Powering a Microcontroller and Motors with a Single Battery Source

@rhammell I could not find a 14500 battery holder with large connectors, the few there are have very thin leads. Why not go up a size to 18650, there are lots of holders with a large tab to solder to. Here is the picture of one.

That is good to know. For my current project I'm trying to get it working with 14500 for size limitations. 

See this image of the current chasis layout: 

I'm not opposed to using 18650, but for now I'd like to stick with the 14500 unless required. 

Hi @rhammell

I may have missed this but what is it that you want your robot to do?  What are you planning?  A good start for an autonomous robot might be one that explores your home and doesn’t get stuck.  Anyway, your circuit looks fine.  Although you might want to add an on/off switch near the battery.

Have you decided what sensors you want to install?  Do you plan to power them via the MCU?  If the MCU cannot handle the current draw you’ll need to go back to your circuit and run a 5V power bus off the voltage regulator.

Have you given any thought to how you plan to debug your software?  Most robots have some way of giving status.  Right?  When my Roomba gets stuck or gets low on power it plays a pre-recorded error message.  I’ve used colored LEDs, sound and LCDs for giving status.  You may want to add some of those to your circuit.

Good Luck   Tom

@thrandell

My bot will include a TFT display on top of it, which allows users to draw a path, which the bot will then follow.

So the user can pick up the bot, draw a path on its display screen and press 'Start'. Then place the bot down on the ground and it will follow the path traced out by the user. 

See a demo video of the interface here: https://www.reddit.com/r/arduino/comments/1hwnz2h/working_on_an_interactive_display_for_a_path/

And a still image here: 

Thats really all, just have the bot follow a path. So I just need it to be able to drive straight and make 90 degree turns. 

We've been discussing the circuit required to power the microcontroller (and display), and motor driver. When you say the circuit looks fine, is it option 1 or option 2 as we've been discussing. This is what my main question is: 

Option 1: 

Option 2:

@rhammell Option 2.

@rhammell One tip re that screen, do NOT try to solder direct to the screen, use headers so you can plug the screen in and out as needed. I can't recall if the screens come with pins, but I think they do so get some female hdrs and solder them to a piece of solderable bread board.

@zander 

Is there are reason you suggest not soldering directly to the screen? I have already attempted this and it works as expected - the display powers on and I can interact with its touch features

@rhammell Many new folks have difficulties, but if you have done it great.

HOWEVER for future reference, what if it breaks? Easier to use a hdr so you can replace it in 3 secs, also better for troubleshooting.

@zander 

Indeed, being able to fix it would be a plus. I do have spares with headers that I have been using to test out on the breadboard, but for this project, to fit within the desired size of the bot chasis I've gone without headers and am just direclt soldering to the pins of each component

Posted by: @rhammell

↑

When you say the circuit looks fine, is it option 1 or option 2 as we've been discussing. This is what my main question is: 

Option 1

Hi @rhammell,

Re: Creating this 'Y' split form one wire into two (for both the positive and ground wire coming out of the battery), is acheived by soldering two wires to the end of the original one, correct? 

Your second diagram is correct.

To be clear, in many cases, someone will adopt your first diagram approach, and it will also work. However, the second approach is better, as it reduces the opportunity for current surges from the motor being fed to the Arduino.

-------

As to whether, that is strictly " soldering two wires to the end of the original one", depends on the length of the wire from the battery holder, and whether the wire from the battery holder is adequate and well fixed.

Some of the battery holders I have seen, admittedly for alkaline cells, from 'low cost' sources, have had extremely thin wires, and a pathetic eyelet fixing between the springy terminal and the wire. These, I have 'resuscitated', by soldering decent wire(s) directly to the spring contact of the holder. Hopefully your holders are of better quality, and don't need bodging!

(Don't try soldering to the cells directly.)

If the original wire and fixing is ok, then I would keep this original wire as short as reasonably possible. Having a long single wire from the battery holder will, (as you suggest), defeat the object of splitting. So you will probably need to shorten it.

I also mentioned, inserting a fuse between the positive battery terminal and the point where your two wires connect, just in case you have a short or similar. Li-ion cells can turn very nasty.

Best wishes, Dave

Hi Ron @zander,

 Re: If it is a technical word other than resistor or capacitor it is either 100% unknown by me ...

In terms of components that might be added, I think resistors, capacitors and inductors (which you also mentioned before) pretty much covered my description.

I have already accepted my inclusion of 'ESR' could be a little unexpected, but is an essential parameter in this context.

I mentioned TI's ubiquitous 74-series logic, which we have discussed before, etc, but only as a typical circuit design part ...

I note that I have been a little lazy in not adding the word 'capacitor', when referring to electrolytics, ceramics, etc, instead of writing 'electrolytic capacitor', etc. ... That is normal conversation style ...

And I also referred to 'decoupling capacitor' ... which refers to the role that the capacitor is playing.

That is a 'coupling capacitor' is one that is transferring the AC part of a signal from one stage to the next, such as anode of one valve/tube to the grid of the next, or collector of one transistor to the base of the next.

'Decoupling' implies a kind of 'anti-coupling' ... in the way you might use the verbs 'couple' and 'decouple' when hitching a trailer to a car and later unhitching it. In the electronic sense, decoupling capacitors are aiming to prevent the current/voltage transitions in one part of the circuit, being accidentally transferred to another part of the circuit, such as via the power lines.

Sorry if I have caused any confusion, which is never my aim ... so please query any specific terms you are unfamiliar with, that I raised ... by doing so, you may help others at the same time.

Best wishes, Dave

Hi Ron @zander,

I am not planning on soldering onto the board, just to the pins or in the case of a 'strip like board' anywhere along the path (the path is HUGE compared to the wire size so an extra 1 cm is irrelevant.

Sorry, but the latter part of this statement is simply not, in general, true.

Wire length is a massive factor in this game.

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

The early Arduino boards are quite small, with only a few chips, so it is not so obvious, but a typical medium-sized logic circuit board, such as the microcomputer boards in the 1980s, may have 10s or 100s of capacitors across the same pair of 5V power lines are a classic example of the problem/solution. So why might they have 50 0.1 microfarad capacitors connected in parallel, instead of one 5 microFarad capacitor?

Simple, each 0.1uF cap had to have a very short wire lengths to the chip it was serving, to be useful. At the clocking frequencies of the day (say 1MHz-5MHz), a few mm of wire (or more likely a few mm of PCB track) was just about acceptable, whilst 100mm across the width of the board was not. A chip handling high frequencies will not 'see' the effect of a capacitor that is connected with 'long' wires ... and 'long' in this context is often a few mm or less.

Of course, most boards also have some lower frequency content, and here a small number of electrolytic capacitors, of say 100 microFarads, can service a wider area, and provide a supplementary benefit, but only for the low frequency transitions.

So in most cases, adding a capacitor on strip board, with wires of 1cm or more, will do nothing useful, to controlling noise with frequency ranges above say 100kHz.

ESP32s, etc have 100Mhz+ instruction processing, and 2.4GHz WiFi signals to handle. At these frequencies, even 1mm can be a long wire.

It may not be obvious, but surface mount components are an essential part of contemporary electronics ... not just because they take less space and are easier to handle on an automated soldering line, but because their wire lengths can be much shorter.

Best wishes, Dave

@davee I think you misunderstood me or I didn't explain well enough. It's NOT wire. I am speaking about a large metal bus like connector many times the cross section of the typical wire connected to the battery. It is much neater and easier to work with than a much larger wire as is often the case. I have done that often for various reasons. Sometimes it is just too difficult to solder on another wire, or the area is cramped and the soldering iron could upset or damage nearby components. In a recent case I had a bunch of these pins so just cut a strip of the pictured board. It is very much like using a big (relatively speaking) bus. Yes, ideally reducing the wire length is ideal for several reasons like power loss, high frequency radiation and perhaps even overall impedance (you taught me what that is). In at least one situation I have also used solid copper wire as big as 10AWG. Just lay it over the stripboard and tack each end with solder. I then was able to use smaller wire to hook up several positive leads and in another case it was done to provide a robust ground. Sometimes a bus approach is better than a star approach as the small wires are shorter, and the bus is 1/10 the resistance of equivalent hook up wire.

Just an FYI, when I designed my RV electrical system I used a HUGE marine grade bus bar for both the positive and negative sides of the battery, inverter, solar connections.

Hi Ron @zander,

  Sorry, but I think we are accidentally confusing each other. And visualising what happens in fast switching circuits is difficult and counterintuitive, even though the underlying principles may seem to be straightforward. 

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

In this case, and not for the first time, I think we may also be having two partly related discussions in the same thread, a confusing the two.

The thread started by asking about wiring of a single power source, and two loads, one (motor+driver), which has a generic reputation for being electrically noisy, whilst the other (microcontroller) can be sensitive to noise. As a precautionary approach to reduce the chance of the noisy load upsetting the sensitive node, I stated it would be preferable to split the load wiring into two branches, as close as reasonably possible, to the source.

I stand by that answer. Part of the reason is that the lower frequency noise, that might be generated by the motor/driver, will often be absorbed by the battery, and hence not passed across to the microcontroller, providing the wiring topology from the motor is more 'direct' to the battery than to the microcontroller. In principle, adding further capacitance close to the point that the battery is connected may also help, but in many cases, it will make little or no difference, as a healthy battery will act in a broadly analogous manner to an electrolytic capacitor.

--------

My recent discussion part of the thread, relates more closely to a subtly different cause that was triggered by your query of including extra capacitors for the benefit of the microcontroller node, which is often suggested to deal with nefarious logic related problems, such as unexpected processor crashes.

Each problem circuit has its own story to tell, but typically this can be for a related, but different cause to obvious noise sources, like motors. And the different cause means that the cure is also slightly different.

--

Fast switching circuits, including microcontrollers/microprocessors are not only sensitive to external noise, but (maybe more commonly) can also be the source of the noise.

That is, the instantaneous current flows arising when an output pin rapidly changes state, can be high enough to create a noise spike. In such cases, the noise spike tends to affect the power input wiring (e.g. Vcc and Gnd) of the chip, in very close proximity to the output pin, both inside and outside of the chip package, when the output pin driver (inside the chip) attempts to source or sink the current associated with each change of logic state.

In principle, just one output pin changing state can cause the problem, but it is more commonly seen when several pins, maybe driving and address or data bus, all change state at the same time.

To overcome this problem, chips with multiple logic outputs usually have multiple Vcc and Gnd pins, physically interspersed between small groups of output pins. (Output pins in this context means all pins which can be configured to be an output, such as GPIOs, as well as dedicated output pins.)

The circuit designer is then 'expected' to provide 'ample and competent' clean power to all of these Vcc and Gnd pins. The exact means of achieving this depends on the demands of the individual circuit design, but the simplest and cheapest approach includes multiple capacitors attached to the Vcc and Gnd pins, with as little wire/track as possible. This is an example where just 1 mm of extra wire can make a difference.

Boards designed to a 'price' inevitably skimp as much as possible, and dev boards, by definition, cannot be tested in their final 'usage' circuit, so it is not surprising, problems may appear. Unfortunately, if such problems arise, the solutions are likely to need capacitors, etc, very close to affected chips. In the case of ESP32s, etc, effective changes could be even harder to achieve, as the actual processor is a tiny device, about 7mm square, hidden under the metal lid! 

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

Your use of metal bars and strip board seemed quite novel, and in some cases, might be a useful approach. For DC power, when the load does not perform high speed changes of current demand, the approach you suggest could be appropriate.

Other cases, such as the input to switch mode power supplies, and power input to processors, especially the higher power consumption ones, can be sensitive to even the smallest amount of inductance ... and every conductor, regardless of thickness, has inductance.

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

Although you may never meet examples, wires and metal bars behave in 'strange' ways with alternating currents, even at lowish frequencies. The impedance of a given 'thick' wire increases with frequency because the current is largely confined to the 'skin' of the metal. High AC currents, requiring 'thick' conductors, can be replaced by 'pipes', of the same external diameter, with little change to their impedance, but with less metal and weight, and corresponding 'resistive' power loss.

Alternate approaches include 'Litz wire', which consists of multiple wire strands, insulated from each other ... Wikipedia has details if you are interested in a read ... https://en.wikipedia.org/wiki/Litz_wire

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

As an aside, I was amused to see a photo of stripboard ... As a 'Brit', I still remember the introduction of 'Veroboard' in the mid 1960s in the UK amateur electronic DIY magazines such as Practical Electronics. A quick Google showed its origin dates back to 1959 patent! https://en.wikipedia.org/wiki/Stripboard , which was even earlier than I expected.

Of course, nearly all of the present product is marketed as "stripboard", to avoid a trademark dispute, but for me, it is still "Veroboard".

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

I hope you find some of these details interesting, even if they are not relevant to your current projects. And apologies if they are tedious.

Best wishes, Dave