Author Topic: Counting the cogs  (Read 4529 times)

fergusd

  • Newbie
  • *
  • Posts: 3
Counting the cogs
« on: April 18, 2020, 07:43:55 AM »
Various sites e.g. Cycle Monkey and including Rohloff's own provide the 14 gear ratios to 3 d.p. as follows
  1     2     3     4     5     6     7     8     9     10    11    12    13    14
0.279 0.316 0.360 0.409 0.464 0.528 0.600 0.682 0.774 0.881 1.000 1.135 1.292 1.467

Can anybody point me to the actual cog:cog integer ratios that are engaging within the hub to get these numbers?
I tried fiddling about with a few e.g. Gear 14 = 44:30 - possibly; also noticed that gears 8,9,10 are respectively the reciprocals of gears 12,13,14 - at least, to 3 d.p.; finally wondered whether the internals might be sufficiently complex for any or all of the gears to involve not just two cogs n1:n2 but three n1:n2:n3; .. ..?  Anybody?

leftpoole

  • Guest
Re: Counting the cogs
« Reply #1 on: April 18, 2020, 11:29:48 AM »
Hello,
I cannot point you to anything inside this hub.
Why on Earth is this interesting?
Well, it must be boredom or being something of a gear freak nerd?  :o :o :o :D
I hope someone other than Rohloff can enlighten because I am going to be worrying if we all should know these things......
Happy days,
John

fergusd

  • Newbie
  • *
  • Posts: 3
Re: Counting the cogs
« Reply #2 on: April 18, 2020, 02:11:45 PM »
>> Why on Earth is this interesting?
Implication: it isn't? or shouldn't be?
Just from an engineering point of view. And given the column-miles of print on derailleur cogs and ratios, surprising that there's nothing. The Wikipedia entry on Rohloff is quite enlightening and explains not just the mutual relationships 8/14, 9/13, 10/12 pivoting around 11 as suggested above, but an identical symmetry 1/7, 2/6, 3/5 around 4. The gears are clearly stepped so that each load (effort) increase n to n+1 is as near as can be contrived identical for all n from 1 to 13.
For me. It would be nothing short of Completely Fascinating to know what cog numbers have been designed into the hub to achieve this remarkable ambition. Maybe this is freaky geeky but, crikey, every purchaser of a Thorn bike, and every single contributor to any of the Thorn forums is somewhere on a spectrum. As is all too evident. And all the better for that .. ..

PH

  • Hero Member
  • *****
  • Posts: 2396
Re: Counting the cogs
« Reply #3 on: April 18, 2020, 04:21:02 PM »
I'm not sure I've fully understood the question, the term cogs is possibly misleading, the hub being made up of sun and planetary gears.  There is no need to understand it, which is a good job as my grasp is limited. The Rohloff website has the detail of which gears/cogs are engaging in each selection, it is fascinating to watch, though it's done more to increase my appreciation than understanding

https://www.rohloff.de/en/experience/technology-in-detail/gear-steps-1-14/

martinf

  • Hero Member
  • *****
  • Posts: 1167
Re: Counting the cogs
« Reply #4 on: April 18, 2020, 09:19:55 PM »
As far as I understand it, the Rohloff is basically a seven-speed hub with an "underdrive" that reduces all seven of the topmost gears to give an identically (in percentage terms) spaced range of much lower gears.

As you say that gears 8,9,10 are respectively the reciprocals of gears 12,13,14, the "seven-speed" is almost certainly a simple one like the old SRAM Spectro 7 rather than a more complicated arrangement like the Shimano Nexus 7.

Correction to the above phrase in italics: the Rohloff isn't a simple seven-speed with 3 epicyclic sets like the Spectro 7, it has 4 epicyclic sets in the first stage.

I did once have enough curiosity to count the teeth on the sun and planet gears (and the gear ring) during disassembly of a Sturmey Archer S5/2, but have since forgotten.

As far as taking hub gears to bits is concerned:

- Sturmey Archer and SRAM three-speeds are easy to completely dismantle. Sturmey Archer S5/2 is a little more complicated as it is necessary to take the timing into account on reassembly, but still relatively straightforward. I strip these hubs down every few years to thoroughly clean out the old oil and check that the internal parts are undamaged. Until recently (not sure if this is still true), it was possible to get any necessary spare part for a Sturmey Archer three or five speed hub relatively cheaply, not that I needed to very often.

- SRAM Spectro 7 doesn't come completely apart like a Sturmey Archer S5/2, for example the planet cage and planets are a modular unit, but is still relatively easy to strip down, clean, lubricate and reassemble so long as you don't try to dismantle the bits that aren't supposed to be taken apart.

- Shimano Nexus 7 and Nexus/Alfine 8 are more complicated, so on these I haven't yet tried to do anything more than removing the internal from the shell and wiping it clean, lubricating it, then reassembling it into the shell. I might have a try one day if/when I break one, but if that happens, given the prices of Shimano spare parts it probably won't be worth repairing, cheaper to replace the entire hub, or at least the internal.

- With Rohloff, I reckon the idea is that it is supposed to be reliable enough never to require being taken to bits by a home mechanic, and too complicated to be done without the factory tools if by bad luck it does break. So I am not curious enough to try and take one to bits to count the teeth on the internal parts.
« Last Edit: September 02, 2021, 07:44:54 PM by martinf »

PH

  • Hero Member
  • *****
  • Posts: 2396
Re: Counting the cogs
« Reply #5 on: April 19, 2020, 01:12:02 AM »
- With Rohloff, I reckon the idea is that it is supposed to be reliable enough never to require being taken to bits by a home mechanic, and too complicated to be done without the factory tools if by bad luck it does break. So I am not curious enough to try and take one to bits to count the teeth on the internal parts.
I'm nothing like brave or stupid enough to try, but stripping a Rohloff  doesn't seem to require any more tools than a pick and a magnet.  It's well designed and made, but not really that complex apart from the gear selector.  There's a few youtube videos of it being done, I think this is the clearest
https://www.youtube.com/watch?v=lS6kj8g5On8


sudo

  • Newbie
  • *
  • Posts: 4
    • https://matt.signorini.id.au/
Re: Counting the cogs
« Reply #6 on: September 02, 2021, 02:01:46 PM »
I don't know why, but for some reason your question about the exact numerical gear ratios was strangely compelling:
  • 119,340 / 428,472
  • 3,818,880 / 12,075,120
  • 443,944,800 / 1,233,999,360
  • 1,326 / 3,246
  • 504,092,160 / 1,086,760,800
  • 4,932,720 / 9,348,450
  • 175,032 / 292,140
  • 90 / 132
  • 2,880 / 3,720
  • 334,800 / 380,160
  • 1
  • 380,160 / 334,800
  • 3,720 / 2,880
  • 132 / 90

It may disappoint you to know that I didn't actually count the teeth on each cog; that is actually harder than you might think to keep track of which tooth you started counting. No, I was lazy and looked up the patent document, which lists how many teeth are on the sun and planet pinions, and gear rings (wonders never cease).

The patent document unfortunately doesn't list the numerical gear ratios so I had to calculate them from the epicyclic gear formula. For gears where 2 or 3 epicyclic sets are involved, the fractions for each set all get multiplied together, so some of them are ratios of huge numbers.

It consists of a 7 speed first stage into a 2 speed second stage. Very brief overview:

First stage
It engages either one of the first two sun gears (or locks them out for direct drive) for 3 or 2 step reduction and either one of the second two sun gears (or locks them out for direct drive) for 2 or 3 step increase. The two halves can be added, eg. -3 step decrease plus 2 step increase is a 1 step decrease. Two combinations (-3 and +3, -2 and +2) aren't used, direct drive in gear 4 and 11 is done by locking out both sides of the first stage.

Second stage
This is simpler. In gears 1 to 7 it is a large reduction ratio, and in gears 8 to 14 it's locked out (directly driving the output to the hub).

First stage ratios are 90/132, 2880/3720, 3720/2880, and 132/90 from the sprocket end sun gear inwards (pick the ratios for whichever sun pinion(s) is/are engaged for the respective gear, and multiply). Sun 3 and 4 are reciprocals of sun 2 and 1 you will note. The second deep reduction ratio is 1326/3246 for the low range (gears 1 to 7), and locked out in the high range (8 to 14). I hope you already have a reasonably good knowledge of how the thing works, otherwise this paragraph probably won't make any sense at all :P

Why might you want this? Well, I guess you can plug the fractions into your Excel gear inch chart for ultra-precise calculations :)

sudo

  • Newbie
  • *
  • Posts: 4
    • https://matt.signorini.id.au/
Re: Counting the cogs
« Reply #7 on: September 05, 2021, 06:13:59 AM »
Update
You can simplify the fractions:
  • 3,315 / 11,902
  • 5,304 / 16,771
  • 34,255 / 95,216
  • 221 / 541
  • 38,896 / 83,855
  • 6,851 / 12,984
  • 4,862 / 8,115
  • 15 / 22
  • 24 / 31
  • 155 / 176
  • 1
  • 176 / 155
  • 31 / 24
  • 22 / 15

Not that I can do that sort of arithmetic in my head mind you. I wrote a small program to print out the list of prime factors of each number, then I was easily able to eliminate the common factors from the top and bottom of each fraction