Servo Motor Driver ( MOSFET )

Interact

I have been experimenting with replacing the guts in a servo with my own design. I have taken apart 4 or 5 servos and looked at the bridge-driver circuits used in these.· None of them use diode protection from back EMF. Now they are all micro servos and take a very small amount of current, but is this a sound design? In my own driver I would like to leave out the protection diodes since there really isn’t room for them on the board that must be stuffed into a very small case. In testing I ran it for an hour with alternating forward/reverse , stalled it for a long time etc.· Basically I tried everything I could think of to blow the mosfets, but I would rather know I’m doing an acceptable thing. Looking at the symbol for a mosfet there is a diode, is this a small protection diode?

Beau Schwabe

In an NMOS device, a reverse biased·diode is formed at both the 'Source' and the 'Drain' through the substrate.· The substrate "tap" or PTAP ( because it is positively doped) is an essential part of an NMOS device that helps guard·against latchup.· This connection is typically shared with the 'Source' terminal (negatively doped). A diode is formed where the positively doped region of the substrate abuts the negatively doped region of the Source and Drain.· Since the PTAP is shared with the Source, this diode becomes inactive (or shorted out), however on the 'Drain' terminal this is a different story.· The diode on the 'Drain' becomes active when·the·'Source' voltage exceeds·the 'Drain' voltage by whatever the BandGap voltage happens to be for that
·process.··Typically the BandGaap voltage ranges from .3V to .7V.
·



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 1/18/2008 4:31:08 PM GMT

Phil Pilgrim (PhiPi)

In theory, these substrate diodes provide a pathway to shunt motor currents when the drivers are turned off, thus limiting voltage spikes. However, their use in this fashion is usually discouraged, with the addition of external Schottky diodes being preferred. Here's a reference: roko.ca/articles/hbridge.php?page=3.

That said, you could check the datasheets for your MOSFETs to see what the current rating is for the body diodes. If they're ample enough to handle your motor's transients, it may be safe to use them that way. However, if you're looking for ironclad reassurance that omitting the external diodes is okay, you won't get it from me.

-Phil

Beau Schwabe

Phil,

Agreed... Notice that the substrate is resistive...this makes for a 'weak' diode. Unless the diode is specifically designed into the component, then I would use an external alternative.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Interact

First, thank you so much for responding to my question. I know it's a bit of a stretch to go against proper design rules just because it works, but hey, it seems to be standard procedure for these tiny (production) servos to be run this way.

Current rating for the body diode I am using is 8A for the P's and 7A for the N's I guess that’s plenty for these little micro motors.
Question: are transistors less prone to damage than the mosfet ( given simi equal ratings ) ?
At least two of the servos I have dissected use what I detect to be transistors. (although telling what a device is when your test lead is twice the size of a pin is quite a challenge).

Phil Pilgrim (PhiPi)

Coincidentally, here's another current thread on the same topic: http://forums.parallax.com/showthread.php?p=702838.

-Phil