Flywheels

[allanferguson]

Has anyone ever contemplated gearing up a flywheel -- say by 1:3? This would have the same effect, in theory, as a flywheel 3 times heavier. There would, no doubt, be practical problems, but in a larger locomotive I'd have thought these could be overcome.

Allan F

[Brian Harrap]

Has anyone ever contemplated gearing up a flywheel -- say by 1:3? This would have the same effect, in theory, as a flywheel 3 times heavier. There would, no doubt, be practical problems, but in a larger locomotive I'd have thought these could be overcome.

Allan F

I did try this Alan. Firstly the motor became severely overloaded, took ages to get up to speed. Also a serious unintended consequence was that the loco (just a mocked up Chassis really) preferred to travel straight on its original trajectory as soon as it came to a curve in the track. And it shook itself to bits. I filed all the bits under 'non-starter'. Brian.

[Petri Sallinen]

Anyone who hasn't driven a loco with a Dynadrive clutch (other clutches are available) ...

That´s interesting to hear that other clutches are also available. Please let me know where.

Petri

[allanferguson]

Has anyone ever contemplated gearing up a flywheel -- say by 1:3? This would have the same effect, in theory, as a flywheel 3 times heavier. There would, no doubt, be practical problems, but in a larger locomotive I'd have thought these could be overcome.

Allan F

I did try this Alan. Firstly the motor became severely overloaded, took ages to get up to speed. Also a serious unintended consequence was that the loco (just a mocked up Chassis really) preferred to travel straight on its original trajectory as soon as it came to a curve in the track. And it shook itself to bits. I filed all the bits under 'non-starter'. Brian.

I do remember the late Professor Eric Laithwaite demonstrating the effects of heavy flywheels in apparently defying gravity. And I do have experience with gyro compasses -- but I wouldn't profess to understand them. Interested that you tried the idea.

Allan F

[Horsetan]

Anyone who hasn't driven a loco with a Dynadrive clutch (other clutches are available) and a LARGE flywheel, as in Petri's example, has a rare treat in store if they ever get the chance to try it, Brian.

I tried ordering another clutch from Brimalm last year, but their webshop has no facility for direct payment by card. Can anybody help me with this? I really need a second clutch for my ProtoDeltic....

[David Knight]

Hobbytown in the US used to do a centrifugal clutch with a flywheel drive for their Alco RS3s. I don't know if they are even still in business but perhaps on eBay?

Cheers,

David

[Re6/6]

Hobbytown in the US used to do a centrifugal clutch with a flywheel drive for their Alco RS3s. I don't know if they are even still in business but perhaps on eBay?

There was this from Squires, sadly after much research, no longer available.
An excellent clutch as Brian will confirm.

001.JPG

002.JPG

There is/was this from ModelTorque in Oz, but unfortunately, I believe that the maker has passed away, so that source has now dried up, so Brimalm in Sweden seem to be the only supplier of any sort of clutch. Unfortunately as Ivan has stated, they seem to be difficult to deal with.

I'm treasuring these two sorts that I've and have yet to decide which loco kit to use them in for best effect.

This is one of Brian's installation in a P87 Roco Dutch 08 which is sublime to drive.

model torque 2.jpg

Now from the sublime to the ridiculous!
This is a Lima Cl 60 with a motor from a 'Blue Streak' ICBM guidance system (according to the bloke that was selling them in France).
This project was a done as an amusement. It's certainly fun to drive and the 'over-run' is remarkable. The driven bogie was from a Hornby Cl 60. The compensated bogie came from John Lythgoe's Formil. Again no longer available. All these drive methods seem to have been overtaken by DCC nowadays, but it's still a shame that all this excellent stuff has disappeared.

002.JPG

[David B]

I tried ordering another clutch from Brimalm last year, but their webshop has no facility for direct payment by card. Can anybody help me with this? I really need a second clutch for my ProtoDeltic....

I have bought one or two things from abroad where I have done a bank transfer. Costs a bit extra but it gets the job done. One one occasion I sent cash €25, not something I would usually do but it solved the problem. I have also found the people on the other end helpful if asked because they have had other customers in a similar position. Try an email.

I suggest you ask your bank what options there are.

David

[Horsetan]

I tried ordering another clutch from Brimalm last year, but their webshop has no facility for direct payment by card. Can anybody help me with this? I really need a second clutch for my ProtoDeltic....

....Try an email....

Tried that, but never had a reply.

[Tim V]

All these drive methods seem to have been overtaken by DCC nowadays, but it's still a shame that all this excellent stuff has disappeared.

Got it in one, these clutch/flywheel tricks have easily been overtaken by what can be done with DCC. There is little incentive to keep producing this stuff.

[Russ Elliott]

This is one of Brian's installation in a P87 Roco Dutch 08 which is sublime to drive.

That big flywheel makes all the difference - I found the ModelTorque's drive characteristic (without a flywheel) absolutely infuriating. Hated it!

OTOH, Dave Haswell's Class 37 had a Brimalm clutch and the largest of John Lythgoe's flywheels, and was great fun - you had to put the motor into reverse to stop the thing.

[Brian Harrap]

This is one of Brian's installation in a P87 Roco Dutch 08 which is sublime to drive.

1, That big flywheel makes all the difference

2, OTOH, Dave Haswell's Class 37 had a Brimalm clutch and the largest of John Lythgoe's flywheels, and was great fun - you had to put the motor into reverse to stop the thing.

1. Yes the flywheel is an absolute must, bigger the better.

2, Can't do that with DCC.

The whole point of real flywheel inertia drives, with or without some sort of clutch is that one is dealing with/controlling REAL mass. I have played with DCC locos and I'm not knocking it, it's very good and has 101 other things going for it but inertia wise it's no contest. Brian.

[Tim V]

Programmable inertia Brian It has to be turned on.

[Russ Elliott]

Tim, I'm something of an ignoramus on DCC matters - in practice, how good is CV03 programming between speed step 'zero' and the first speed steps '1' and '2' etc?

[Natalie Graham]

Tim, I'm something of an ignoramus on DCC matters

So am I, which leads me to wonder, given posts above contrasting the two, are the clutch and flywheel type drives incompatible with DCC?

[grovenor-2685]

in practice, how good is CV03 programming between speed step 'zero' and the first speed steps '1' and '2' etc?

I don't really understand your question here. CV03 adjusts the acceleration rate, so far as I know the decoder will move through the speed steps at a constant rate depending on the value you put in the CV until it reaches the new speed that you have set with the controller. Cv03 = 0 will have 0 delay, CV 03 = 4 a noticeable delay, CV03 = 128 a very long delay.
Normally you set the voltage of speed step 1 with CV02 so you don't waste a lot of speed steps without the loco moving. And if you want a non-linear response curve you can either use Cvs 05 and 06 to create a knee in the curve or use the speed tables, CVs 67 - 94 to make a fully customised curve.
Keith

[Tim V]

Russ makes a good point.

I will say that the perception of a loco changing its position from one of rest to one in motion - that moment where it starts to move (and stop), is something that is very difficult to achieve, but can be programmed into a chip. However, the more expensive (high end - CT, Zimo) chips will more easily deliver this.

DC can deliver this with very careful use the regulator knob, but my observation at shows is that under exhibition conditions, operators rarely achieve this. Invariably there is a jerk start/stop.

Once programmed, DCC will deliver this every time. It is a combination of CV2 (start volts), CV3 (acceleration) and CV4 (deceleration). Nigel Cliffe of this parish gives some guidance on the 2mm website.

Use of huge flywheels slows down the motor - so can achieve the same thing, but as has been pointed out, to get a decent sized flywheel in results in other problems. Mechanical clutches go some way, but are now scarce.

And it is also possible to turn inertia on and off in DCC - very difficult with flywheels.

[Russ Elliott]

I understand the purpose of speed step voltage setting with CV02.

What I do not understand is how a CV03 setting, which if I understand Keith correctly is a delay setting determining how long it takes for the controller to change its command from speed 'X' to its next speed 'X+1', can be construed as an acceleration between those two speeds.

[Alan Turner]

What I do not understand is how a CV03 setting, which if I understand Keith correctly is a delay setting determining how long it takes for the controller to change its command from speed 'X' to its next speed 'X+1', can be construed as an acceleration between those two speeds.

Because to move from an initial velocity to a new velocity there has to be an acceleration.

Alan

[grovenor-2685]

acceleration = (new speed - old speed)/time.
So the longer you make the time the lower the acceleration.
Keith

[Russ Elliott]

Well yes, of course, Alan - I see lots of DCC things parading themselves like (a), where the acceleration alternates between zero and a high value, and I suppose my question is whether the CV03 setting can play any useful role in setting an acceleration that is a more constant value.

cv03-query.gif

[grovenor-2685]

If it is visibly doing your (a) and its not the driver doing it then I would suggest its probably set to 14 speed steps and an excessive amount of delay. For smootrh acceleration use 128 speed steps then reasonable values of delay will not exhibit such symptoms.
Some decoders use half steps when adjusting speed but that is just a design feature, not something that can be set by CV3.
Keith
PS. one can always add a flywheel to smooth out the acceleration curve

[nigelcliffe]

If the chip is fairly crude, then it will step up the speed range through just the speed steps selected. This might only be 28 (user option on throttle). With some locos this will give a noticeable change in speed when moving from one to another. Equally if the lowest speed possible on the chip is actually quite high, this will result in speed 1 being a noticeable jump from speed zero; I tested various chips against analogue control before deciding to chip my locos - some makes of chip were significantly worse than analogue in this aspect (so not worth the change!), but a few were much better (so it is worth the change if you choose carefully!).



If the chip is more sophisticated, then it uses rather more "internal" speed steps than the nominal number transmitted by the command station. So, the chip when moving from 0 to 1 (or 1 to any other speed) will interpolate rather more steps. ESU use 256 internal steps, mapped to the externally received speed instructions to smooth things, so moving from step 0 to step 1 (of 126) will involve one intermediate internal step for an ESU chip, or about 16 internal steps if the throttle is using a 28 speed step range. I'm fairly sure I've read that Zimo and/or CT may use 1024 internal steps, but can't put my hands on a reference for this (CT and Zimo share quite a lot of common design). With such smoothing, I doubt a watcher can detect the transitions in a reasonably decent loco setup (ie. appropriate gear box ratio for motor and loco, and an appropriate chip selection).


If the chip is really sophisticated, then you can more with the acceleration curve. Zimo have a notion of "exponential" acceleration (and matching deceleration). This allows the stretching of acceleration over part of the speed range. TCS have different acceleration rates between at different points on the speed curve (essentially allow three different rates through the speed range, or alternatively, two rates of deceleration depending on whether a "brake" button is used). Other makers no doubt have other solutions.


If the chip is well matched to the loco (both chip choice and setup of CVs), then the transition from stationary to moving should not be noticed by any observer. I've just measured a couple of 4mm locos on step 1; 30 seconds to move from one sleeper to the next. With a sensible setting for acceleration, this results in no perception of a jump between "stop" and "moving", which I find is essential to any illusion of mass in a model.

This smooth transition from stop to moving is a combination of decent mechanical design (in my case they are often High Level kits) and appropriate chip and chip settings. A bad exception in my fleet is the rather low gearbox ratio in a Railbus (Dapol body, Branchlines chassis) with a Lenz decoder. That does have a noticeable "jump" from stop to moving. Its a compromise because I can't be bothered to rebuild with higher ratio gearbox and the Lenz chip is there for it's internal auto-shuttle feature (so another chip with better low speed control of the motor isn't an option).



If mechanical flywheels (and absence of worm drives) are used, then Zimo have motor braking options available. These allow the motor to be used to apply reverse load to add braking to the flywheel. Or, again with Zimo and those into making mechanisms, a servo motor could be employed to actually pull brake blocks against wheels or flywheel....
If there is drive back-lash to be taken out before moving off in a new direction, Zimo have a feature for this as well.



So, its my view that superb control is possible with good mechanical parts and a good chip. That many DCC layouts have locos which lurch from "stop" to "3mph" without an in-between speed is down to less critical modellers who are using a sub-optimum mechanical designs, sub-optimum chips, have the chips set up poorly, or poor quality controllers.
The same lurch in speeds from stop to 3mph is often seen on analogue layouts; there are of course exceptions to this, often because the locos have careful choice of gearing, suspension, weight distribution and a decent analogue controller well matched to the motors being used.


( the regular mention of Zimo may be because their chips really are superb, or it may be my bias in tending to use them..... )


- Nigel

[Russ Elliott]

Thanks Nigel. I had a feeling that proper acceleration could be replicated only by using interstitial PWM states.