dsPic-Servo

[slp_prlzys]

It's been a while since my last post.

Here, as if there's not enough projects here to keep some of us busy.
I found this http://members.shaw.ca/swstuff/dspic-servo.html
haven't tried it yet tho, doing something else for the moment.

If any1 ever build this please post result.

Similar Threads:

• Servo Dynamics, West Amp, Servo driver, Anilam, 1400 relay board 901-175
• couple of sets of Indramat Servo Controller+Mains filter+Servo Motor 9Nm, 4500rpm
• Galil Motion Controller, with NEMA 23 Servo Motors, Servo Drive, Encoders, Etc.
• first steps with dsPIC + IRAMS AC servodrive

[tfmacz]

It's been a while since my last post.

Here, as if there's not enough projects here to keep some of us busy.
I found this http://members.shaw.ca/swstuff/dspic-servo.html
haven't tried it yet tho, doing something else for the moment.

If any1 ever build this please post result.

I have 2 of these running the X and Y on my cnc router.

[fikretduru]

thanksss

[tfmacz]

There is some further development on this project under way.

I have a working prototype of a higher current version that uses two op-amps in parallel to increase the current capacity of the drive.

I have some larger motors that I would like to use that require more current than a single op-amp can provide.

Also, I am building a version that uses two op-amps as an H-Bridge to give me higher voltage capability. Hope to have a working prototype in a few days...

There is a U-Tube video of the servo drive in action.

[ame]http://www.youtube.com/watch?v=RJEXbEXn0CQ

Cheers...
Ted

[babinda01]

Hi Ted,
I have been looking over this project and am quite interested. Do you feel there is much work required to get this to work with a HBridge????
Regards
Andrew

[tfmacz]

The present design relies on the op-amp power stage to look after current limiting. The dsPic controller outputs a pwm signal which is filtered to a voltage proportional to the required output current.

The new design uses 2 op-amps as an h-bridge. The design should be capable of up to 60 volts and 7 amps, which are the op-amp specs.

I have not fully tested the new design but initial test look good.
It runs the servo and didn't let the smoke out of anything.

[smarbaga]

hello
perhaps check out some of these op amps
http://www.ife.tugraz.at/datashts/nsc/h8710.pdf
http://www.ortodoxism.ro/datasheets/...wn/mXuussw.pdf
(opa447,548,549)
perhaps a few lm3886 gain clones ($6.00 each)
_________________________________________
e_mail recieved about lm3886
Cool, glad you liked it!

Yes I think the National chips would make great servo drivers. Not that the
LM1875 I had in the drawer ever got used for that. Just treat them as an
ordinary opamp and as long as you run them with a high enough gain (above
10 to be safe) and good power supply decoupling they should work as
expected without stability problems.

One thing I noticed though is they don't like a lot of reverse current
(regenerative braking), as this places maximum stress on the output devices
(I think). I think they are protected to an extent but it would pay to run them
well below their power limit and with really good heatsinking, or at least drive
one hard and see if it survives whatever mode of operation you want to throw
at them. A really well protected chip will run without a heatsink, it'll just cut in
and out.

Best of luck!

- Antony.

--------------------------------
From: "smarbaga"
To:
reguards garry

[waleed alhadidi]

A PIC 30F4012 is used in this controller
Can I use an alternative PIC instead of this one ? like the 16FXXX to do the same job ?

Thanks

[tfmacz]

The 30F4012 has built in provision for the quadrature input and hardware counter for the input from the encoder.

the 16F pics do not have this.

[epineh]

The 30F4012 has built in provision for the quadrature input and hardware counter for the input from the encoder.

the 16F pics do not have this.

How about the dsPIC30F2010 ?

I am interested in trying this and using the PWM drive signal and direction signal, converting this using logic into PWM and NOT PWM and feeding this straight into a discrete H-Bridge.

Any thoughts would be appreciated.

Nice work so far btw.

Russell.

[ladzhari]

I'm (slowly) working to convert the output routine to h-bridge operation compatible with a hip4080/81 bridge driver so I can make some with only junk box parts.

The way it's put together should make it really simple to convert, it already does the PWM by itself so it's just a matter of getting a "direction" signal out of the PIC, coming up with a way to limit the current, and making sure that the deadband operation is OK. I think all of this stuff is in a single function in the code.

kudos to Lawrence/Tom for putting together and sharing such a well documented and great project.

Also:
I think at some point I read that someone else was doing the exact same thing, but I lost track of it?

[epineh]

I would be interested in how you go, I have put together a board, and logic to use the output from the DSPIC to drive a discrete bridge, but I think the PWM output needs to be inverted, as 4.75 V is zero current and 0V is full current from the original Op Amp configuration, I got distracted and have not been back to try again. I also couldn't get any serial comms so there were other issues with my board, I don't think I set the clock configuration correctly.

I have a friend that used the DSPIC servo with a different op amp (smaller) to drive +- 10V industrial drives and it works very well, just have to get it working with discrete's.

Russell.

[ladzhari]

It's been a while since I looked at the thing after moving my workbench but IIRC there was just one function that set the duty cycle of the PWM based on the error calculated by the PID loop. I think that floating at 2.5 was the "zero error" (Motor is in the correct spot) place and nearly full-off and full-on was the max and minimum error. Changing it to whatever we want should just be a matter of modifying the one function.

The project already has the hard bits done, the PID math and serial-port tuning code.

I'm aiming for something with similar specs to the Gecko g320/340. It could be that they are using these same HIP4080 drivers since the maximum voltage is 80V, which is the same as the HIP parts.

Also, I've got already from a previous project a slick single-sided PCB foil layout for an H-bridge that should be pretty hobbyist friendly to construct.

[H500]

Why not use the pwm signal to drive a discrete mosfet power stage directly, instead of the linear op amp? The efficiency, current and voltage capability would be so much higher.

[epineh]

It's been a while since I looked at the thing after moving my workbench but IIRC there was just one function that set the duty cycle of the PWM based on the error calculated by the PID loop. I think that floating at 2.5 was the "zero error" (Motor is in the correct spot) place and nearly full-off and full-on was the max and minimum error. Changing it to whatever we want should just be a matter of modifying the one function.

From what I understand the output is sign magnitude, from the schematic "RE2 sets direction and PWM on RE0 sets current" so this almost all we need to drive a H Bridge directly, thinking of it it may be as simple as putting the PWM signal through an inverter, say a 74HC14 to swap the max/min value and voila ! I can try this on my board once I get back to it, maybe this weekend.

Aparrantly sign/magnitude is a little better than locked anti-phase for driving a full bridge, but I don't know enough about it to say much more than that.

I have a friend who mentioned a line of the code that actually inverts the output PWM and he thinks commenting this out will do the same thing, even better...no hardware change

Why not use the pwm signal to drive a discrete mosfet power stage directly, instead of the linear op amp? The efficiency, current and voltage capability would be so much higher.

That is exactly what we are trying to do, I mentioned the small op amp configuration used to drive industrial amps just to point out how effective the drive functions are.

Russell.

[H500]

I experimented with an ir2104 based h-bridge servo some time ago. The DIR signal was AND'ed with the pwm signal and droved the left half bridge. The inverted DIR was AND'ed with the pwm signal and drove the right half.

The same approach should worked with this servo also. As you mentioned, the pwm signal need to be inverted either by software or hardware.

[tecfacet]

i dont know alot about dspics but i am very good with the other pics.

i also can help with pcb's and such. i am well versed with altium.

please keep me informed.

roger

[barum]

tecfacet, where to find dspic lib for altium?

[tecfacet]

i just make the library parts as i need them. in some ways, altium is a pain but .... once one dives into it, the process is not too bad.

roger

[barum]

Firmware is so huge. Maybe removing all those printf calls can make them fit in smaller Motor Controller dsPIC like 2010. I dont have dsPIC30F4012 samples so ill try to compact hex size first.