Barrel Stabilization

[Vince Cutajar]

Now that I finished my Tiger 1 I find that I don't have a new project to start for now.

Something that I have been thinking about these last few days is barrel stabilization (not for the Tiger 1). The more I think about it the more I am getting interested. At the moment it is all on the drawing board and would be just as a proof of concept. Also, there's no guarantee that things will actually work out in the end.

The idea at the moment is to use an Arduino and a motion tracker something like the MPU6050 or MPU9250. Ideally stepper motors would be used but I will try to investigate the possibility of using the Armortek supplied geared DC motor instead. So the DC motor would be supplied by the Armortek module for normal operation and also by the stabilization circuit for stabilization purposes.

I already have an Arduino Uno to play around with. I also have an original Armortek geared DC motor that I haven't used so all I need is to get my hands on a motion tracker to start experimenting.

I am not an electronics whiz kid so I already encountered my first problem. As the same motor will be supplied by two separate 24v supplies I want to provide some protect to the circuits. If I insert diodes in the supply lines (see below) would that be OK or is there a board that can do this?

Vince

[Adrian Harris]

With both diodes facing the same way, the current will leave the Armortek module, pass through the first diode, pass through the motor, then be blocked by the second diode. For current to pass around that loop, the diodes have to face in opposite directions in the diagram (so they're the same direction in the current flow.

But that then means that the motor will only turn one way, as it will be impossible to reverse the voltage to turn the motor the other way, as both diodes will prevent that.

In any case, I don't think the geared motor will be fast enough to cope with the changes, which means it will always lag behind the control input, creating an inherently unstable system.

I've used a servo on the Grant - there's a preview in that section somewhere - but that only works with very strong servos and perfectly balanced gun barrels. I used the Arduino to control the stab, plus listen to the receiver to see if a control input had been requested, plus listen to see whether stab has been enabled or disabled. With the 6 axis board i was using, I seem to recall it took up to 10 seconds for the board to reach a steady knowledge of it's position in space. Any movement during that time would create errors which multiplied, until the thing became unstable. I think a production system would probably have to have a delay built in to give the system time to learn its orientation, before the tank could be moved. Probably best to generate a sound once the errors from the board were within tolerances.

Getting the turret stab working will be a much harder challenge, due to the weight and inertia in the system. Getting servos or brushless motors which are powerful enough to spin the turret whilst driving will be the major challenge.

[Vince Cutajar]

You're absolutely right about the diodes. Good thing I asked. So I guess it has to be some form of relay.

I remember the experiment you did and the video. It was the same time I was experimenting with my Arduino recoil.

As I said, this will just be to test ideas. I don't plan to make a fully working system because I don't have a use for it.

Vince

[Gerhard Michel]

Hi Vince,

there are ready-to-mount modules for turret & gun stabilizing using a simple DC-Gearmotor. Reinhold's huge Leopard II in 1/4 scale (total weight about 400 kgs!) uses such a module, available at the SGS shop

[Vince Cutajar]

Gerhard

there are ready-to-mount modules for turret & gun stabilizing using a simple DC-Gearmotor.

Do you have an internet link? I tried to do a search but was not successful.

Vince

[Stephen White]

Vince, you may not have seen my post on the Chieftain Thread.

Vince, Youngjae

Gun stabilisation is a holy grail for Armortek models. It's very much more complex than it looks. I hope with all the expertise we have on the Forum and at Armortek, we can develop a workable solution. I know it's been on the Armortek agenda for some time. I did a lot of experimentation on my Centurion with partial success, enough at least to understand where some of the challenges lie. Here are some of the issues:

- how many axes? Is traverse more or as important as elevation?

- how do you accommodate recoil, which moves the centre of gravity of the barrel in relation to the mantlet? Gun balancing is essential in any event, as is minimising friction with a linear bearing.

- how do you make the drives sufficiently responsive in order to eliminate lag? The current Armortek lead screw elevation system is unworkable (too slow). Servos may work but a stepper motor arrangement might be the final solution. The turret is a big mass with a lot of inertia. The traverse mechanism is a major challenge.

- how do you prevent collisions with parts of the hull which could foul? Do you ignore the problem or fit limit switches or.....

- how do you give the gun a point in space as datum? I'm now very certain that key to this is a so-called six axis sensor package. You need accelerometers, rate gyros and GPS to drive the system. Fortunately, these are available very cheaply. The stabilisation arrays available for model aircraft are not appropriate (you're only stabilising the platform itself rather than stabilising a sub-system in relation to the platform). Camera gimbals are more analogous and offer some promising potential solutions but - they all rely on brushless motors and cameras have nowhere near the same mass as our tank turrets and guns.

- how do you integrate control of stab with the RC system? - the most promising solution I've seen is based on Arduino, which allows the vital tailoring of response which will be required.

There's no doubt with dextrous hands and one stick drive control, you can to a degree simulate stab but a working system would be a big step forward. I've got a copy of a PhD thesis on gun stabilisation in the full size which shows just how complex the physics are so it's worth bearing that in mind in order not to be disappointed when a new, simple solution fails to deliver.

For my part, I got a prototype working on the Cent but found that it struggled with the mass of the turret and gun and quickly got out of phase with the hull. That led to a potentially self-destructive induced oscillation which was fun to watch but not what I was trying to achieve. I think that camera gimbal technology will eventually work for us for the sensor and control part of the package and the big challenge will be the physical drives.

Good luck to anyone trying this.

I'd add that these guys may have controllers and IMUs worth looking at. I haven't seen any evidence of anyone doing this successfully beyond the 1/16th toy tanks but Gerhard may be able to enlighten us. Delighted to see anyone working on this, as I said above, it's a holy grail for one sixth.

The big challenge with stabilisation is to overcome lead lag, ie the stabiliser controller becoming out of phase with the mass being controlled. It needs a very responsive and accurate feedback loop and actuators sufficiently responsive, accurate and powerful to ensure the control is deadbeat, ie with no overshoot or free play. That's not easy, as Adrian says, given the mass being moved.

Stephen

[Vince Cutajar]

Yes Stephen I have seen your post and actually that is what made me think about it. One could say I'm a sucker for these type of challenges and whether it works or not I'm not bothered because it would have kept me occupied. All I need to get is the motion tracking sensor.

Gerhard's friend model is 1/4 scale and that's a good lump of metal to move around so I would be interested to know what's being used. Unfortunately my search on the internet did not turn up anything.

Vince

[Adrian Harris]

Vince, you have to think in German, so if you search for "SGS modellbau" it throws up www.sgs-electronic.de

My browser attempts to translate foreign language websites, but doesn't understand tanks at all, so turrets become towers and barrels become pipes

Adrian.

[Stephen White]

Looks like this one:

https://www.sgs-electronic.de/webshop/2 ... TVC-GSU-12. The product description is:

This module is suitable for all models and stabilizes the pipe cradle and the tower. A gyro sensor and an acceleration sensor are used for this. The pipe (=barrel) cradle is controlled by a servo, the tower (=turret) rotation by a geared motor. A speed controller with 5A continuous load is already integrated to control the tower turret rotary motor. In addition, the module has a servo output for barrel retraction, which is triggered with a shot. The module can be operated directly on a receiver or operated via the Scalebus on an FO module.

It lacks an integrated GPS which is essential if you want the stabilisation to maintain the gun lay on a point in space. The IMU has got rate gyros and accelerometers, which is a start. Although it claims to be suitable for all scales, I suspect it's really intended for 1/16th, given the output. I've seen a reasonably successful implementation on YouTube using this but only on a small model tank. It would be interesting to know what Gerhard's friend has had to do and what functionality he has achieved.

If I had to start anywhere, I'd probably go for something like this:

https://www.basecamelectronics.com

With all the expertise here on the Forum, we must be able to crack this?

Stephen

[Phil Woollard]

As I am out of my depth with this one I shall keep very quiet but I just wanted to add that the operator will have to slow the model down considerably and drive in a more scale like manor, or there is not a chance in hell that the stabilization control system will ever keep up! Also the motors or actuators will have to be immensely powerful, and will therefore be quite large.
If someone can pull this off it sure would look the business!

[Gerhard Michel]

Hi Guys,

well, Reinhold (the organizer of the Hausen tank meeting) told me that Mr. Stelzer, owner of SGS, had modified his module for the heavy Leopard turret in 1/4 scale. I don't know which modifications he made, but I am confident that he would do the same for other customers. Therefore I think that an inquiry to Mr. Stelzer could clear all boundary conditions for using this module in 1/6 scale full metal models.

I for myself still didn't see the real functionality of the module in Reinhold's turret (overshooting etc.). Perhaps it is possible to make a video in future.

[Youngjae Bae]

I have been changing the main gun stabiliser and the recoil at the same time and have ordered all related orders through cheap AlliExpress and have been waiting for delivery for nearly a month now.
I finished the design with a little bit of the past experience of making automated robots in the company, and I've implemented the behavior in design, and I'm going to do it with a simple commodity available on the market.
Due to the Korona virus, the production is delayed due to delayed delivery, but I am sure that the reproduction of motion will make progress. Tomorrow, my laser workpiece will arrive.

Youngjae

[Vince Cutajar]

That looks interesting Youngjae. Looking forward to more updates.

Vince

[Gerhard Michel]

Reinhold confirmed that he uses the SGS stabilizer, originally made for 1/16 scale and modified by SGS for his Leopard turret with 85 kgs!! Turret rotation is made by a brushed electric gearmotor, barrel elevation by a strong servo. This servo needs its own power suppy (not via receiver) up to 8.4 volts, like all strong servoes do.

[Youngjae Bae]

I did some experiments to control the torque of the main gun and selected a suitable electric actuator.
The actuator has been considered with force and speed to process the Gyro signal in real time.
This whole process doesn’t necessarily guarantee success, but it seems to be worth the effort.

Youngjae