Tronic's AMC motor motion-controller with pwm/servo output

Fortunately wiper motors cost just 15Euros

I must congratulate the greek posting office... the parcel with RN-VN2 DualMotor arrived yesterday, but I can't pick it up until Monday morning!! :evil:

Me and my luck...

Hi to all,

I have good news!

First of all, I picked up the RN-VN2 this morning. Intact!

During weekend I had some progress on the controller. I programmed it to recieve on 38400 and move two simple servos...

The resuls are good. Its very fast and picks up even the smallest changes on acceleration data! Next is programming the servo code for the HB-25 with feedback pots, and try some PWM driving now that I have the RN-VN2 motor controller...

BTW, using the LCD helped me alot to debug the output!!

I'll use simple potetiometers for feedback for now leaving the use of optical encoders for next week (I hope). Another thing I want to do also is to find a proper crystal to be able to archieve 115200bps baud without sync errors!


Regards, Thanos

@yobuddy

perhaps you'll get the motors cheaper on the junkyard.

@ yobuddy

As Ego said it would much cheaper to get these wiper motors from a car junkyard!

And HB-25 is a nice motor controller. But I think you should wait a little to see the results with the PWM RN-VN2 motor controller that will have more direct and faster responce to full reversing situations!

Thanos

Btw, it should pe possible to control some servo valves with your interface :happy:

I had this in mind when started working on the old interface with the simple servos, but the prices of these servo valves were too high to go this way. Another one could attach a simple servo on a hand operated valve to have simular effect... :?

RaceRay mentioned this method some month ago. This was the initial idea of my post. Yes you are right!

But afaik RC servos only have a turn angle of 60 degrees!??
I have some manual valves, but they need ~15 turn to fully open/close.
Perhaps with some bigger sized valves than my 1/8 ones.

But this is another topic...

BTW, do you need some translation of the RNV2 driver module pdf-file?

ego

Well a normal servo won't do, but a modified one can!

You can use a normal DC motor with gear reduction box, and a multiturn potetiometer (I have seen some big metalic ones with up to 20 turns on a local electronics shop) that are driven by a HB-25 motor controller same way I do in the old interface.

Or you can even use the standard electronics of a simple servo if you use the same size motor altering only the potetiometer with a multiturn.

The real challenge would be the way of the mounting of the motor on the hand-controlled valve along with the pot. :

If anyone is still using pneumatics, and is willing to try this method I could start making such controller (in a new topic) that is easier to implement. I have all the nessasary code parts ready for this purpose...

@ Ego,
Thanks, I already translated all the parts I needed of its datasheet!

Thanos

AVR motion board development news

Hi to all,

I had some progress on the interface. I fixed various issues with inteferences during ADC and archieved 115200bps stable serial communication speed by overclocking the chip from the standard 16Mhz to a 24Mhz crystal!!! Its running very fast now! It responds instantly compared to the old BS2 interface! And its even faster if I ommit the display of the data on the LCD! I roughly measured the entire cycle of the code to a 2ms!!!! WOW! That was fast! And adding the rest of the code (PWM, encoders, etc) it won't go more than 5ms.

In the video you can see how fast is the position control of the simple servos, that by the way are a little slow from their nature, to follow the position commands! I control the servos with 5ms refresh rate (via interrupt) and they consume 4 times more power than in the BS2 setup due to constant positioning that occurs many times in a x-sim cycle data packet... I wonder if the same happens on the wiper motors when using PWM!

http://www.youtube.com/watch?v=S5RaAWBxb08


I haven't much time to also try PWM driving since I had to resolve the astable ADC function to have some feedback about the position. Now the ADC has 9 bit resolution (0-1023). For these tests I'll have to construct a demo setup with the smaller motor (w/encoder mounted on its axis to also test encoder driving).

Another thing I must do is to chance the ATmega8535 to a ATmega16 that have available 4 hardware PWM channels.

See the test table I use to program and debug the controller! Ugly isn't?




Until now the new controller seems promising even if many parts of its functions need to added and cooperate toggether without conficts.

Thanos

Yeah! :gut:

I hope it will be possible to use 16bit encoder reading.

Hi Thanos,

First of all i wan't to thank you for doing this great project..

I folow your work for some time now and i'm impressed by your work.

When i look at your test table it's starting to look rather complicated for me, I have no knowledge of electronics. I hope it will be possible to build one myself. Like yobuddy i can't wait to start building.

I love the hardware accident..............

No, I thank you, and all other for your patience waiting...

The final controller won't look like this at all. It will be a small PCB board!

And if you can't build it for yourself you can give it to an etching service to prepare the board for you. Then its only placing the compoments on it and programming the firmware with a serial cable...!

This hardware accident wasn't the only one, just the funniest!

Thanos

I hope the same too! Or else I'll have to dedicate the encoding task to some ATtiny13...

Thanos

Let me explain why we need that much resolution.

For example. You have an actuator with high pitches what causes that the screw does not brake itself when cutting the current off. So the positioning has to be holded by the motor and its PWM.

If the controller has to hold the position, it has to fix the actuator between two encoder-steps. If these steps are too far away from each other, then the actuator will rumble much.

I think it will be better to do it like you mentioned before. Using some Atinys for 16bit decoding and PWM.

regards
ego

Hi Thanos

I agree with ego,

Example from trials: With my motor planatery gearhead setup it freewheels if it is not held by PWM power. With the analogue pots, I have tried 10k and 100k, it still searches slightly for centre. (sirnoname mentions this is a common trait with ADC) So the result is it jitters while being held even if opposing forces are set at 255 pwm on a total motor brake. But then again I do not think I am using 9bit, my steps only go up to 255.

So I am keen to trial digital encoding to see whether I can remove the jitter while centering, although I think analogue will still have its uses for people with different motor systems - and it is cheaper and less complicated!

I also have a linear actuator (WIP) that is being modded by a mate. Target speed is aimed at 150mm/sec+. The base plate has been made and now I need to source a motor to drive the pulleys. The pitch of the screw is 10:1. I doubt this will freewheel as fast as the planatery gearhead system but there will still be a bit of 'sponginess' because instead of a wiper motor (like yours) rotating the screw it will be ..say a.. 2800rpm 200W motor and 3:1 pulley arrangement (300rpm=50mm/sec).
I would like to test this with your controller board too, your board is going to be a very versatile board once it is perfected! :top:

Evan

I believe that 9bit ADC resolution (0-1023) and 10turns multiturn pot will perform the positioning very accurate! If you consider that I use only a half turn pot (90 degrees) in my simulator, it will be more than enouph for driving and holding accurate positions with PWM.

After all some ramping will be used on the PWM driving to avoid overshouting and oscillations. Don't worry the ramping will be very small maybe 10 values of the 1023 of the ADC... and I can make this ramping adjuastable for various types of motors or even totaly disabled if the motor gearing can handle this.

I bought a 10 turns multiturn pot to use with the test table setup and it will be connected to the DC motor that will be driven with PWM. These multiturn pots are kind expensive (I got mine for 13euro!!!) but will be ideal to use in the place of optical encoders for start.

One other thing i'd like to ask you (evanf & ego), what is the best format for the PCB board? I design them in PCBexpress and printing the trace lines side on PDF. Is that enouph for you to be able to replicate it? And if you had to give the PCB board to an etching service, a PDF print would be enouph? Or they also need more info (for the holes position etc?).

Regards, Thanos