powering a servo

Lee Marshall

hi, i need to power a micro-servo from a 13.5v DC line. i take it that the 13.5 would fry the servo.
what should i use to step down the voltage? a plain 7805? or something like a switched-mode reg(please provide an example part number).

Thanks in advance for any replies
Lee.

Leon

How much current?

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Zoot

A 7805 or 7806 (if you want 6 volts) with a moderately sized heat sink would be fine. I would use a 7805 to prolong the life of the servo(s).

I've run up to 5 standard-sized servos through a 7806 being fed ~13 volts w/o thermal shutdown.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
When the going gets weird, the weird turn pro. -- HST

1uffakind.com/robots/povBitMapBuilder.php
1uffakind.com/robots/resistorLadder.php

Lee Marshall

this is the problem...servos are apparently unpredictable when it comes to current.
from what ive read, a micro servo can draw, anywhere between 180 and 340mA operating under no load.
the servos i want to use(cheap esky EK1-500's) arent listed anywhere as far as i can tell.

Lee Marshall

Zoot, what type of current rating do you have on your 7805s?
would 500mA cover it?


EDIT: referred to the wrong guy(said leon, meant zoot).

Post Edited (Lee Marshall) : 10/4/2008 12:07:42 AM GMT

Zoot

7805s in a TO-220 case are rated to handle 1 amp running, much more peak -- with decent heat sinking depending on current draw and voltage in. Remember it's voltage drop (regulated) * current = watts, so a servo pulling 500ma through a regulator that drops 13.4v to 5v will generate approx. 4.1 watts -- pretty warm. It's not ideal, of course, because all that heat is really wasted power that has to come from somewhere, but it certainly works, and is simple and practically indestructible with a decent heatsink.

That said, the best thing to do really is just measure the current draw of the servo -- then you'll know. In theory my standard servos draw up to 1100ma each from stall, but when I measure all 5 starting at once, as a total set they were drawing less than 1.75amps, and less than half that running.

You can also get regulators in TO-3 cases which require a larger heatsink and can handle much more wattage, but at that point a switching regulator might be in order.

You can get integrated switching regulators with a 780X series pinout (single board, ready to use) for about $15, e.g., search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=PT78HT205V-ND

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
When the going gets weird, the weird turn pro. -- HST

1uffakind.com/robots/povBitMapBuilder.php
1uffakind.com/robots/resistorLadder.php

Lee Marshall

thanks for the replies.
i have built and PCBd a prototype circuit, and it is currently working. the firmware is incomplete, but i have 2 servos moving to a preset position.
the actual application is a set of miniature 'robot' lights which use high intensity LEDs(REALLY HIGH) to project onto a semi-transparent silk-type material suspended above the front garden in the dark.
the effect being "ghosts" flying around. this halloween is gonna be awesome.

I'm going to use a computer to send commands to each light unit(distributed around the front of the house) via a 19200baud one-way serial link in a multi-drop bus setup. the framerate will be 20Hz.(each unit receives 20 commands a second)
the problem is that each light unit needs power.

i measured the current of the servos, and it hits about 250ma when moving.

would there be a problem with using a 3-core cable with 13.5VDC, GND, and SIGNAL, as the "bus" for the light units?
i am worried that the fluctuations in current would smoosh my signal up if if they are in the same flex.
the signal is to be 0-5V to be fed directly into each microcontroller's UART. could i just use a max232 as a driver for all of the unit's UARTs(there will be about 5 light units in all)?

Lee Marshall

ok, ive done some reading on this stuff and it seems like the cheapest thing to do to avoid interference would be a twisted pair cable with opto-isolation on the receiver-ends.

any thoughts?

Post Edited (Lee Marshall) : 10/8/2008 1:01:44 AM GMT