Armortek

Hi Mark, I have been discussing the very subject of programmable gear changes and therefore more realistic driving techniques/limits with Liam from down in OZ and he has achieved this very concept ..good work lets hope something comes of our concerns and ideas, it can only help the hobby develop regards Phil

Very interesting stuff. I'm currently programming Tiger 1 gear ratios into a raspberry pi to get the sound of gear changes/revs... It didn't occur to me "limit" the motors via this. It would actually make the sound more realistic too? Otherwise either the sound will be out or the engine will seem to rev/gear change far too quickly.

I'm still a long way off the sound only though. I've got rough indication of torque via the amp meters on the motors but I have no measurement of revs.

So many ideas and so little time

Cheers

Sam

A final thought before i retire to bed .. in the interim ....can we insert a resistor in line to the motors to lower the gear ratio and so deceive the sound unit into giving us higher revs for a slower speed?

The tanks we have don't have gears. At low speeds, the voltage to the motors is being chopped significantly, hence they have very reduced power. For a full size tank driving over an obstacle, you'd put it in the lowest gear and rev the engine up to the best torque output. For the DC motors in our tanks, full torque is only available to full power.

You can limit the end travel of a channel on the transmitter, but I'd have to check if it is possible to do each end separately. From memory, I think it is. That was you could configure forwards to be 0% - 100% and backwards 0% - 25%, but it will be gutless going when backwards.

A simple resistor wouldn't be possible as it would have to dissipate so much power it would need its own cooling system.

I've tried adding gear changes into the mix, using an Arduino, but the vagaries of RC signals and ensuring that the changes occur at exactly the rights points made realistic driving a real chore. Easy enough when driving in a straight line but when trying to perform manoeuvrers where different speeds are required, it was impossible to keep in the right gear. Rather than making the driving more realistic, it made it worse. If you're too quick on the sticks with the Benedini, it's already possible to do an apparent step change from 200rpm to 3000rpm, without anything between, Try doing that in a Challenger.

It also affects the granularity of the sounds. Without getting into it too deeply, on an 8 bit system there are 128 steps forwards and 128 steps backwards for a change from 0% to 100% speed. There's a dead spot in the middle, where the engine is at idle, so say 120 steps each way. If you wanted 5 gears changes, that means you have to split the 120 steps into 6, or 20 for each gear. This means the transition of engine speed will change from a relatively smooth 120 steps over the full range to just 20, which I found to be very noticeable and jarring.

The way I run the sound on the T-34 is to control the motors with the right stick, in single stick mode, and use the left stick just to control the Benedini speed input.

This has several advantages:

1) You can move around at idle, which seems to be quite common looking at the YouTube videos.
2) You can play around with the revs without moving the tank.
3) You can really rev the engine when approaching an obstacle, without it affecting the tank speed. It's very easy to simulate something like this (at around 2:10)

You do lose what would be the elevation control, but that's another story

At the risk of being seen to blow my own trumpet, by using one of my fan control modules, it's a piece of cake to link the smoker fan speed to the audible engine speed, rather than the physical motion of the tank.

Adrian.

I'm not sure I'll ever get any of it to work and even if I did if it would be any good but...

I wasn't thinking of limiting the stick movement. Rather send the stick movement via the onboard computer (bypass when computer says no). The computer would then have a model/simulation of the engine/gearbox that could limit how quickly it could rev up etc.

I'm just going to do the sound first which in some ways might be more difficult as I'll have to work back from what the tank is doing (power/revs) and fit that to the model of what the engine/gearbox is meant to be doing.

I love being able to be such a jack of all trades in this hobby. I'm also glad not to have to worry about weight like I had to with helis, planes etc.

Cheers

SAm

Sam please keep us posted with what you come up with ..good or bad its all very interesting regards Phil

Adrian Harris wrote: It also affects the granularity of the sounds. Without getting into it too deeply, on an 8 bit system there are 128 steps forwards and 128 steps backwards for a change from 0% to 100% speed. There's a dead spot in the middle, where the engine is at idle, so say 120 steps each way. If you wanted 5 gears changes, that means you have to split the 120 steps into 6, or 20 for each gear. This means the transition of engine speed will change from a relatively smooth 120 steps over the full range to just 20, which I found to be very noticeable and jarring.

Adrian.

That is one of the reasons behind my suggestion. Rather than 6 gears sharing 100% of stick movement, each gear could have 80% of stick movement, the upper band being used to trigger an up gear change, the lower band triggering a down gear change.
Presuming an 8-bit control of the motor as you say, the motor driving the model may only have 20 steps per "theoretical gear", but that is exactly what it has now. However each gear would have roughly 100 sound steps over the full range. 600 changes in engine speed sound, revving up, dropping down as new gear selected, revving up again as opposed to what we have now.
Is anybody likely to notice the engine sound has increased /decreased a little but the model has not speeded up / slowed down at all ?

Well, I solved the problem of too fast alterations of driving speed only by programming my transmitter. Modern computerized transmitters have a function like "servo speed delay", which allows to adjust the maximal speed a channel can react to a sudden moving of a control stick. This is a kind of life ensurance for my tanks when I am busy with transmitter and camera for the same time. More then once by following my tank I pushed the large an heavy DSLR against the throttle stick without intention; one time the stick even broke at this occasion!

My throttle and my steering channel are programmed to need at least 5 seconds from 'full forward' to 'full backward' and from "full right' to 'full left' position, independent of my speed of moving the sticks. This amount is alterable, but 5 seconds are enough to keep the running gear safe from damage.

I can understand what you're saying, but the granularity of the sounds are still controlled by the movement of the stick controlling the motor.

So the sounds can go from 0% to 100% as the stick goes from 20% to 40%, but whatever processor is listening to the changes from 20% to 40% is still only going to see 20 steps, so each step will be 5 times the size of the step as compared to when the sounds follow the stick motion from 0% to 100%
The way I would code it would be to read the incoming stick movement then, for instance, if this changes from 40 to 41, then the pulse controlling the sounds has to change from 72 to 79.

When it changes from 41 to 42, the sounds change from 79 to 86.

Any jitter in the movement signal will be magnified due to the bigger step changes in the sound signal.

The table above shows the progression of stick movement and sound changes, but the granularity of each "gear" would still be limited by the small number of changes over the range of the stick movement. I've used steps of two to limit the size of the picture, so it would be better by a factor of two that this might suggest.

Adrian.

Adrian Harris wrote:I can understand what you're saying, but the granularity of the sounds are still controlled by the movement of the stick controlling the motor.
Adrian.

What I was suggesting means that the sound would only be affected by a minimal amount, maybe 100 steps instead of 120 steps for each gear, but 600 steps in total, min to max rpm in each gear. The electrical motors driving the model may only have 120 steps in speed but using a seperate take-off, the sound unit being driven through a "virtual engine and gearbox" could go to 600 steps (up / down 6 times).
The stick(s) that was /were controlling the motor(s) now control the engine rpm, with the "virtual gearbox" converting it to the signal(s) to the motor(s)

With any of these won't you need some "input" for torque?

Full throttle in 1st gear will sound very different on smooth ice surface vs thick mud?

Interesting the idea of using "gestures" to change gear. I've only got old FM system with 20ms frame rate so would be hard to tell glitches from flicks up and down in gear?

Cheers

Sam

Different soundings of the motor depending on external load is a real problem. Electronical 'gears' however are graphical programmable by simply moving the knots in the diagram of the Beier sound module, which allows up to 5 gears forward an back. In each gear the throttle can be adjusted from 50 to 300 %. The pic shows my 2 gears forward and back, with the speed change 'jump' between the gears. Also switching sounds can be programmed.

I think I need stronger coffee to understand that diagram.

Stupid question: Tiger 1 driver would definitely go into lower gear and higher revs in anticipation of hill or rougher ground? ie before a full-automatic transmission would?

Maybe need three modes:
Semi-automatic (I only learned what this meant yesterday)
Full-automatic
Sports mode - for destroying tracks (to get back to original topic of this thread)

The diagram shows the rpm's (the vertical line) according to the milliseconds of the servo pulse (the horizontal line). A servo pulse of 1.5 ms means the middle of the amplitude of a joystick ('neutral' or 'stop'), 1 ms means fully back and 2 ms means fully forward. The knots mark the points of rpm's when switching the gear.

E.g. my chosen adjustment: tank stands and idles at 1.5 ms at 70 % of standard driving sound (the sound chosen for the slot 'driving sound' in the module; means 1,000 rpm). By moving the stick 'forward' the rpm's increase to 250 % of standard sound (= 2,500 rpm) at half the way the stick can move, then go down to 140 % (1,400 rpm) by automatical 'switching the gear'. Then the rpm's increase again ('2nd gear') to the maximum of 300 % (= 3,000 rpm) till the end of the stick path is reached (2,000 ms). Up to 5 gears and all points and rpm's are freely selectable. The only problem ist to move the stick so slowly, that all gears can be switched correctly up and down. This also was the reason for me to choose only two gears forward and back.

For considering variable load situations (uphill, downhill, sudden acceleration, braking) I think over a potentiometer driven by a pendulum in the tank hull, which controls / overdrives the volume of the sound amplifier. This would be rather easy to be installed.

Gerhard, I'm still trying to digest your posts properly. I'll try again tomorrow.

So I can mess about with ideas (and generate mind bending graphs) I've added fake inputs for amps and axle revs to my simulator and a chart to output calculated engine revs and throttle.
I've not done it yet but matching axle power to engine power seems like the best bet? Then an optional full automatic that tries to keep the engine in a preferred power band?

Axle power seems easy? amps * 24v * 6^3 ?

Cheers

Sam