I have a circuit that has a voltage drop about half way through.
The circuit is spanned across 3 soldered bread boards.
The whole circuit draws a little over 2 amps. (20 servo motors).
If I grab my DMM and measure at the first servo I get 4.9v. And by the last servo I have 4.12v
I’ve tried various power supplies. In the end I borrowed my mates variable 5a power supply which also didn’t solve the problem. I’m happy to say that I have ruled out under amping the circuit.
My best guess is that the wires I’m using are too long, or too thin, or adding too much resistance. Thing is I am getting continuity throughout the board, so the resistance can’t be that high.
What are some common reasons for voltage drop so i can debug this?
Also consider the topology of the wires.
e.g. Are the cascading from one board to the next or each running direct back the the power supply (star configuration).
With higher currents you do need to plan your power and ideally have the high current back to the main power supply.
If you really need to split the wire, then do the math on the wires.
At the moment I also have my 3 boards connected power and ground.
If I’m giving each board it’s own supply, should I remove the common power cable?
I don’t think It would make much difference.
Thinking about this, I guess each power supply may not be exactly 5v so I would be creating a voltage differentials across the wire if I left the Vcc connected between boards of different supplies.
You may not need 3 supplies of the one you have should be able to supply the power you need. Just star the power runs back to the power supply. this way each draw is over its “own” wire.
If that is not ideal, then just consider the wire size as you add more current to it (e.g. as you get closer to the supply)
This is were its kinda nice of you have access to a thermal camera, you can see whats going on.
so something like
Vcc1 Vcc2 Vcc3
\ | /
\ | /
\ | /
Main Power Supply
Connection
If you can’t measure a resistance but you do see a voltage drop, then the resistance is dependent on the current. Some part of the circuit can’t handle that current without heating up and increasing its resistance. Are you using the breadboard traces? Adding a run of solder along the traces (a trick often seen on old power supplies) or beefing them up with a length of wire would be a good place to start. What wire gauge are you using and what are the lengths?
And all are dependant on each other. Any one (or both) of the first 2 will produce the 3rd situation.
As Michael pointed out if all items are in a string the first bit carries all the current and the wire carries progressively less current as the string progresses.
It all boils down in the end to Mr Ohm and Mr Kirschoff and I thought by now you would have been pretty familiar with these 2 gentlemen.
I also thought that by now with all the discussion regarding powering long strings of LEDs you would have been familiar with minimising this by powering at the centre or both ends or even both of these.
Anyway that is the obvious solution. Increase wire size or look at alternate powering points.
Cheers Bob
I just saw these other posts, arrived while I was having coffee.
DO NOT connect power supplies in parallel. For reasons outlined previously and I am not going to go into here.
Connect in a star configuration certainly using ONE supply or connect multiple supplies to their own groups only connecting the GROUNDS together. Mainly for control data integrity.
Note I am only talking about servo power supply here. Not controller power, this can be quite separate.
I’ve used some jumpers, some solid core, and some propriety cable. I think the smallest wire would be 28 or maybe even 30awg. It’s far too small in retrospect.
I think connecting the 3 boards’ ground together and then powering each of them separately is the solution. Next time I’ll use 22awg.
Hi Gerard, Pix
I have a fairly old Hartland catalog which unfortunately only goes down to 28AWG which says this size has a suggested current rating of 0.5A, stranding 7/0.12, area 0.079 sq mm and for tinned copper 242.34 Ω/km.
I suppose I could look up more details but so could Pix so I am not going to bother as I have no need to know at this time.
The only comment I have is that these wire sizes are not the sort of thing I would contemplate for any sort of power distribution.
Cheers Bob
Hi Pix and ALL
There is another insidious problem using small wire over a few metres.
If you have say a PWM motor speed controller on the end of this wire you could have a very real problem with the inductance of that wire. Smaller wire has appreciably more inductance per unit length than large wire.
Any wire but will have an effect depending on how much current is being switched. This current through the wire inductance forms a charge pump action which is added to the supply voltage thus increasing the supply by what could be several volts. As you can imagine this could play merry hell with any voltage sensitive electronics connected to the same supply.
A very good reason to have a separate motor supply in any circumstance
If this is unavoidable and circumstances are such that this motor supply wire has to be much more than 150mm or so the fix is to connect a capacitor of a few thousand µF RIGHT AT THE CONNECTION of this wire to the motor speed controller. Actual size of this cap will depend largely on how much switched current is involved but it could range up to 10,000µF or more.
Of course if you use a separate supply for all the sensitive bits you minimise any problems from this source.
Cheers Bob
I think you might have oversized those capacitors. Energy (joules) stored by an inductor is E=½LI²
A meter of 0.5mm wire has an inductance of about 1µH, stored energy with a current of 1A is 0.5µJ
When the current in the inductor falls to zero, this energy is dumped into the capacitor. Energy stored in a capacitor is E=½CV².
Assume the capacitor has no volts at the start, and limit the voltage rise to 1V, invert the formula to find the C required (V²=1 so disappears) C=2E, E=0.5µJ so C=1µF
But if the supply was say 12V, with 1µF the voltage rise is 1/144th of this, a few mV.
In practice, maybe 0.1µF is adequate. 10,000µF might be needed where switching hundreds of Ampere, such as in an Electrical Vehicle.
Hi Alan
I first had a major problem with this when trying to set a low voltage shut down bit of add on circuitry on a friend’s golf buggy with a variable voltage supply. Impossible while this effect was interfering. For a load on the buggy motor I had 2 similar motors coupled, one as a generator with a bunch of 12V light globes to vary the load (bit hard to put a load on a buggy while sitting on the bench).
Anyway the value I came up with about 10A switching current was 6800µF. This figure was due to the fact that I happened to have one handy and the smaller ones I had did not quite get rid of enough volts. I agree that 10000 is an overkill but it does as an example which I was trying to point out as a lot of hobbyists may not be aware of this little item. I might add here that the wire size used was about 1.0 sqmm. The wire in the buggy was something like 2.5 sqmm and battery to controller about 150 - 200mm. Quite short.
On another note this stray inductance can play merry hell with RF applications. I have seen an occasion where a switch wired with thinner wire made a transmitter impossible to tune. Replacing the switch wiring with the correct wire rectified this. Some applications use flat copper strap in lieu of circular wire as this also increases surface area and thus reduces RF resistance.
Cheers Bob
Anyway we are digressing a bit here. Pix original post was concerning voltage DROP, not increase. I just threw this in to warn about connecting sensitive items to the same supply which is being switched like PWM. The point I was making was that this could cause an INCREASE in supply voltage which could cause problems in sensitive devices and a separate switched (PWM) supply should be used. My personal opinion is this should be the case anyway.
Edit:
I just found that actual controller. The value was 1000µF, sorry but was a long time ago.
