# How to isolate from back EMF of brushed DC motor?

#### riemanndiy

Joined Jul 25, 2020
8
Hi I'm working on a brushed DC motor with permanent magnets. I'm giving it 12V as input voltage (with a bench power supply) and it produces a back EMF that has a range of about (-90,90)V that I've seen in my oscilloscope, since here nothing is strange.

The power supply and the oscilloscope are conncted to the same 230 AC (Europe AC voltage) through the same power strip...
I tried to completely disconnect my scope from the motor and the channel of my scope is detecting the back EMF anyway(same frequency of the DC motor), but in this case the range goes from (-5,5)V.

Since I will have to test the same kind of motor but also more poweful than the one I'm testing right now, I'm concerned about a way to completely isolate the bench power supply from the power strip so that the oscilloscope won't detect any back EMF as it should be and also to protect my other instruments/electric device.

How can I do that? Can an isolated transformer do this?

#### KeithWalker

Joined Jul 10, 2017
2,925
That is not back EMF, it is EMI (electro magnetic interference). Back EMF is a DC voltage generated by the motor in opposition to the applied voltage.
What is connected to the input of your scope when you see the EMI?

Joined Jul 18, 2013
28,023
A DC motor generates a voltage that is the same polarity as the applied voltage, the level of which depends on RPM and Load.
With no load applied, this generated voltage almost equates the applied voltage
Max.

#### kennybobby

Joined Mar 22, 2019
75
... and it produces a back EMF that has a range of about (-90,90)V that I've seen in my oscilloscope,
What is the frequency of this +/- 90 V signal? Is it a clean sinusoidal?

you may be picking up stray electric field from the mains AC.

If you have a 12V battery, you might use it to repeat the test to see if the noise is still present, turn on/off the power supply, overhead lights, etc to see what causes the noise to disappear.

#### LowQCab

Joined Nov 6, 2012
3,569
Here is what you need .......
.
.
.

#### riemanndiy

Joined Jul 25, 2020
8
That is not back EMF, it is EMI (electro magnetic interference). Back EMF is a DC voltage generated by the motor in opposition to the applied voltage.
What is connected to the input of your scope when you see the EMI?
I connected the scope using negative voltage for ground clip and positive voltage (+12V) of the power supply (that has no ground connection).

So here is the update of what is happened between yesterday and today since now: the Oscilloscope was in the first spot near the strip. I tried to move the Oscilloscope to another power strip ""to increase the distance from the power supply"" and this noise disappeared.
I tested this different times with different motors of the same type, and the maximum peak of this spikes (there's nothing sinusoidal or similar), it's about (-4V,4V) from its average voltage value: so let's say I apply 12V i would see (8V,16V) spikes, sometimes smaller than that, but not bigger than that.
The following measuring were made with 10X setting of the probe and oscilloscope:
The first image is from my scope, with the motor at 12V input,

The second image is with 5V input

You can clearly see that the spikes are not so big. So nothing scary here, infact yesterday I was able to connect an Arduino (powered by a 9V battery) to measure the voltage through a voltage divider (the voltage range [A0,gnd-Arduino] stayed ALWAYS between 0.5V and 4V). So I was happy.

So today I decided to try to change position to my scope and the power supply, to inspect if anything would change. Bad Idea... The scope started to show huge voltage peaks, also without the motor rotating: infact due to the high voltage i didn't want to try to run the motor with 12V.

To make you understand what I'm saying: this is the voltage spike when I applied just 1V (from bench power supply) and rotated the shaft with my fingers. It ranges about (-60V,60V).

I tried also with a 3V battery, to see if an isolated voltage input would behave the same and this showed the exact same result... Huge Spike!
So at the beginning I thought it was my bench power supply fault, I think now this is not the case.

I really can't figure out why yesterday everything was behaving in a way (first two pictures), and now it behaves completely different with huge spikes of voltage.

I will post the last image (same time interval of picture 2) of the motor driven with 5V input it has (-51V,85V). You can clearly see
the difference with the picture 2 that has almost no spikes.
From this last image I think we can see the similarities of the waves between the picture 2 and this one below. But this one has the spikes, while the picture 2 no.

I know this can be a really noob issue, but I have no idea why the same motor showed two different behaviors.
If this, from now on, will be the default behavior of the motor, i will need to design a circuit to protect any device that I could connect to (for example an Arduino in my case).
Anyone has some ideas of why there were two behaviours of this (I don't think it's rigol DS1054Z fault)? Thank you

Last edited:

Joined Jul 18, 2013
28,023
Have you tried a capacitor across the motor, start with a 0.5uF

#### riemanndiy

Joined Jul 25, 2020
8
Have you tried a capacitor across the motor, start with a 0.5uF
No I didn't thought it at all, I used a parallel capacitor to smooth the signal for the input voltage for A0 Arduino pin.

Since this is not "a real DC voltage" it has a duty cycle maybe this can work
You are saying to use the capacitor in series with the motor, correct?

#### kennybobby

Joined Mar 22, 2019
75
But what is the frequency of these signals? i can't see any gridlines and there are a bunch of useless measurements of min,max, mid, avg voltages. Set your scope up and identify the timebase, vertical v/div and set zero volts clearly identified, measure the frequency and we can read the rest.

#### riemanndiy

Joined Jul 25, 2020
8
Every image has a timebase of 50ms, only the third image has a time of 1us (that is the one where I spinned the motor with my fingers).
The first image has 12V input so it's the only image where you will not see the zero volts clearly identified.

This is the image with frequency and period. The zero volt line is where there is the little "2" on the left, 5 volt per division. From this image it can't detect the maximum and minimum spikes, but they are: (-51,84)V. The average voltage of this image is clearly 5V

i can't see any gridlines and there are a bunch of useless measurements of min,max, mid, avg voltages.
No thery are not useless (lol). Because repetitive reverse voltage can easily damage an Arduino. The whole issue is about the voltage peaks of the motor generates. So Min and Max are what I'm trying to normalize.

Whatever is the frequency, the motor will generate anyway, with no exception, high spikes!! So what do you find from the frequency data information?
Hint: the motor doesn't reach 10.800 rpm, but only 2400 rpm.

Last edited:

#### kennybobby

Joined Mar 22, 2019
75
What is your mains frequency--50 or 60Hz? The frequency measurement is showing 180 Hz, which is the 3rd harmonic of 60. Unless the number of commutator slots and the speed you were running just happened to fall into that frequency range, this is quite a coincidence.

Don't be adding capacitors if you don't know what they are supposed to do, or why. It is not clear if you were seeing AC noise from the mains in your OP as we never saw a scope trace.

If you want to identify the back EMF of the motor, connect your scope leads across the motor leads with no power supply connected, then spin the motor shaft by hand for one revolution and capture a trace for that revolution. If you have a drill motor then you could use that to back-drive the motor at a higher speed than you can do by hand to get another scope trace.

How about a scope trace at 1ms timebase so we can see several pulses--going from 50ms to 1 usec is too much

Can you see the brushes and commutator--are you seeing sparking and arcing when the motor is running?

Last edited:

#### LowQCab

Joined Nov 6, 2012
3,569
It's not uncommon at all to have RFI Spikes that Range up into the Ghz Range,
this may be too high for your Scope to process, so it may not show-up at all.

This is caused by the Sparking that is caused by the Brushes in the Motor,
and gets much worse as the Current gets higher.

Back-in-the-day, I built a PWM Blower-Motor-Controller for my Truck.
It continuously smoked MOSFETs left and right, sometimes within minutes of run time.

The Circuit Schematic that I provided earlier is what it took to
stop smoking PWM Transistors and Free-Wheeling-Diodes.
It "might be" over-kill for your application, but it definitely works "as-is".
My Blower Motor was ~25-Amps @ ~14-Volts.

The very minimum I would recommend is to place several
.
.
.

Joined Jul 18, 2013
28,023
You are saying to use the capacitor in series with the motor, correct?
No, in parallel
Here is what is installed on a popular make of T.M. leads.
Ferrite toroids
Max.

#### Attachments

• 310 KB Views: 9

#### LowQCab

Joined Nov 6, 2012
3,569

#### KeithWalker

Joined Jul 10, 2017
2,925
All the scope traces that you have shown are perfectly normal for a brushed DC motor running on a not very well regulated supply. If you don't want that EMI on your motor supply, take the advice of those who have taken time to offer it and suppress it with a capacitor, ferrite beads, a diode or all of them at the same time.
If you are controlling the speed of the motor with PWM, don't put large capacitors across the motor terminals. If you are reversing the direction of the motor with the controller, make sure that any diodes connected across the motor are also reversed at the same time.
When motors are used in conjunction with micrcontrollers, it is normal practice to connect the controller and motor to the supplies with separate sets of wires right back to the power sources. This reduces the possibility of introducing motor noise into the controller through the power supply leads.

#### shortbus

Joined Sep 30, 2009
10,037
When motors are used in conjunction with micrcontrollers, it is normal practice to connect the controller and motor to the supplies with separate sets of wires right back to the power sources. This reduces the possibility of introducing motor noise into the controller through the power supply leads.
Isn't that the reason they make what is called a "motor shield"? https://store.arduino.cc/usa/arduino-motor-shield-rev3