next topic »

[on:]

Hi all,

I have the following robot and I notice that the servo rotates smoothly with no issues if I power the arduino from the usb cable but it's much slower and quite jerky when I use my power circuit.

Robot -

1 9.6v 1800mAH battery pack connected to an  Adafruit motor shield which drives 4 x 4.5v motors on a Sainsmart 4WD chassis
http://www.adafruit.com/products/81
http://www.sainsmart.com/sainsmart-4wd-drive-aluminum-mobile-robot-platform-for-robot-arduino-uno-mega2560-r3-duemilanove-silver.html

1 9.6v 1800mAH battery pack connected to
    Arduino via barrel jack
    Raspberry pi via 5v BEC to usb port (3A UBEC 5V Max 5A 3 Amp Lowest Switch-Mode RF - datasheet unavailable)
    5v supply for Servo via BEC
    I also have a switch and led in this circuit to show power

The Arduino is powering -

3 home made passive IR sensors ( based on http://www.societyofrobots.com/schematics_infraredemitdet.shtml)
1 HC-SR04 ultrasonic sensor, servo mounted
1 sharp ir sensor, servo mounted

The arduino and pi communicate via serial run through an adafruit voltage converter - https://www.adafruit.com/products/395

The servo signal pin is also connected to the Arduino

Any help here would be greatly appreciated. I'm thinking I might need to decouple my power circuit or change out the BECs but I don't know where to start troubleshooting. It could even be bad battery packs

[Reply #1 on:]

My wild guess is that your battery pack isn't able to provide enough current to drive the stuff connected to it. Have you measured the voltage of the battery pack while under load? Remember the Raspberry Pi uses almost 1A at 5V. How efficient is your voltage regulator (or BEC how you call it)? Is this a linear or a switching type?

[Reply #2 on:]

In order to test the voltage under load I assume I should be measuring across the pins coming out of the pack itself instead of across individual circuits?

I don't know about the BEC to be honest, I got them for $4 each on ebay it takes + and - up to 25 volts with a 5v or 6v jumper and outputs the 5v and ground as well as a white wire, I assume for signal? They are generally used in model planes. Measured output is 4.98v on one and 5.27 on the other

If I were to buy proper voltage regulators should I look for linear or switching?

Thanks for the help

[Reply #3 on:]

If I were to buy proper voltage regulators should I look for linear or switching?

Switching regulators are much more efficient but the linear ones are easier to integrate (usually just one single chip, while the switching regulators need additional parts including inductors).

I would first try to just insert a not too small capacitor after the BEC (100µF and a 100nF from GND to 5V). That may be enough to flatten the current spikes and establishes a hopefully more constant voltage.

In order to test the voltage under load I assume I should be measuring across the pins coming out of the pack itself instead of across individual circuits?

Both. If it's the battery the voltage drops what's the minimum of your voltage regulator. The you probably also see voltage drops in your 5V area.

I guess you don't have a scope to see the voltage changes over time?

[Reply #4 on:]

I just have a pocket combimeter so it's pretty difficult to see changes over time. With your capacitor suggestions you are saying to have an electrolytic 100muF in series after the BEC and a ceramic 100nF linking the 5v output to ground before this, if I read correctly. Like I said I'm new to electronics. Could you explain why the circuit doesn't short out via the ceramic capacitor or maybe link to a circuit similar to what you are proposing?

Thanks again

[Reply #5 on:]

http://en.wikipedia.org/wiki/Decoupling_capacitor