Go to Website Homepage Webmaster: Phil@scmec.us
Want to know more about the Club?

 

Back to last page

9 Speed Forward, 2 Reverse, Dual Shift Gearbox implementation for Grove TM9120 Truck Crane Gearbox.

Click on the appropriate thumbnail to see the relevant photo. Use you browser's 'back' button to get back to this page.

The Grove TM9120 Truck Crane was designed with a 'Roadranger' Eaton-Fuller 9 speed Transmission, giving 9 forward and 2 reverse gears, coupled to an auxiliary gearbox giving high and low range speed. The 9 speed unit is comprised of a 5 forward speed + 1 reverse speed section followed by a 2 speed section, giving 10 total speeds and 2 reverse. Only 9 of the 10 forward speeds are used.

Given my Meccano model will be very heavy and I plan to have it moving around the floor at a 'reasonable' speed, I decided that I would need some form of gearbox capable of near constant power flow to the rear axles, to overcome the usual problems of the model grinding to a halt during a gear change, which is a well known problem in models where the ratio of rolling resistance to air resistance is totally different to that of a 'real' vehicle.

1) Try #1. Planetary. dual 4 speed solution. An interesting puzzle. Nice idea, but didn't work!

I initially decided to try a planetary drive gearbox as I knew such a gearbox could be designed with near instantaneous gear shifts. Given the packing density required for this complex a gearbox, even in a 9.2:1 scale truck, where the gearbox unit could be approx 1ft long, my attempts failed. I tried using the modern Meccano plastic roadwheels with internal toothed ring (187c). This design ended up fitting and getting close approximations to the correct gear ratios, but the loss was too high, due to a) using the plastic parts and b) the basic power flow. Let me explain....

My scheme was to use a pair of 4 speed gearbox sections - a 'left' side 4 speed and a 'right' side 4 speed. Each of these 4 speed gearboxes would be a sequence of a 2 speed, followed by a 2 speed gearbox. 2x2=4 possible speeds per 'side'. By the use of a a differential to combine both 4 speed 'sides' and carefully selecting the interconnecting gear ratios, using Excel spreadsheets, I was able to 'cherry pick' 9 speeds of the 16 allowable (4x4) that would match my target speeds. Very much a 'suck it and see' design approach, taking about a month's worth of evenings 'messing with' Excel spreadsheets! Further, 2 of the 16 combinations of speed yielded reverse speeds.

The power from the motor was split and then fed through the two 4 speed 'sides' of the gearbox, to be joined together at the output shaft with the planetary differential. Mathematically, it was complex and all looked good! Meccano solution wise, it all looked good! Power efficiency wise, it failed!!

Planetary gearbox viewed from above. My plan was to use 2 sets of 4 Meccano motors as servo drives, one motor per 2 speed section. I never got around around building up the 2nd rack of 4 servos.
Each 2 speed planetary box was controlled by a pair of small brakes, one per gear. In this early view, Meccano train wheels were used as the brakes.
One 4 speed selector sections. Cams would drive the brakes.
Contrate reduction from the motor, followed by worm drive to the output cam would select a gear in less than 1/4 of a second!

The problem with this scheme was conceptually simple, but it never occurred to me until I had got as far as I had. The power from the motor effectively passed through the two sides of 4 speed gearboxes to be combined prior to output. When one side of the gearbox was set to minimal reduction and the other side to maximum reduction, the planetary gears would be 'back driven' up the 'maximum reduction' side to get all of the gears rotating at their desired speeds. This resulted in multiple thousands of rpm being fed back into the plastic annular ring based planetary unit nearest to the motor and the effective friction was massive - so much so that even with a 50W electric motor driving the gearbox, it would stall at the faster output speeds. So... major flaw in the design. :-(

2) Try #2. DSG with a pair of 8 speed gearboxes. Works like a charm!

After much puzzling, and a little coincidence, I found the solution. Alan Wenbourne in the UK had demonstrated the DSG (Dual Shift Gear) principle, as used in high performance cars these days. This principle is elegant, and somewhat simple. Assume one wants a 6 speed gearbox. Split it into a pair (side by side) of 3 speed gearboxes. Let's call them side 'A' and side 'B'. Design for 1st, 3rd and 5th gears to dive through side 'A' and 2nd, 4th and 6th speeds to go through side 'B'. A pair of clutches pass power from the engine into side 'A' or 'B'. If one is in 1st gear and power is passing through the 'A' side, then the gearbox can pre-select 2nd gear from the 'B' side. Simply redirecting power through the 'B' side by the clutch can yield a gear change as fast as 10mS without power flow to the vehicle's wheels. In 2nd gear, side 'A' can be set to 3rd gear (as no power is flowing through side 'B' at this time) and then clutch movement brings the power flow back to side 'A' to enable 3rd gear. Thus, one 'walks' up the gear selections, moving from side 'A' to side 'B'. To go back down the gears, to reverse sequence is used.

The basic configuration as shown by Alan Wenbourne suffers somewhat from the same speed multiplication issues that I ran into with my planetary design. As one increased the number of gears required, the gearing-up problems becomes a real issue. For a 14:1 ratio from top to bottom gear, as required in my Grove design, if I were in bottom gear I would need a 14:1total gear ratio, compared to a 1:1 ratio for top gear. If I were to use this configuration, whilst in 1st gear, the gearbox output speed would equal the input speed. Working back up the other side of the gearbox, under worse case solutions, I could have a 14x speed increase up the non-selected half. Ouch! For a car, with a much smaller top to bottom gear range, this is no where near as much of a problem.

My solution was to turn the gearbox configuration 'on it's head' and have both sides of gear box driven from the motor with 'forward' power flow and then move the side-selection clutches to the output shaft end. The problem this gives is that the output clutches have to be able to work over a 14:1 speed (and torque) range, yielding the requirement for very low loss, yet powerful clutch design.

2.1) The DSG Solution to the 'Roadranger' gearbox.

The block diagram above shows both the gear lever shift pattern to achieve 9 forward and 2 reverse as well as the basic power flow. If you click on the diagram you can download a larger format version.

Input power from the motor is supplied on the left hand side at point 'X'. Power is split to the upper and lower sections. Each section comprises three sequential gearbox arrangements, each comprising of 3 series 2 speed gear sections. For example, power from 'X' passes into the divide by A/B section. 'Switch' 'a' selects which divider ratio. So, a = X/A, or X/B. Similarly, speed 'b' works out to b = a/C or a/D, yielding 4 combinations for speed 'b'.

b = X/(A*C), X/(A*D, X/(B*C) and X/(B*D).

It takes a pair of servos, one controlling the A/B choice and the other the C/D choice, to yield 4 gear options. Add in another 2 speed section, E/F and we have 3 servos to get 8 speeds. Do the same for the lower G/H, I/J and K/L sections and one ends up with 6 servos controlling 16 speeds, all combined up by the final servo driven switch to select speeds 'd' or 'g'. Divider rations E/F and K/L act like the 'back end' 2 speed reduction unit in the Roadranger gearbox. So, units a, b, e and f give 8 possible speeds and the final d/g speed options act as high/low range to double this choice to 16 possible speeds.

As you can see, there is another mathematical block shown. Motor speed is always divided by M and arranged to give a reverse direction feed to a differential that adds speeds 'b' to 'h' to output speed 'c'. This is how reverse works. Dividers A and C, for example, can be set to yield speed 'b' to be less than speed 'h', so the differential output at 'c' is reverse direction. Once 'b' revolves faster than 'h', then forward direction is yielded.

With the correct choice of gearings, one gets:

Meccano match                
Path   Ratios Ratio Meccano RT-14709H g b c d Y
H-J-K   A 3   1 1.05       1.05
B-D-E   B 1   0.74626866   1.4 0.74375 0.74375 0.74375
G-I-K   C 1   0.54644809 0.525       0.525
B-C-E   D 0.75  3/4 0.38910506   1.05 0.39375 0.39375 0.39375
H-J-L   E 1   0.28011204 0.28       0.28
B-D-F   F 3.75  1.5*2.5 0.20833333   1.4 0.74375 0.198333 0.198333
G-I-L   G 4   0.15313936 0.14       0.14
B-C-F   H 1   0.10917031   1.05 0.39375 0.105 0.105
G-J-L   I 0.5   0.07524454 0.07       0.07
A-C-F   J 1   -0.0719942   0.35 -0.30625 -0.08167 -0.08167
A-C-E   K 1   -0.2570694   0.35 -0.30625 -0.30625 -0.30625
    L 3.75  1.5*2.5            
    M -1.6  4/2.5            
                     
    X 1.05              

This shows, to normalize the scaling we have to assume top gear gives Y = 1.05X. Only 9 ratio combinations, like path H-J-K through the block diagram gives 1.05 as the output speed, compared to input speed X. Similarly, at the slowest output speed end, path G-J-L gives 0.07 times X, or 14.3:1.

This table also shows how the mathematical speeds are implemented in Meccano. For example, to get M = -1.6:1, I use a 4:1 gear set to slow down the drive, followed by an Exacto 2.5:1 set to speed up the drive, giving 4/2.5 = 1.6:1. The high/low range gear sets on both upper and lower paths are 1:1 (high range) and 3.75:1 (lower range). The 3.75:1 is made from an Exacto 1.5:1 gear set followed by an exacto 2.5:1 gear set to give 3.75:1.

The column 'RT-14709H' shows the gear ratios in the Roadranger RT-14709G 9 speed gearbox. When plotted against the Meccano version (Y), one see a very close match.

2.2) The Meccano Implementation.

Here are a few pictures of the final solution. More will be added later, together with a more detailed explanation. Suffice it to say here, that the electric motor is a windscreen (windshield for us US folk!) wiper motor, capable of outputting at least 50W of power. 6 miniature servos control the constant mesh, dog clutch selected gearbox sections and one large, brass gear based servo operates the output clutches.

The motor/gearbox finished unit. View from the rear of the truck.
From the front of the truck.
Output clutches and clutch selection servo. Meccano helical gears are pressed into 177a space-set seats to give a pair of very solid brakes!
The gearbox 'dropped' into the chassis frame - just to prove it fits!

 

Last Updated: February 18, 2010