Simple battery question

[Luke Keyes]

Hey everyone,

Here's a simple questions (and hopefully one I haven't asked before,
if so I apologize).  Somewhere on my road of building robots, I've
picked up that you should keep your power sources for your motors and
your electronics separate, due to voltage/current fluctuations when
stopping/starting the motors.

However, after picking up two batteries for the ATX Hackerspace robot with
the plan of hooking one to the motor controller, and the other to the
carputer power supply, I've had quite a few people telling me that it
is unnecessary and inconvenient, and that everything should just be
run off of one battery with the other battery being a spare.

Can anyone with more knowledge/experience speak to this?

Thanks,

Luke

[Chris Albertson]

Using two batteries, one for electrotronic and the other for motors is
a simple and fool proof method. Alo it allows each to work at a
different voltage..

A more sophisticated design will likely use switching mode step up and
step down regulators to supply the voltages and these regulators along
with some filters, will also provide a great amount of isolation
between motor induced transients and the electronics. The two
battery solution is technically very easy and I think suitable for a
smaller project. A bigger project is going to want to run the
rechargeable battery over it's full range from full to minimum charge
and so you'll need the regulation and step up/down in any case.

> --
> You received this message because you are subscribed to the Google Groups
> "HomeBrew Robotics Club" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to hbrobotics+...@googlegroups.com.
> To post to this group, send email to hbrob...@googlegroups.com.
> Visit this group at http://groups.google.com/group/hbrobotics.
> For more options, visit https://groups.google.com/d/optout.



--

Chris Albertson
Redondo Beach, California

[Wayne C. Gramlich]

Luke:

The answer to this question is pretty long and involved and I won't be
able to respond until this evening at the earliest. There actually
multiple valid ways of dealing with the "motor noise" issue. Once
you understand what the issues are, you can decide which of the various
solutions you want to pursue.

"I'll be back!"

Regards,

-Wayne

[Michael Ferguson]

The one battery vs. two is a bit like C++ vs Java: some people think you should never do memory management, but a large group of people have obviously built programs that do memory management successfully, of course, they might have spent considerably more time trying to debug memory issues.

Within in the club, there are a number of very smart people on both sides of the fence when it comes to one vs two batteries.

With the single battery approach you have some upsides, like less wiring, less risk of your batteries running out at different times, etc. But, you have to take more care in how your electronics are designed. Extremely noise motors, lack of filtering on power supplies, etc will cause issues with your electronics. If you just slap a 7805 regulator on the 12V rail with no capacitors -- it's likely not going to work, but with proper filtering caps/inductors/etc, or a properly filtered DCDC converter, you can often be just fine (all of my robots have always been single battery).

With the dual battery approach, you are typically now routing two different power busses through the robot. Depending on your choice of motor driver, you might need to put optical isolation between the driver and the micro controlling it. Of course, once you have the proper isolation and wire routing, it is extremely unlikely that you will get motor-induced issues in the logic side of your robot.

In the end, the best solution really will come down to your skillset, and what you feel comfortable working with and debugging.

-Fergs

[Wayne C. Gramlich]

On 07/02/2014 09:27 AM, Luke Keyes wrote:

Mobile robots are self powered from self-contained power sources.
The most common power source is one or more batteries. The BEAM
robotics folks tend to use solar cells. It does not take much
imagination to think in terms of a fuel cell as well.

Mobile robots frequently have power distribution and management
issues. Inexpensive brushed DC motors tend to have two issues:

1) They can be "bursty" in the current draw.
2) The brushes tend to generate some high frequency electrical
noise that cause issues due to EMI (Electro-Magnetic
Interference) and inadequate power supply filtering.

DC motors can change their current draw very quickly. The simpler
motor control software is notorious for going from a dead stop
to full forward (or reverse) and causing a large spike of current
draw. These current spikes can do several things:

1) The current spike can cause a "ground bounce". The ground
return path from motor H-bridge has a resistance. The increased
current causes an increased voltage drop along the current
return path. This shows up at the microcontroller as a
drop in the power supply voltage. With enough ground bounce
the microcontroller will decide that it is in a brown out
situation or reset situation. This is more of a problem
with linear regulator chips (e.g. the 7805) than it is
with DC to DC converters.

2) The current spikes can cause ground loops where current
is flowing back and forth between the electronics in the
system. A bad ground loop can exceed the current capacity
of a wire and melt it.

It should also be mentioned that DC motors are not the only source
of large current spikes. For example, the Sharp GPD120 was notorious
for generating 250mA current spike with a quiescent current of 9mA.
Likewise, if you have a bunch of LED's that go from all off to all
on, a fairly surprising current spike can result.

The first decision point is whether to have one or two power
harnesses.

1) The single power harness solution has a single battery and
a single power harness to distribute the power. Your
logic electronics and your motors share the same power
terminals.

2) The dual power harness solution has one power harness to
power your devices with bursty current loads (e.g. DC Motors,
GPD120's, LED arrays, etc.) and another logic harness for
your devices with relatively smooth current draw (e.g
microcontrollers, bump switches, etc.) By keeping these
two power harnesses separate, you minimize the amount of
spurious microprocessor resets and brownouts.

The choice between one harness vs. two harnesses is a personal
choice. If you go down the one harness route, you need to be
more careful with your electronics design. Almost all
commercial products use the one harness design. They also
typically have one or more electrical engineers on staff to
get all the electronics right. If you go down the two harness
route, you have a different set of choices. In particular
you have to worry about communicating from devices on one
of the power harness to the other. This is typically done
with opto-isolators, pulse transformer, high impedance
operational amplifiers, etc.

In general, if you are an electrical engineer, you get to
make the call. If you are not an electrical engineer, I
recommend the two harness route for floor based robots or
larger. For table top robots, it probably does not matter
that much.

For the two harness strategy, you get to make some more
decisions:

1) Use two separate batteries and keep the power harness 100%
electrically isolated.

2) Use two separate batteries, but tie the ground wires of both
harnesses together at exactly one point that is very close
the minus terminals of both batteries.

3) Use one battery, but put a good DC to DC buck converted to
provide power to the logic harness. The motor harness is
attached directly to the plus and minus terminals of the
battery.

4) Use one battery to power both harnesses by connecting both
harnesses to the same plus and minus battery terminals.

Personally, I've done 1), 2), and 4). I tend to use 2) the most
since I can run the motor harness off of either a 12V or 24V
battery. The Logic harness tends to be run off of a 7.2 RC
battery. I've heard that 3) works pretty well, but I have
no direct experience with it. I no longer encourage anybody
to try 4), batteries are sufficiently cheap that there is
no reason not to use two. If you really, only want to use
one battery, go with 3). Pololu sells reasonably priced
step down voltage regulators.

When it comes to communicating between the two harnesses
there are a variety of technologies that can help out:

1) Classic opto-isolators. They are cheap, readily available
and easy to use. Their primary draw back is they tend
not to be very fast. They are just fine for driving an
H-Bridge.

2) There are higher frequency electrically isolated devices.
Analog Devices sells some devices that are based on pulse
transformers. In addition, Silicon Labs sells some that
are basically really small RF transmitters. The Silicon
Labs are good up to mega bit per second.

I mostly use opto-isolators, they are cheap and work just fine.

Now when it comes to EMI, the issues get fuzzier. Basically
our robots are made out of wires and those wires are perfectly
happy to behave like antennas and pick of radio frequency
signals.

In general, for EMI, there are several issues:

1) Signal Frequency. The higher the frequency the bigger the
problem. Remember AM radio occurs approximately from 500KHz
to 1500KHz. FM radio occurs from approximately 90MHz to 110MHz

2) Signal Slope. The squarer the pulses, the more high
frequency that comes out. Good signal transceivers are
slope controlled.

3) Signal Run Length. The longer the signal run, the more
opportunity there is to radiate and receive EMI.

4) Signal Transmission Strategy. In general, differential
transmission over wire pairs works better over longer distances.
For shorter distances, you can get by with simpler single wire
signal transmission works.

5) Shielding. The longer the signal run, the more you should consider
some sort of wire shielding. I used to live a 1/4 a mile away
from a fairly high powered radio transmitter and it caused a fair
amount of grief on my electronics projects. I tended to encase
my projects in metal boxes to provide a kind of Faraday
cage to keep the evil RF signals at bay.

6) Filtering. Sometimes you can put some filters in place to help
minimize high frequency noise. There is a whole sub area of
electrical engineering called Snubber circuits to try and
minimize the noise at the source. I use a simpler strategy
of putting ferrite beads on my motor power leads next to
the motor.

Dealing with EMI issues can be a bit like black magic.

------------

Now comes the section that is opinion. Note that below is
*my* opinion and reasonable people can have alternative opinions
that are perfectly valid.

For small Arduino based robots, I recommend the two battery
approach. Battery packs are cheap. Some of the Arduino
clones out there have poorly designed linear regulators,
so the dual battery approach may save some grief.

For larger robots, if you are an "electrical engineer", you should
understand the issues enough so that you get to make the call.
If you are not an "electrical engineer", I recommend the two harness
route. By the way, I put "electrical engineer" in quotes
to mean someone that is comfortable working with electronics.
I could care less if you have an official BSEE from some
sort of educational institution.

Again. These are my opinions. You are free to disagree.

Regards,

-Wayne

[Butokim]

Wayne,

Great reply. You really gave this some thought.

James "Jim" M. Geidl, K6JMG
D.B. Cooper, you have a message.

[Luke Keyes]

Thanks Wayne for the well-thought out reply.

My gut was telling me to go with 2 harnesses, and your response has made me decide to stick with that direction. 

A quick recap of the project -  This will be a large floor-based robot that is built around a donated Balder power-chair.  A Sabretooth 2x25 will drive the motors, and be controlled via an Arduino that is hooked up to a Dual core PC running off a M4 ATX power supply.  So it's going to be big (6' tall) and heavy (>200lbs) with some big current fluctuations.  We've already got it working via an RC reciever/transmitter pair for driving around.  Next step is putting in the brains.

Given that even though it's a collaborative effort, I'm the primary contributor, and I'm more of a "software guy". I'm going to say that I don't have enough electronics knowledge to feel comfortable going with a 1 harness method.  The people who have given outside input certainly do have the knowledge, but since I'm the one doing the work, unless I can get discrete instructions, I'm just going to have to ignore the input.

Thanks again to everyone who responded.  I definitely appreciate it.

Luke

[Butokim]

Luke,

 

I have a Jet 2 chair sitting in my garage with brand new batteries in it.  I also have a Sabertooth 2 x 60 to power the motors, a Arduino Uno for brians, and a GY521 Accel/Gyro to use with it.  My hopes are to built a homebrew Segway using the GY521.  Have you got time to answer a few questions?

 

James "Jim" M. Geidl, K6JMG
D.B. Cooper, you have  a message.
 

---

From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of Luke Keyes
Sent: Thursday, July 03, 2014 3:25 PM
To: hbrob...@googlegroups.com
Cc: wayne.gra...@gmail.com

Subject: Re: [HBRobotics] Simple battery question

--

[Jacob Thomas]

HI,

Why do we need filters for the regulators because battery provide pure DC...

--
Regards,
Jacob Thomas

[Chris Albertson]

On Thu, Jul 3, 2014 at 10:43 PM, Jacob Thomas <jacobth...@gmail.com> wrote:
> HI,
>
> Why do we need filters for the regulators because battery provide pure DC...

Because it is NOT pure DC. Try a simple experiment where you have a
battery, a switch and a light bulb all in series. Close the switch
and the voltage a gross the battery changes rather abruptly. This is
because the battery has some internal resistance. Think of a real
battery is being an ideal battery in series with some low value
resister. So by Ohm's law the voltage depends on the current.

If you attach some large motors to the battery they will draw current
in short bursts and by varying amounts so your battery voltage will
have all kinds of AC-like transients. Some of these might be enough to
effect other components on the robot.

[dpa]

Wayne,

Great Post!  Very thorough.  As these are problems that most experimental robot builders encounter, you might consider publishing this under a more general robotics wiki, for other not-list builders to find.  A great resource.

I've built robots using single supplies and dual supplies, and had good success with both.  A couple of additional points to add from my own experience:

1.  To help ameliorate motor noise, be certain that your metal motor cases are grounded.  

Given the popularity of plastic and 3D-printed robot chassis, this problem is more common than might be expected.  Run an extra ground wire from the supply to the motor mounts if needed.
The R/C eletric car/heli/airplane guys also often add a capacitor between the motor leads right at the motor, and one from each motor lead to the metal case of the motors, to help dampen motor-generated EMI noise.  More expensive motors often have these caps built in.

2.  For dual battery supplies, another method for signal connections between the two power systems which is somewhat simpler and pretty robust is to connect the grounds of the two supplies together through an inductor. 

This allows common ground between the two supplies, while preventing noise from the motor supply ground from propagating onto the logic supply ground.  I've found something like a 1 uH inductor works pretty well.

3.  In the realm of personal opinion, I have gravitated over the years to using single battery supplies.   Mostly because of the hassle of keeping track of the charge on two batteries, which always end up needing to be charged at different times and intervals.
Most of my recent robots run 24 volt motors from a 24 volt battery pack, and use DC-DC converters to provide the 5v logic and sensor supplies.

A pretty robust and simple solution for this that is widely available is to use the regulators which have been developed for the electric R/C industry, which they call "BECs" (Battery Eliminator Circuits).   Here's one from Castle Creations that can supply 10 amps at 5 volts.   I think I paid about $20 for it.  It is used on one of my newer robots that runs from a 6S LiPo pack at ~22volts.

<http://www.rctoys.com/rc-toys-and-parts/CC-BEC/RC-PARTS-VOLTAGE-REGULATORS.html>

These are often quite small.  Here is that one mounted on a robot using a bit of double sided tape.  It's the little package wrapped in blue heat-shrink mounted just over the PULSE battery pack.

These take a wide range of input voltages, and many, this one included, are programmable so that the output voltage is selectable .  (The heli guys, for example, often run their servos on 6 to 8 volts for quicker response.)   If you don't have R/C gear to do the programming, most hobby shops will program them for you.  They are physically robust and very good at isolating the supply noise from the output, giving clean logic power to run your microprocessors and sensors when powered from the motor battery.

happy roboting!
dpa

[dpa]

Oops forgot the link: 

<http://www.geology.smu.edu/dpa-www/robo/rcat/castle-bec.jpg>

dpa

[Wayne C. Gramlich]

David:

(Sorry about the delay responding, I got busy...)

On 07/04/2014 11:23 AM, dpa wrote:
> Oops forgot the link:
>
> <http://www.geology.smu.edu/dpa-www/robo/rcat/castle-bec.jpg>
>
> dpa
>
>
> On Friday, July 4, 2014 1:20:19 PM UTC-5, dpa wrote:
>
> Wayne,
>
> Great Post! Very thorough. As these are problems that most experimental
> robot builders encounter, you might consider publishing this under a more
> general robotics wiki, for other not-list builders to find. A great resource.

My intention is to publish this information in the HBRC wiki when it
comes back to life.

> I've built robots using single supplies and dual supplies, and had good
> success with both. A couple of additional points to add from my own experience:
>
> 1. To help ameliorate motor noise, be certain that your metal motor cases are grounded.

It can be pretty challenging to ground some motors. I personally use
radiator clamps when I do that. I wish motor vendors would put a grounding
tab on the motors, but it is not going to happen any time soon.

> Given the popularity of plastic and 3D-printed robot chassis, this problem
> is more common than might be expected. Run an extra ground wire from the
> supply to the motor mounts if needed. The R/C eletric car/heli/airplane
> guys also often add a capacitor between the motor leads right at
> the motor, and one from each motor lead to the metal case of the
> motors, to help dampen motor-generated EMI noise. More expensive
> motors often have these caps built in.

I've also had some good luck with ferrite beads. I put a ferrite bead
on each motor wire as close to the motor as possible.

> 2. For dual battery supplies, another method for signal connections
> between the two power systems which is somewhat simpler and pretty
> robust is to connect the grounds of the two supplies together through
> an inductor.
>
> This allows common ground between the two supplies, while preventing
> noise from the motor supply ground from propagating onto the logic
> supply ground. I've found something like a 1 uH inductor works pretty
> well.

It sounds like a good idea to me. Again, I also like ferrite beads
for this sort of thing. I view ferrite beads as "it can't hurt".

> 3. In the realm of personal opinion, I have gravitated over the years
> to using single battery supplies. Mostly because of the hassle of
> keeping track of the charge on two batteries, which always end up needing
> to be charged at different times and intervals.

My work around is to always charge both at the same time. I usually
have two different chargers because the "logic" battery has a lower
voltage than the motor battery. I understand the hassle tho'.

> Most of my recent robots run 24 volt motors from a 24 volt battery
> pack, and use DC-DC converters to provide the 5v logic and sensor
> supplies.
>
> A pretty robust and simple solution for this that is widely available
> is to use the regulators which have been developed for the electric
> R/C industry, which they call "BECs" (Battery Eliminator Circuits).
> Here's one from Castle Creations that can supply 10 amps at 5 volts.
> I think I paid about $20 for it. It is used on one of my newer robots
> that runs from a 6S LiPo pack at ~22volts.
>

> <http://www.rctoys.com/rc-toys-and-parts/CC-BEC/RC-PARTS-VOLTAGE-REGULATORS.html <http://www.rctoys.com/rc-toys-and-parts/CC-BEC/RC-PARTS-VOLTAGE-REGULATORS.html>>

>
> These are often quite small. Here is that one mounted on a robot
> using a bit of double sided tape. It's the little package wrapped
> in blue heat-shrink mounted just over the PULSE battery pack.
>
> These take a wide range of input voltages, and many, this one included,
> are programmable so that the output voltage is selectable . (The heli guys,
> for example, often run their servos on 6 to 8 volts for quicker response.)
> If you don't have R/C gear to do the programming, most hobby shops will
> program them for you. They are physically robust and very good at
> isolating the supply noise from the output, giving clean logic power to
> run your microprocessors and sensors when powered from the motor battery.

Having a separate battery eliminator or a good DC-to-DC converter can
eliminate the second battery. It is still a good idea to be aware of
the motor current/noise issues.

One of the reasons why I nudge beginners towards the two battery
approach, it is forces them to ask "is this power connector for
motors or logic?" After a few robots have been built, the comfort
factor is higher and different strategies can be used.

[much snippage]

Regards,

-Wayne

[robotMaker]

Thanks Wayne for your suggestions, came just in time, since I'm having a problems with one of my noisy motors, I should try first grounding the case because that's the easiest. I've had problems like this before and solved it by grounding the case, but I forgot about doing that. Makes me wonder what other things I've forgotten that would get me quicker results with my bots.

-Cesar


--------------------------------------------

On Wed, 7/9/14, Wayne C. Gramlich <wayne.gra...@gmail.com> wrote:

Subject: Re: [HBRobotics] Simple battery question

To: hbrob...@googlegroups.com
Cc: wayne.gra...@gmail.com
Date: Wednesday, July 9, 2014, 10:45 PM

[Steven Nelson]

I would suggest using three .01 or .001 uf ceramic capacitors rated at eh 600 vdc on your motor power leads One from the motor hot lead to the motor case. One from the motor ground lead to the motor case. One across the motor hot lead to the motor ground lead. Actually this is just a darn good idea even if you are not having a noticeable problem. Here is a nice drawing of how to use capacitors with DC motors. http://www.beam-wiki.org/wiki/Reducing_Motor_Noise

--
Inspection time: The period of time required to inspect a recently welded piece of metal using your bare hand.
The hotter the metal the shorter the period of inspection time

[Dave Hylands]

This is where I found some discussion about wiring techniques, which includes some background on why twisting the wires is a good thing:
http://www.4qd.co.uk/serv/appnotes/machinewire.html

--
Dave Hylands
Shuswap, BC, Canada
http://www.davehylands.com

[Maddox.Robert]

I am a harness designer and component engineer – if you have need of more specification on what and why regarding twisted pair I’d be glad to explain.

 

From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of Dave Hylands
Sent: Friday, July 11, 2014 9:58 AM
To: hbrob...@googlegroups.com
Subject: Re: [HBRobotics] Simple battery question

 

This is where I found some discussion about wiring techniques, which includes some background on why twisting the wires is a good thing:
http://www.4qd.co.uk/serv/appnotes/machinewire.html

--

[Andy Jang]

Hi Luke,

I have have also found in robot project blogs that people tend

to separate the controller power from the motor power source.

My intuitive view of this is that motors draw a lot of power when

switched on and off. Also, there is noise from connections in the

motor.

The controller is the brain of the robot so you would not want

fluctuations or noise that would disrupt the digital circuits. If

the robot is climbing up a steep hill we would not want the

fluctuations in battery voltage to reset the controller.

The PC power supply is good example of isolation to separate

sub systems. The mother board has a dedicated power cable

separate from fans and peripherals.

I found this article interesting:

http://robotics.stackexchange.com/questions/267/why-are-capacitors-added-to-motors-in-parallel-what-is-their-purpose

Regards

On Wed, Jul 2, 2014 at 9:27 AM, Luke Keyes <luk...@gmail.com> wrote:

--

[Chris Albertson]

The debate here is really about trading off engineering time and
complexity for cost, weight and ease of assembly. Generally if you
are building just one of something you want to minimize the
engineering even if this means you add a few pounds the to project.
Using two battery system does just that. It is simple and easy.

But if you were building flying robot of even a high performance
hexapod you'd care about every gram of weight and about the power
density of the batteries. When you care a lot about performance you
want to use one high power battery and then get the various voltages
you need from switching DC/DC regulators. The regulators also give
you a hugh amount of noise isolation too. But this eats up a lot of
development time

[Steven Nelson]

If you want to go the single battery route, I have found that these Automotive M3ATX power supplies work really well. They are small, lightweight, smart, reliable and designed to deal with input voltage surges.I've used them in large PC based robots and we solved several battery load/engine cranking problems with them on the All Power Labs, Power Pallet gasification systems that are Arduino based.

 http://www.logicsupply.com/components/power-supplies/automotive-power/

[Chris Albertson]

This is more of what I was thinking of.  It is small, costs well under $2 and will step down a large battery to what you 5v or 3.3v controller needs.  It will pretty much isolate the logic from the motor and take care of voltage change over take because of battery discharge or transients.DC-DC-3A-Buck-Converter-Adjustable-Step-Down-Power-Supply-Module-LM2596S

There is another version of the same circuit that is smaller Mini-3A-DC-DC-Adjustable-Step-down-Converter-Standard-Power-Supply-Module-LM2596-

[anfederman@comcast]

So it is a not so simple question.   Big 12v brushed motors can draw a lot of current, those sparks create WiFi interfering EMF noise and need to be controlled through relatively expensive PWM controllers.  So while a two battery solution will work, the expense side of the equation is in monitoring and maintaining the the batteries.  On the one battery side, it is easier to maintain and keep one set of batteries charged, but the expense is in isolating the various voltages and in motor controller circuitry.

 

So you can pay Paul or pay Peter, but in the end you have to pay someone.

From: Chris Albertson

Sent: Saturday, July 12, 2014 10:38 AM

To: hbrob...@googlegroups.com

Subject: Re: [HBRobotics] Simple battery question