derailleur?

John

+1 on that john; I'm dying to get independent assessment of NuVinci efficiency. NuVinci+Matt's gearbox+RC motor all coupled closely/stealthily to the rear wheel of my tadpole trike

[TylerDurden]

It is better to have even/odd cimbination because it is quieter and less wear of chain. When you have odd/odd os even/even conbination the little rollers on chain hit sprocket tooth simultaniusly and amplify the noise.

Seems like the chain would do this regardless of odd/even, as the driven sprocket aligns to the gaps in the chain.

[John in CR]

...
Maybe I should just bite the cost and relatively small efficiency bullet and get a Nuvinci. Don't they have close to a 3:1 range? Knowing my luck I'd get one that slips, or is that just some rumor started when one guy got some slippage with a max'd out Etek? How about some actual efficiency numbers instead of a comparison to a dirty derailleur?

John

+1 on that john; I'm dying to get independent assessment of NuVinci efficiency. NuVinci+Matt's gearbox+RC motor all coupled closely/stealthily to the rear wheel of my tadpole trike

Paging Mitch.....He was selling the darn things for a friend, so he should have more than just marketing hype by now.

John

[John in CR]

It is better to have even/odd cimbination because it is quieter and less wear of chain. When you have odd/odd os even/even conbination the little rollers on chain hit sprocket tooth simultaniusly and amplify the noise.

Seems like the chain would do this regardless of odd/even, as the driven sprocket aligns to the gaps in the chain.

Early in the morning and I got a shot already of my learning something new every day. I googled and ended up at the Stand Handbook of Chains in google books, http://books.google.co.cr/books?id=BIJjtXB_qHQC&pg=PA130&lpg=PA130&dq=odd+and+even+number+of+teeth+on+sprocket&source=web&ots=R8OJ5AUxCI&sig=Y7JLXFUWn9rKvPZl_r-3QiD99yw&hl=en&ei=6PyOSey2OJW6tweixdGNCw&sa=X&oi=book_result&resnum=2&ct=result#PPA143,M1

The long and short is always an even number of links in the chain. Also, odd numbers of teeth on smaller sprockets is better, because the pin link elongates with wear and the roller link doesn't. With an even number of teeth, each tooth hits the same type of link on every revolution increasing wear. With motors we will really be stressing chain compared to its design use, so this stuff is important. For a single speed setup, Sheldon Brown suggests even numbers of teeth on the sprockets, so the chain and sprocket wear together and match longer, but that also requires you properly match phase of the chain and sprocket over the long term.

I can't find anything on your point HAL, regarding one odd and the other even. I understand how they will contact the chain at a different time, but I can't find anything to support that's better. Do you have a link? The uneven wear issue tells me always odd on the sprockets for a single speed, but with multi-speeds the chain will change position anyway.

John

[HAL9000v2.0]

Do not know the link ..it is something that I remeber form years ago.. maybe I am wrong.

[HAL9000v2.0]

Just checked in my "machine elements" book from faculty"... smaller sprocket is preferably even number of teeth..." The book dont say why!?!?
...
Also see your book page 178 on the bottom.

It seams that we can go with identical custers...
It will be nice to find 8 or 9 speed cluster or cassete form 11 to xx with linear grovth of teeth.

We need FW0713 Dura-Ace 10 cassette...

[Miles]

It will be nice to find 8 or 9 speed cluster or cassete form 11 to xx with linear grovth of teeth.

Hal,

You can also build your own using the individually available Marchisio sprockets:
http://clemenzo.com/index.php/content/v ... 4/lang,en/
http://clemenzo.com/index.php/content/v ... 7/lang,en/

[John in CR]

I wonder if the gearing needs to be so gradual. A four speed with 11-15-19-23 teeth or a 5 speed 11-14-17-20-23 should both be pretty easy to find and stack and still give over a 4:1 range.

I'm thinking that the tracking bar for the shifting mechanism should be angled in both dimensions. Then a combination of the rollers acting like a front derailleur, and the spring arm and 1 idler of a rear derailleur will do the trick. My goal would be to put the 2 cassettes closer together and very compact, even if it means only shifting just one gear at a time. With a steeper increase of ratio changes I don't see a need to jump so many gears anyway. Fewer gears would seem easier to pull off anyway...maybe even just a 3 speed with 11-15-19 giving a 3:1 range is sufficient, and maybe get away with just a normal front derailleur and rear derailleur tied together on a single shifter.

John

[John in CR]

Now that I think about it, why even bother with the separate gearbox for a street bike? I'm into easy and the efficiency of a derailleur is fine, so why not just go for the main benefit of the Phaser for our use, a straight chain to reduce wear and sound? Use a normal cassette at the wheel and in inverted one on my out of wheel hubbie. Use normal derailleurs, and so what if the front one is limited to 3 or maybe even 4 out of a front derailleur using cheap 7 speed cassettes. Then I can still get 3 or 4 speeds and a straight chain, and with one of those Mega Range cassettes on the rear, have a good range of speeds. If I get lucky enough to match total tooth count for each speed and successfully tie both derailleurs to a common shifter, then I might even be able to get rid of the "S" in the rear derailleur.

Before I was already thinking of an inverted cassette on the front, and even if I save Phaser attempts for rainy season experimentation, some important issues arose in this thread about the use of bike chain with electrics, since bottom bracket drives are known to wear chains and sprockets more quickly. They are:

1. Align everything correctly for a straight chain at the most used speeds.

2. Use sprockets with odd numbers of teeth on at least the most used sprockets, so the chain and sprockets wear evenly.

I think this will go a long way toward mitigating my concerns about sound and dependability for BB drives.

John