Electronics > Projects, Designs, and Technical Stuff

Projects to find the limits of what is possible right now?

<< < (8/10) > >>

TrickyNekro:
Now I will actually be doing some shameless self promotion but hey, it´s not often I get that excited about something I try to build.

Now by any means, this project is not going to fly us to Mars but it sure breaks two major headaches people had with hobby servos.
I designed a board able to replace the standard PCB in hobby servos, it adds feedback, comms over CAN and a truly "HV" option,
being able to run the servo off 4S LiPo packs directly. Still in the development but I recently made a small BOM for the thing.

For 10 pieces the materials would cost me around 18€ tops, probably around 16€, boards included and of course the outrageous 24% Greek VAT.
So say you buy an MG995 for around 20€ you would have a nice digital servo, probably compatible with CANOpen... Also open-source (still non commercial though),
so you can try to build your own program around it. And I suppose if I ever wanted to start manufacturing it, it wouldn´t cost more (lower prices for high volume parts and manufacturing included).
Try to find a servo that comes for that cost and it´s digital with feedback. Hardly any...

The funny thing is that all these are just a byproduct, it´s been years I wanted to create something like that just for fun, but hey! It´s progressing!

Nominal Animal:

--- Quote from: TrickyNekro on January 25, 2019, 03:52:03 pm ---I designed a board able to replace the standard PCB in hobby servos, it adds feedback, comms over CAN and a truly "HV" option,
being able to run the servo off 4S LiPo packs directly.
--- End quote ---
I've looked at OpenServos, as the analog nature and timing-based standard servo control seems .. complicated? to me.  Sure, using a time-based control signal makes the servo itself simple, but it does so by shifting all the complexity to the controller.  I mean, for most use cases, telling the servo the desired state and getting back the actual state in digital form would be so much more useful.

:-/O Something like a SPI but with a shift register, so you could chain any number of servos together, and control them with a single SPI bus, would be really nice.  Sure, others do need acceleration profiles and whatnot.. but a simple serial shift register control with state feedback would be lovely.  Hmm.. Is there a cheap Padauk chip suitable for this?  ;)

About the "HV" option: Do you use the same ground for both digital I/O and motor?

I have a SmoothieBoard (an LPC1769 with four A4988 stepper controllers, a couple of MOSFETs, and various I/O), but the way they used the same ground plane for both digital logic and motors makes it a pain in the butt in my opinion. (If you connect it via USB, your computer ground is now the motor ground. That isn't good.  Yeah, I can isolate the USB 1.1, but why would you design it like that?)

It matters less with hobby servos -- definitely so for anything battery-powered! --, so perhaps I'm just obsessed over keeping the digital logic isolated from the noisy motor world and avoiding ground loops that so easily burn out my precious little toys, though.

coppercone2:
This thread is proof engineering work drives people crazy. Discuss your projects

Discusses eating rat organs. :-\

TrickyNekro:

--- Quote from: Nominal Animal on January 25, 2019, 04:54:10 pm ---
--- Quote from: TrickyNekro on January 25, 2019, 03:52:03 pm ---I designed a board able to replace the standard PCB in hobby servos, it adds feedback, comms over CAN and a truly "HV" option,
being able to run the servo off 4S LiPo packs directly.
--- End quote ---
I've looked at OpenServos, as the analog nature and timing-based standard servo control seems .. complicated? to me.  Sure, using a time-based control signal makes the servo itself simple, but it does so by shifting all the complexity to the controller.  I mean, for most use cases, telling the servo the desired state and getting back the actual state in digital form would be so much more useful.

--- End quote ---

Nah, they use PPM and then something like a de-serializer at the receiver, and for analog servos they use op-amps in a PID like configuration (if not only P), which you can totally achieve with a single op-amp
some resistors and some caps, plus the drive stage.


--- Quote from: Nominal Animal on January 25, 2019, 04:54:10 pm ---About the "HV" option: Do you use the same ground for both digital I/O and motor?

--- End quote ---

In the hobby market "HV" means around 7.8 - 8.something... In my design I go absolutely max at 18V.

Well about the ground thing... Yes it uses the same ground for everything... BUT and that´s the beauty of CAN Bus!
CAN Bus does not need ground to operate, cause it´s a differential signal. You can have a separate "power ground" and a separate "signal ground".

If you want to get extra freaky for whatever reasons there are also isolated CAN Bus drivers, but that should be on the controller side, it´s plain stupid to put it in the servos.
Plus by no means I had the space for something like this. That´s something you do on controller side.

TrickyNekro:

--- Quote from: coppercone2 on January 25, 2019, 05:01:07 pm ---This thread is proof engineering work drives people crazy. Discuss your projects

Discusses eating rat organs. :-\

--- End quote ---

Bruh! I got this hanging out of my door... I came prepared bruh!  :-DD :-DD :-DD




Edit: As a fellow RA2 fan I expect understanding and compassion in this matter. :-P

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod