[Solved] Question about 2nd power source and servos

[Redcap]

Hello everyone,

Background
I'm planning to use two SM-S4303R Continuous Rotation Servos controlled by an Arduino uno.

At the moment, I have it hooked to the laptop via usb and the coding and motors work fine, but I wish to have it run freely.

I know a standard 9volt battery is enough to power the Arduino, but won't supply enough amps for the servos and as such, I'll need another source.
I'm going to source/buy a 6v NiMH battery pack to power the servos as my servos run from 4.8v to 6v, but the data sheet doesn't show how many amps they draw each.


Questions
How do I find out how many amps they are (each) drawing? I do have a multimeter, but have only used it for testing DC voltages and resistor values.

What is the best way to connect the servos to a separate power source?
I know I'll need to connect the negative/ground wires of the power sources together, but should I connect the actual servos in series or parallel?

I was thinking of connecting them in parallel so that a full 6volts (at max charge) will be delivered to each servo.
If doing so will still supply the servos with enough amps, are there any If's or But's I should know about?


Any help you wondrous people could lend me will be greatly appreciated

[garrettg84]

The most common current number being slung around the forums is half an amp. You should be able to find specs on the servos online though provided they aren't cheap knockoffs. As for good power sources, NiMH is good but heavy compared to something like a LiPo battery.

Definitely connect the servos in parallel. They should all sit on a common power and ground rail.

[Redcap]

Ahh, much thanks, Garrett
So, probably look for a battery pack that can supply at least 1 amp (or two battery packs if lower amperage)

[rbtying]

Just about any AAA-size or large NiMH pack will be able to source >1A: look for the ones designed for RC usage, they're designed for running servos.

[vinniewryan]

I've got a 3 pack of 700Ma AAA NiMH cells running with a draw of 1.1A. It gets warm, but never hot. Most batteries can handle a discharge rate of 1.5*C with no problem (with a few exceptions).

Check out this page, it has a lot of information regarding battery discharge and such.
http://batteryuniversity.com/print-partone-16.htm

Cheers!

[Redcap]

Thanks rbtying & vinniewryan.

I was thinking about this battery pack, as I can get it from a local supplier, and adding a 5v limiter on it so I don't fry the servos.

I (quickly) read over the page/link you sent, vinnie, and from what I understand, it means this battery should be able to run at a 1C discharge for a little over half an hour?
(Will be sure to re-read at a normal pace when I have more time on hand).

[rbtying]

You'll probably want what RC enthusiasts call a "BEC" (battery eliminator circuit), which is usually a switching-mode power supply that regulates to 5v or 6v.  You could also use your own SMPS - but for dropping 9.6v to 5-6v at servo current draws (>= 1A for both servos), a linear regulator won't cut it.

[Soeren]

Hi,

I was thinking about this battery pack, as I can get it from a local supplier, and adding a 5v limiter on it so I don't fry the servos.

Nickel based cells are up to 1.45 fresh out of the charger and is considered flat at 0.9V

The battery you link to is only 650mAh, but since it's an 8-cell battery, it will go from ~11.6V down to 7.2V on a full discharge.

If you get plain AA NiMH cells, of 2200mAh to 2700mAh, 7 of them is good for from ~10.15V down to 6.3V over their discharge cycle (8 cells as the above NiCd battery, but still 3..4 times the runtime).

Which is better, 7 or 8 cells, depends on how you plan to keep the voltages in check. I assume you need 5V for controller and sensors and 6V for the servos.

With linear regulators you need to use LDOs with a drop out of max. 0.3V for the servos (or accept a slightly lower servo drive voltage near the End Of Discharge (EOD).
For the 5V line a drop out of <= 1.3V will do and any LDO able to handle the current will do.
The average losses will be something around:
  ((1.2 x n) - 5V) / I_controller
Plus
  ((1.2 x n) - 6V) / I_servos

(Where n is number of cells).
It's only a rough estimate, as the exact number will vary with usage pattern and a few other parameters.

Assuming, as an example, 300mA for controller and sensors and 1A for the servos, 7 cells results in a loss of 3.4W, while with 8 cells, you lose 5W. The average used power in both instances is 7.5W, so the efficiency will be 68.7% (7 cells) and 60.1% (8 cells) IOW, you loose either 31.3% or 39.9% of the power supplied from the batteries.

In another perspective...
A 2400mAh nickel cell has got around 2.88Wh, but you'll only be able to spend 1.98Wh (7 cell) or 1.73Wh (8 cell) per cell, as the remainder is lost to global warming.
Still though, the run time you'll get will be 3..4 times what the 650mAh NiCd battery can provide, which is 6.24Wh total or 0.78Wh per cell (and with the 60.1% efficiency factor, you're down to 0.47Wh/cell).

And regular AA-cells can be charged in any standard charger and can be swapped out for alkaline primary cells in a pinch.


With switching converters, you could get efficiencies up to ~95%, depending on parts selection and going that route, I'd recommend either 8 cells and a buck converter or eg. 4 cells and a boost converter, either one set to around 6V output, feeding that directly to the servos and using an LDO as a post regulator to get a nice 5V from that.

[Redcap]

Ah, thank you both very much

Soeren, as always, you have given much food for thought, as well as a solid starting point for any working out/decisions. Thanks for taking the time to reply

I'm assuming the switching converters you mentioned is along the same lines as the "battery eliminator circuit" that rbtying brought to my attention?

- I haven't been getting much sleep lately, so something may be clicking wrong in my brain there  


Edit: At the moment, I have the servos hooked up to the arduino which has its own 5v limiter.
Later I am hoping to further reduce this to just the microcontroller and not the entire board.

[rbtying]

Yes, switching power supplies are synonymous with "battery eliminator circuits".  As it happens, many beginners into robotics come with an RC background, and thus switching power supplies are often marketed as BECs.  They're probably also the cheapest ones you can find, and are designed usually to handle one to six servos (up to around 5A total is common). 

[Redcap]

Ah, awesome, thank you

[Redcap]

I've had a quick look on Ebay and found these batteries which have a capacity of 2600mAh.

Due to them being loose, I can buy/modify a battery holder for them as needed and recharge them safer with the charger I already have (instead of modifying one as per the tutorial and timing it).


Questions
I have some more questions, who's answers may be rather obvious. (Again, due to the lack of good sleep. Sorry if I'm getting frustrating  )

I'm planning to run the arduino & sensors off another power source, so only need to take the two servos into account.

5 cells x 1.2v batteries will give me ~6 volts to work with when fresh, and ~4.5volts when considered flat.

Will the 5 cells x 1.2v be sufficient?
I'm assuming it won't be due to the voltage when flat only being 4.5v.

Or should I still go with 8 cells (and thus more volts when flat) and use a buck converter?

Thanks in advance,
Red

[Soeren]

Hi,

Yes, switching power supplies are synonymous with "battery eliminator circuits".  [...] switching power supplies are often marketed as BECs.

That's stretching it a bit!
If you reverse your sentences, it becomes true, i.e. most BECs are switchers, but as you put it, is like saying that a four wheeled vehicle is an Audi (or whatever).

A BEC is any circuit (however the topology) that makes an extra battery superfluous.

They're probably also the cheapest ones you can find, and are designed usually to handle one to six servos (up to around 5A total is common). 

Even cheaper, it could be made from junk-box parts 

[Soeren]

Hi,

5 cells x 1.2v batteries will give me ~6 volts to work with when fresh, and ~4.5volts when considered flat.

Will the 5 cells x 1.2v be sufficient?
I'm assuming it won't be due to the voltage when flat only being 4.5v.

Or should I still go with 8 cells (and thus more volts when flat) and use a buck converter?

Nickel based cells charge to 1.40..1.45V/cell, so 5 cells fresh from the charger will be up to 7.25V.
They first bit down to ~1.35V/cell (6.75V) is pretty fast, but this may still be too much for the servos (read the datasheet carefully).
If you think your servos will handle 5 fresh cells, then go ahead - no need to tinker with a switcher if you can cope without it.

If you need a switcher, it would be better with at least 6 or 7 cells.
With an ideal (lossless) switcher (which is only built by a faerie in never-never land unfortunately), each extra cell (within reason) means around 2.6Ah x 1.2V = 3.12W extra power to spend -  in our world, around 2.8W..2.9W is a more realistic guess.

[Redcap]

Ahh, cheers Soeren
Yea, I think the datasheet is somewhat lacking. It does say 4.8v to 6v, so going to play it safe and implement a switcher.

Thank you for clearing that up for me