To minimise unreliability I’m supplying my circuit from a regulated external 5V power supply. At present this supplies the entire circuit. But should I use it only for the servo 5V/0V line, with another supply (such as the PC USB port, or a 9V supply)? The maximum current now, when the servo is pressing a heavy button for 3 seconds, is about 180 mA.
You can use one power supply as long as it can supply the total current. Sometimes motors can generate noise on the power supply line but it's rare that you'd need completely separate power supplies.
The same 5v (at say >= 1A) power supply should be able to power both the servo and the Arduino.
What Arduino do you have ?
Avoid using the external 5v supply and USB (FTDI) power at the same time.
A properly regulated 5 V supply will be able to power everything.
Note the need to connect power (and ground) separately from the output of the regulated supply to each part, the servos that use more current and the "5V" pin for the logic section.
Then the control wires from the logic must track back as a bundle with the power wires to the power supply and from there as a bundle with the power wires to the servos so that there are no open loops in the wiring.
I would ( and do for my servo projects ) use a single supply but I add some good decoupling between the motor powering parts and the processor and sensor parts.
In particular I use a Pi circuit like the last one shown on this page.
http://www.thebox.myzen.co.uk/Tutorial/De-coupling.html
Thanks all. I’ll try applying those recommendations in the morning.
The only part I’m having trouble getting my head around is this:
Then the control wires from the logic must track back as a bundle with the power wires to the power supply and from there as a bundle with the power wires to the servos so that there are no open loops in the wiring.
In particular I’m not sure what an ‘open loop’ is?
I’ll draw or photograph my single servo layout tomorrow If that helps? Actually, in case it matters for this discussion, I have two layouts: one is using the Arduino Uno plus breadboard for experiment and the other (with an identical sketch) I’ve transferred to veroboard with its ATMega328 and placed in a small case.
Also note that the servo connections (5V, 0V and output pin 9) are connected by three wires about a metre long for the breadboard version, and 3 m for the finished veroboard version.
What is the servo's STALL current? What is the size (gauge) of that 3m of wire? What is the current rating of the 5V supply?
Also note that the servo connections (5V, 0V and output pin 9) are connected by three wires about a metre long for the breadboard version, and 3 m for the finished veroboard version.
That is a long way. You need at least some decoupling capacitors at each end of these wires, and possibly something like a 100R series resistor in the signal line to try and cut down on reflections on the line. This long wire will act as an antenna radiating any interference kicked up by the motors.
JCA34F:
What is the servo's STALL current? What is the size (gauge) of that 3m of wire? What is the current rating of the 5V supply?
JCA34F:
What is the servo's STALL current?
It's a Hitec HS-311, whose stall current is rated as 800mA.
(Its 'No load operating current' is 180mA.)
What is the size (gauge) of that 3m of wire?
It's 'standard' ribbon cable which looks like (and I assume is) 26 AWG.
(It's 'No Load Operating Current Draw' is 180mA.)
What is the current rating of the 5V supply?
4A
I read that 26 AWG has a resistance of 134mohm/m so the total for the 2 x 3m = 6m would be about 0.8 ohms, which implies a loss of about 0.4V during a stall. I don't think the servo is stalling but I'll get back to the shed and check this morning. Assuming I'm right, and the max current through the servo is 180mA (measured at my supply unit as "0.18A", leaving aside the noise issue for the moment, does that all look OK?
Terrypin:
To minimise unreliability I’m supplying my circuit from a regulated external 5V power supply. At present this supplies the entire circuit. But should I use it only for the servo 5V/0V line, with another supply (such as the PC USB port, or a 9V supply)? The maximum current now, when the servo is pressing a heavy button for 3 seconds, is about 180 mA.
I very much doubt the peak current is only 0.18A - budget at least 1A per servo. 95% of servo
problems seem to have is inadequate power - the high current is drawn in narrow spikes that simply
won't be visible on a multimeter at all.
Use a separate supply to the Arduino 5V too, the shortest of spikes on the supplycan cause havoc to
logic circuitry, or even cause damage - not worth the risk...
Servos are designed for 6V to 7.2V typically anyway, sometimes 4.8V, so 5V isn't necessarily going to
work reliably for the servo anyhow.
If you have a 'scope and can check the cleanliness of the supply rail you are in a position to decide if
the servo supply is suitable for sharing with logic - if not you are guessing/hoping it will be OK, and
then the best approach is using substantial decoupling (470uF or more perhaps).
One way to isolate logic supply from main power supply is with a buck converter - 7.2V servo supply also
drives a buck converter to give 5V for the Arduino - then the converter acts to filter spikes and
drop-outs to a certain extent.
Grumpy_Mike:
That is a long way. You need at least some decoupling capacitors at each end of these wires, and possibly something like a 100R series resistor in the signal line to try and cut down on reflections on the line. This long wire will act as an antenna radiating any interference kicked up by the motors.
Thanks. I was hoping to avoid any extra components at the servo end. It's outside my shed pressing the button on a CUBE camera, reasonably secure from the weather under a table. I didn't locate the circuit itself there too because bench-testing etc would then have been impossible. Instead I'm bringing the servo and camera into the shed workshop for testing and code editing; the three servo wires are then about 1m from the supply and circuitry.
Screenshots for background:
(The latter was a photo taken when the servo was being tested indoors.)
Thanks @MarkT, I’ll follow up those suggestions too.