OpenServo.com

[robert.r]

Dear community,

for some time i'm making experiments with an Servo, but not getting any stable result.

I am using a Futaba-Servo with PCB replaced to OpenServo V2, bought from watterott (sorry, but I can't type link on my first post).

The Servo is powered by a 12V and 3A PSU and connected to an OSIF. Since there is an 5V powerlane in OpenServo V2 PCB the AVR is powered by USB from OSIF, if connected. Otherwise the AVR is getting about 4.3V from 12V lane through integrated voltageregulator.

So now the Problem i can't understand:

When the motor is powered up or suddenly changing it's direction i see a spike on the 5V lane, breaking down to about 2V (seen on an old analogue oscilloscope) causing a brown out.
Of course powering up a motor leads to a spike on 12V lane; and i could understand, that the voltageregulator wouldn't supply 5V to the AVR if it's input voltage falls below ~6V. But even when the 5V lane is powered by the USB from OSIF there is a breakdown in Voltage on this lane.

How does that breakdown on 12V lane cause the same effect on the 5V lane? The only connections in between are

the common ground,
the (diode protected) voltageregulator and
the gate-capacity within the H-Bridge.

A big capacitor (100uF) at 5V lane didn't solve that problem.

Has anyone an explaination for that behavior?

Thanks in advance

[Cliff]

Hi robert.r,

Try checking your common ground for excessive IR drop: Scope ground at common point of power supply and OSIF and probe ground on the OSv2, during motor current spike.

Small and/or long ground wire and/or poor connector contacts could be the cause.

Cliff

[robert.r]

Hi Cliff,

thank you for your reply. I've done some measurements.

During motorspike I noticed a voltagedrop of about 0.2V between OSIF and servoboard on common groundline. This doesn't look critical for me. My multimeter shows me a resistance of about 0.5Ohm for this connection.

For experiments i removed the protection diode between 12V-5V regulator and 5V lane, leaving them unconnected. So now the only connections between 12V and 5V are the gates of the MOSFETs.

But still, on full-acceleration or inversion the voltage on 5V lane drops down to 4.7V (on my slow multimeter), causing brown-out.

I just can't get it, why a drop on 12V lane causes a drop on the 5V lane - even now, when they are independent except of common ground.

[robert.r]

Problem solved:

I soldered a 47uF Tantalum capacitor directly across the Vcc and GND Pins of the microcontroller. (This was a hint from a experienced fellow.)

From now on there are no more brown-outs

[BasicFox]

a capacitor across the motor leads is also an extra protection.
See the construction tutorial of Jared:
http://www.openservo.com/ConstructionTutorial_MG995#head-06085380728a53ad1be7c8266fc889ffdfc743a5
_________________
Do try this at home!