bec diy esc, how much current is needed?

[plane_spotter]

I have made a few of Mike Norton's esc circuits with pleasing results. I have made the unit now completely smt apart from the voltage regulator. I now that it has to be Low dropout. Is it necessary that it supplies 1A. I now it depends on overall load, servo quantity, reciever, no of servos. Has anyone got some rough figures for currents of various items. I would like to replaye the to220 reg for a smt one to keep up with everything else. Up to now I can find one in d2pak config giving 800ma max

any feedback would be great.

cheers

[Comatose]

d2pak isn't really that great a package for our uses. Dpak is much smaller and lighter, and has pretty similar power handling abilities. D2pak is great if you have a couple square inches of board space to devote to heatsinking. On a minimal pad they're quite similar.


I'd look at the NCP1117DT50 from On Semiconductor. Budgetary price is 37 cents, Dpak, 1A, 20V max input, LDO

Its also availiable in SOT-223 if you want to be more thermally brave.

[vintage1]

I clocked some HS55's at 250mA apiece when moving. Maybe more under stall or aerodynamic load.

[AndyKunz]

A stalled servo, even a little one, can pull 1A.

I use TO-220 format on the ESCs I manufacture with BEC. Put a decent cap on the output to the receiver.

Andy

[Bruce Abbott]

Here's some measurements I made of various servos:-

===================
Servo Current Tests
===================

Tested with GWS servo tester, analog ammeter.

GWS Pico (STD,F-BB) 200mA
Naro (STD,BB) 200mA
Naro HP BB 250mA
Naro Pro BB 150mA
Micro STD 120~150mA
Micro2BB-MG 150~300mA
Mini L BB 100~400mA (1.5A peak)

Hitec
HS-55 150mA
HS-81 250~350mA

Multiplex
Super FL Digi 100~300mA

Futaba
S3101 150~200mA (400mA stall)
FP-S143 100~200mA (1A peak)


250mA per servo looks like a safe value to suit most popular types. Receiver current is usually only about 20~30mA.

A TO220 regulator with no extra heatsinking can dissipate about 2.5W before going into thermal shutdown. For 8 cells this corresponds to an average current of about 550mA. Luckily, most models don't need to have all servos moving continuously, so this would probably be good enough for 4 servos.

The thermal resistance of smd packages is much higher, so the pcb must be used to conduct away heat. It may also be possible to use the battery wires to provide heatsinking, provided they aren't getting hot from motor current! A disadvantage of using the pcb as a heatsink is that the switching MOSFETs are also adversly affected by high temperatures.

Two commercial solutions that might be worth copying:-

Kontronics Rondo 400:- Uses a TO220 regulator mounted flush to the pcb, but with the heatsink facing outwards (low profile and keeps heat off the pcb).

Castle Creations Pheonix 45:- has a separate pcb with 2 smd regulators wired in parallel. The entire pcb is copper clad for better heatsinking.

For my next design I will be using the L4941BDT, which is a 1 Amp LDO regulator in TO252 packaging. It is available from Mouser Electronics (http://www.mouser.com)

[jojonz]

Thank you Bruce for the servo test.
First time I have found such data.
Does the GWS servo tester place a defined load on the servo arm while testing,or is the data no load?

[Bruce Abbott]

Except for the stalled servo, all the tests were done without mechanical loading. To get the highest reading I wiggled the servo tester dial back and forth as fast as possible, causing the servo to reverse direction continuously. This actually draws more current than when it is moving from one extreme to the other.

Peak readings were taken with an oscilloscope. I only did this for a couple of servos that were causing very rapid meter movements.

[jojonz]

Thanks for that.

[Bill Glover]

No load figures are interesting, but IMO you should take the stall current to be on the safe side.

[vintage1]

Current isn't the problem. Heat ultimately is. A stalled servo will very quickly overheat any BEC.

Its easy enough to do 2A peak, but on say - 12v motor pack, dropping to 5v, thats 7x2=14W of heat to get rid of. Not at all easy without a substantial heastink and some decent cold air around it.

What we really want is servos that work on - say - 12v.....