How come nobody builds their own ESCs?

[jstapels]

I was curious why it seems that everyone uses the off-the-shelf ESCs rather than build their own version? As far as I can tell, the standard Turnigy Plush ESCs seem like complete overkill (in terms of features) for a quadcopter setup using MultiWii.

Specifically, it looks like all one would really need is a MOSFET driver that can handle the proper amperage rating of the motors used. It seems that this would offer not only a cheaper solution, but also more control over the motor itself since the MultiWii software could use PWM as fast as it wants to.

ESCs appear to be designed to talk to the receiver which isn't done when using something like MultiWii. In fact, most of the ESC extra functionality appears to just get in the way. MultiWii has to fake the start-up sequence, the battery protection is dangerous since it just cuts out (in LiPo mode), and the brake function appears to be disabled by default.

That said, I'm a software engineer and only pretend to know hardware . I'm sure I'm missing something, which is why I pose the question.

[mr.rc-cam]

Specifically, it looks like all one would really need is a MOSFET driver that can handle the proper amperage rating of the motors used.

It would be as simple as that if we were using brushed motors. However, there's a lot more going on in a 3-phase brushless ESC and the typical MWC's CPU does not have the spare resources.

[Rob]

Building your own esc is not lowcost.
This is my own Nfet ESC using good components.
Buying in china is cheaper.
The intension of Multiwii is to build a cheap good working quadrocopter.
If I make 1000 esc's it's still not lowcost.
http://qynx.nl/index.php?title=Motor_Regelaar

[copterrichie]

Building your own esc is not lowcost.
This is my own Nfet ESC using good components.
Buying in china is cheaper.
The intension of Multiwii is to build a cheap good working quadrocopter.
If I make 1000 esc's it's still not lowcost.
http://qynx.nl/index.php?title=Motor_Regelaar

What about Quality? Does that account for anything these days or is Cheap the only way to go?

[jstapels]

Building your own esc is not lowcost.
This is my own Nfet ESC using good components.
Buying in china is cheaper.
The intension of Multiwii is to build a cheap good working quadrocopter.
If I make 1000 esc's it's still not lowcost.
http://qynx.nl/index.php?title=Motor_Regelaar

I guess I'll have to take your word for it as you didn't really explain why it's so expensive. My somewhat naive assumption was that an ESC was essentially three half-h bridges, meaning it's a microcontroller hooked into six high-power MOSFETs. I know how to program AVR micros and I figured I could pick up six relatively inexpensive FETs from a place like digikey.

Naturally there's a lot I don't understand, but a part like the NTD5867NL can handle 20A and runs around 60 cents a piece. Generally the passive components (like resisters, caps, and diodes) are a dime a dozen. So unless there's something obvious I'm missing, I don't see how one of these wouldn't cost more than, say, $15 to put together.

Maybe in the end there wouldn't be any advantage but it seems to me that making an ESC designed for a quad would allow some nice options such as running it via I2C and being able to get speed data back from the motor.

[Rob]

I only answered Your question........ There are people who build their own ESC.
But most people choose to reflash the $8,- ESC instead of making one of $15,-

I like a good quality, thats why I choose to make my own ESC.

I have test a lot of ESC'S ......... from lowcost to highcost

And if a $8,- good produced ESC is flashed with the right software there is no problem, thats why not
everybody building their own ESC.

And not everybody can make their own ESC........

Rob

[jstapels]

You did answer my question, so thank you for that. If a decent $8 ESC can be turned into a good ESC with a quick reflash, it is pretty hard to justify rolling a DIY. That said, could you recommend a good solid ESC? In the build logs I've been following it looks like Turnigy Plush and Hobbywing are pretty popular. Are custom firmwares available for both those models or is there a better choice?

[copterrichie]

I have only seen one attempt at making a multiple ESC in the same package. Present ESCs are designed for a single engine aircraft, with multirotor copters, we are talking 2,4,6,8 and now more. That is a lot of ESCs on one aircraft. Seems to me, a good niche would be an ESC with four or eight ESCs in one package.

[Rob]

I thinking 4 ESC on 1 board is for production ok, for Octo 2 boards is easyer than 8 ESC's.
And I preffere only N fet's.

I design a PCB with 4 of my ESC's

Rob

[jstapels]

I thinking 4 ESC on 1 board is for production ok, for Octo 2 boards is easyer than 8 ESC's.
And I preffere only N fet's.

I design a PCB with 4 of my ESC's

Rob

I would be more than happy to help in this effort. Rob, I looked around on the wiki you linked earlier but I couldn't find any PCBs or BoM for your design. Would you be willing to start a thread and allow others to help you out. I'm not much of an EE but I do how to program Atmel's and would be more than happy to offer some assistance.

It might also make sense to start with a dual ESC before a quad ESC. That way it offers a little more flexibility for those who want hex copters, plus, more importantly, it should help with the heat dissipation. That many FET's all in one place could be a serious heat issue. My other thought is that while today's micro are pretty darn powerful, the closer the PWM pattern is to an AC signal the smoother the motor control should be, so controlling 12 FETs might make things a little easier assuming we're going with a single micro to save money/space.

I haven't dug through the MultiWii code but would it make sense to have the ESCs use I2C for easier control as well as using the back EMF to send speed data back to MultiWii? This kind of information might be useful in allowing MultiWii to be more stable as it would now only know how fast the motors should go, but how close they are to getting there.

[copterrichie]

Here is a very interesting power distribution board, something along this line with one I2C input would be idea in my opinion. The MWC does not support I2C ESC but it would not be a difficult modification.

http://www.rcgroups.com/forums/showpost.php?p=21116841&postcount=51

[jstapels]

Here is a very interesting power distribution board, something along this line with one I2C input would be idea in my opinion. The MWC does not support I2C ESC but it would not be a difficult modification. [/url]

That's a cool looking board but it would cost a lot more to manufacture that PCB over a square one. Most fabs charge per sq in, and that area in the middle is a bit of a waste. For a first go on this it would probably be best to stick to something more squarish.

As for updating MWC to support I2C, I know it already has support for reading the sensors on I2C, so all the libraries are already there. It's just a matter of telling the motor control code to use that. Another advantage to using I2C is that it would allow hex/octo copters to be able to use a Pro Mini instead of upgrading to a Mega.

That said, I2C is slower than SPI (but I don't know how much). I would imagine I2C is still plenty fast, but in the event that it's not, we could also consider using SPI instead.

[copterrichie]

Yes, SPI is faster but interfacing with the MWC controller would be a problem right now. Not saying it could not be done, just some rewrite. I agree that power distribution in its present form is a waste, BUT!!