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

Thanos mentioned to use 8 FETs/motor in order to increase the potential of the driver.

And what about the IRFZ48N ones?

@Thanos
Hmm, I think you should open a new Thread for the DIY motor driver. Its a really great additional project, but it lets the controller thread go off topic.

regards
Christian

Yes you can use as many FETs in parallel as you want, making sure each bank has equal numbers of course!, this can be implemented to share the Current loading as opposed to using just one, BUT if one blows the whole bank can blow! yay joy :?

If the IRFZ48N is balanced with the IRF4905 in terms of similar low Gate threshold for example, then it's a good question, but why reinvent the wheel when there are many H-Bridge designs out there that use the IRFZ44/IRF4905 combination. There must be a good reason! :wink:

Hi

I've splitted the thread at this point in order to keep it more clear.

Thanos DIY DIY dual mosfet h-bridge thread is continuing here

regards
Christian

Btw: Almost finished

Re: h-Bridge

Hi Dirk,

The H-bridge I designed (atleast, its layout) is ready! You can take ideas from it to design an even bigger one by using parallel placed mosfets...
Can't really tell you more details on the how they work of the mosfets as I have not much knowledge in analog electronics. There is lots of information on how they work around the internet. Google it!
Edit: here is a quick find relative to your problem:
http://www.parallax.com/dl/docs/cols/nv ... l/nv52.pdf


As far as I can tell you should use a simple 270 degree potentiometer with the gears you have placed. With a 10-turn pot placed i assume that you are using only 10% of the resolution of the pot resulting in huge stepping!

In my AVR controller I read the pots in 10-bit resolution which means I have 1023 points of position feedback using just a 270 degree pot! Here is what my controller is capable to do: Reads the data from the serial port in 115200Bps, reading four pots in ADC 10-bit resolution, calculates the proportional servo drive values and drive four motors with 10-bit PWM at 24kHz frequency! Its fast! Its entire program cycle lasts less than 5ms!

Atmega16 is nice but If I were you I would try the much better Atmega644!!! Its a beast! More memory and variables, and better hardware PWM support for SIX channels!!!

If you have the courage to read all the 25 pages of this thread, you will see the benefits of my controller!

Regards, Thanos

@ego,

nice job, can you build me one too! :

just kidding, but I'm still waiting on my parts to arrive.

Hi all,

First of all a very happy new year to all.

As i understand we don't need the RN-VN2 anymore with the completion of the simple Hbridge by thanos.

i've made a simple image how the wiring will be. Is this correct or did i miss something.




[/img]

Hi Rovor,

Yes, its correct!

But it will look cooler if you place the dual hbridge motor controller the other way around to be able to read easier the description of the ports!

Happy new Year to all!

Regards, Thanos

Hi Thanos,

I've etched my first motion simulator pcb, and i must say after 1 failure i'm very pleased with the result.

pcb.gif (0 KB KB)

Now i can start building.

I want to start etching the Hbridge too, will there be any modifications or is this the final version, will you be placing a partslsit for the Hbridge too?

RoVoR

Hi Rovor,

Good etching job!

Be patience with the drilling process!

Yes, you can build the dual H-bridge motor controller too. Its layout looks like its going to be as is for a while. I haven't builded it yet and you might be the first that actually test it first! ;-)

I will prepare a partlist for the Dual H-bridge too, later. I want to be sure for the high power diodes first! You can figure the parts as they are named on the schematic and the silk layer...

Regards, Thanos

Hi, everyone

I think I found the most proper quadrature decoder chip to interface my optical encoder kits. Its the HCTL-2016 Quadrature Decoder/Counter chip with 16-bit operation and 14Mhz clock operation.

Now the bad news: It works in parallel! Which means that except its 2 control signal lines, you receive the data on 2 sets of 8 parallel bits (on 8 lines). This makes us 8 + 2 = 10 lines. And considering that I can attach only one encoder kit on this chip and two are nessesary the number of ports keep increasing, which is 10 + 10 = 20 lines! And even if I try to share the sames data lines (D0-D7) I would still need 11 lines to read 2 encoders with 16bit resolution.

Its impossible to attach on my AVR motion controller another 11 lines! To do that would mean to use all 4 R/C servo motor lines and all 8 potetiometer lines! But there are some other ways to reduce the control and data lines to just 3 lines:

Option1: I could use the PCF8575C Remote 16-bit I/O expander for I2C-bus. The PCF8575C has 16 I/O that are controlled by I2C commands. So I could easy handle up to 6 HCTL-2016 chips with shared data lines!

Option2: Use a small AVR, perhaps an Attiny2313, to read the HCTL-2016 chips and send the 16-bit measures with serial signal to the master Atmega8535 of the AVR motion controller.

Now the really news! I asked their price on the local store and got as answer that the HCTL-2016 costs 50 euro each!! Ouch, its price is too high! Just consider that I must pay 150 euro alone for three such chips for a 3DOF platform! :-(

So I'm back to the start again searching of another solution... Perhaps a slave AVR to do the job... Unless anyone can suggest a cheap 16-bit quadrature decoder!


Regards, Thanos

Hi Thanos,

I was looking at the layout and comparing it with the partslists. I think i have found some missing parts in the partslist.



I can't find the 1n2 capacitators in the partslist, or am i overlooking something in the list.

Greetz RoVoR

@Rovor

Those capacitors are 1,2nF

regards

Hi Rene,

Thanks for that! It will simplify the optical encoders output enouph to be able to use an small AVR to do the counting of the pulses!

Regards, Thanos

Hi,

Another AVR motion controller board (1.5v) is ready!







There are some holes with no compoments on them near the TX-RX connectors. I placed them in case someone wants to simply solder some wire pieces to do the connections permantly instead of using molex.

Its ready to accept the dual H-bridge board on it!

Regards, Thanos

Hmm, I analysed the compoments I need to form an circuit to work like the HCTL-2016 and I discovered I need 10 chips that consist of flip-flop's, counters and latches (4-bit each)! Ouch! They are cheap but I will need a PCB in size of the AVR motion controller to place them all. And one such board will count 16-bit values of just one encoder! This propably justifies the high prize (50 euro) of an HCTL-2016...

Now I'm searching for ways to decrease the number of chips needed to 3 or 4 by using 8-bit or 16-bit compoments, if there are such.

Will see how it will end up... ;-)

Regards, Thanos

Hi all,

I've allmost finished my 1.5 board.
The etching worked out fine as did the soldering. I must say i was a little scared to start this project (this is my first project ever), but i'm pleased with the results.

Here some pictures.





I've ordered the parts for the Hbridge and want to start building as soon as they arrive. But i think i have to be a little patient and wait for Thanos to complete his work with the H-bridge.

This morning i went for two whiper motors, they have raised the prices for them, 35 euro's a piece. @#$#@ one month ago they were only 15 euros a piece, is my whole town started building a simulator or what...... :wink:

RoVoR

As always Rene comes to the rescue!!! Good find! I was so concentraded on the 16-bit resolution that I didin't bother to look on the bigger members of HTCL family! They are 32-bit but can also give any other lower resolution!

Now I must go and ask their prize here in Greece and get them if they are so low, or else I guess I have to arrange another order... ;-)

Thanks!
Regards, Thanos