Motor Noise Suppression - In deep and need advice.

Thread Starter

daba1955

Joined Apr 27, 2019
84
Hi all, Please excuse the long post, but there is so much I need to get sorted.

i'm working on a project that has a brushed 12V DC Motor, and believe it or not, two choices of speed controller.

One is an H-Bridge controller that runs off a manual switch/pot for local control, and the other is a MOSFET PWM driven by an Arduino Nano, controlled by an EV1527 learning receiver, for remote control. The motor could be changed to brushless, but they have been purchased in quantity.

I have to be sketchy about the application, as it is a commercial product, and patents will be applied for, but I have attached the schematic.

Now I have the problem that when the electronics is all packaged in its aluminium tube housing, electrical noise is affecting the 433MHz RF signal for the Arduino control.

I am attacking the problem in several ways.

1. Add a grounded tinned copper shield around the EV1527 module and the Arduino
2. Modified the PCB layout to minimise the cable length from main PCB to Motor.
3. Add LC filtering to the MCU supply line.
4. Add LC filtering to the motor (I have opted to do this on the motor terminals, to try to keep the noise away from the feed wires)
5. Experiment with the Arduino PWM frequency, which is currently 490 Hz, whereas the H-Bridge frequency is 21kHz. I can go to 3.9kHz or 31.3kHz. The traces show the waveforms across the motor terminals for both controllers, you can clearly see that the H-Bridge is much cleaner, so going to a higher frequency must suit the motor better.

So I have to add LC filtering to the motor to try and squelch the noise to a level that the EV1527 can cope with, and following recommendations from Kerry D Wong's "A Short Guide On Motor Electrical Noise Reduction" have designed a PCB to add to the motor rear-end containing the necessary components. I've obviously gone for the dual choke/capacitor scheme, as one of the controllers is an H-Bridge, and thought I had chosen suitable 10mH inductors but I have noticed they have a high DC resistance of several ohms, which I don't believe will be acceptable in this application.

Looking around for that sort of inductor with low DC resistance, I am finding that they are all way too large to be usable in this project, they would even be too big for the main PCB. But I don't wan to put them there, as the motor wiring is the interference culprit, so want to kill it at the motor.

I would be grateful for any suggestions or corrections, I have to nail this interference issue.

TIA

MOSFET_Driver.jpg H-Bridge_Driver.jpg Schematic.jpg MainPCB.jpg MotorPCB.jpg
 

Sensacell

Joined Jun 19, 2012
2,438
Most cheap RC toy cars have two ceramic disc caps soldered to both motor terminals and the metal motor case, short and tight.
These serve to bypass the RF energy generated by the brush arcing, close to the source.

Ground loops are also critical, have a careful think about where the ground currents are flowing.
 

mvas

Joined Jun 19, 2017
538
Now I have the problem that when the electronics is all packaged in its aluminium tube housing, electrical noise is affecting the 433MHz RF signal for the Arduino control.
You have the receiver inside of an aluminum tube ???
 

Hymie

Joined Mar 30, 2018
708
I would add to Sensacell’s suggestion that you also place a ceramic capacitor across the supply terminals of the motor.

The capacitors need only be of the order of 10nF.
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
You have the receiver inside of an aluminum tube ???
Yes, but if you look at PCB layout there is a landing spot for an M6 stud, insulated from the case, onto which screws the 433MHz stubby aerial. The end of the PCB is slotted into a nylon housing that also serves as one of the battery retainers, and the stud makes contact with that PCB land.

I had to dramatically reduce wiring to the barest minimum, as the original scheme had two PCBs, the DP3T power switch, leads from the battery, leads to the motor, and a wire to the base of the mounted aerial. It was absolutely impossible to assemble and get it all inside the tube. This new PCB made assembly relatively easy.
 
Last edited:

Thread Starter

daba1955

Joined Apr 27, 2019
84
I would add to Sensacell’s suggestion that you also place a ceramic capacitor across the supply terminals of the motor.

The capacitors need only be of the order of 10nF.
The test I linked to says "But in general, the capacitors on the motor side are several hundred nano-farads and the inductors are several milli-heneries. Higher order LC filters can further improve the noise-filtering performance but are rarely necessary."

I have updated my schematic with what I am proposing,

2019-05-19_144647.jpg
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
I am most concerned about the DC resistance of those Inductors I have earmarked for the motor PCB.

They are CoilCraft RFS1412-106KE
The datasheet says they are 9.9 Ohms DC resistance, will that be too much for my motor which is a Mabuchi RS555PH-3255
The supply voltage is 14.4V NiMH, so could trade-off some top speed if I used those inductors.

The motor has very little load, and I have measured the current at 520mA at full speed.

Is the flywheel diode on the motor necessary ? Both controllers drive the motor the same way. And that has opened up another possibility that I don't need 2 LC circuits, just one, as the motor never goes into reverse.

Ahhh, so many questions, sorry guys....
 

MaxHeadRoom

Joined Jul 18, 2013
19,014
I would be looking at a small common mode choke, Digikey have them for this application, both through hole and surface mount.
Max.
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
I would be looking at a small common mode choke, Digikey have them for this application, both through hole and surface mount.
Max.
Can't find any 10mH chokes that will do 500mA that could be called anything but "small".

What part number are you thinking of ?
 

Hymie

Joined Mar 30, 2018
708
I would try the capacitor only fix first, if successful it will be much cheaper (and simpler) than the cost of the inductors – try 10nF, they will have a better performance at high frequency compared to the 100nF caps.
 
I see several issues with the pcb layout. You have your low current and high current sections intermixed. Move the battery and switch closer to the motor control section. Place all of the low current devices (processor, radio, regulator) on one end and all of the higher current devices at the other.
Your high current traces of opposite polarity are too widely spaced. For example, the motor + and - are on opposite edges of the board. This increases the inductance and will radiate a magnetic field. They should be along the same edge and on opposite sides of the board. The same thing goes for the battery input pads. Move them away from the processor and place them closer together. (near one edge) Power and ground should always be kept close together when possible. Place a small filter for the power and ground to the voltage regulator instead of trying to filter the current to the motor. (i.e. separate the low current from the high current both physically and electrically.)
All of the shield you mentioned will work for electric fields but won't do anything for the induced interference from the magnetic fields on the motor. You may need to use mu metal to shield the magnetic field at the low current circuit.
You can increase the motor leads but make sure you twist the leads to reduce emissions.
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
I see several issues with the pcb layout. You have your low current and high current sections intermixed. Move the battery and switch closer to the motor control section. Place all of the low current devices (processor, radio, regulator) on one end and all of the higher current devices at the other.
Your high current traces of opposite polarity are too widely spaced. For example, the motor + and - are on opposite edges of the board. This increases the inductance and will radiate a magnetic field. They should be along the same edge and on opposite sides of the board. The same thing goes for the battery input pads. Move them away from the processor and place them closer together. (near one edge) Power and ground should always be kept close together when possible. Place a small filter for the power and ground to the voltage regulator instead of trying to filter the current to the motor. (i.e. separate the low current from the high current both physically and electrically.)
All of the shield you mentioned will work for electric fields but won't do anything for the induced interference from the magnetic fields on the motor. You may need to use mu metal to shield the magnetic field at the low current circuit.
You can increase the motor leads but make sure you twist the leads to reduce emissions.
Thanks for the comments Charles, much appreciated. I will see what I can do with the board design, but some of the locations are fixed due to the tubes already having been drilled. Certainly the physical location of the battery pack, circuit board, and motor are fixed due to design configurations - the battery has to be removable for replacement with a charged one, and the motor has to be removable for maintenance. The picture shows the arrangement.
2019-05-20_062201.jpg
One of the problems is the length of the motor leads. They have to be soldered to the motor before insertion into the tube, so the excess wire goes under the PCB. This has to be where the interference is coming from, and a re-think on assembly order can prevent that, at the expense of making motor replacement more difficult, as it would necessitate a complete strip-down of the whole assembly. Let us hope that maintenance of the motor is not a regular occurrence. I am currently working on making the leads as short as possible, and implementing as many of your PCB changes as I can....
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
Charles, I have done a re-design of the PCB layout, tried to segregate Power and Signal as much as possible, and incorporated some of you suggestions. One of the biggest changes was a 180 rotation of the relay, that has helped a lot. I think I stated that the toggle switch cannot be moved

I am putting a grounded tinned copper screen above the Nano and the RF Receiver, and the motor leads will now be about 3" max.

I am concerned with the resistance of L1, 9.8 ohms, so can try it first with a wire link. Same goes for the LC networks I am going to put on the motor lugs, see the round PCB layout from post #1.

I would appreciate your comments on this new design.....PCB Ver 2c.jpg
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
One thing I have realised is that nowhere does the aluminium tube get grounded, and thus the motor casing is not grounded either. I believe I need the tube to be grounded to act as a ground plane for the aerial. The new PCB for the motor will allow me to ground the motor casing, and the tube.
 
I don't think the motor lead length is a problem, make them as long as you need to for easy assembly but make sure the cable has twisted conductors. This will help minimize radiated emissions.
The new board looks much better, still not perfect but the physical constraints limit the possibilities.
I think you can do without L1, I don't believe it provides any benefit and may actually cause problems.
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
I don't think the motor lead length is a problem, make them as long as you need to for easy assembly but make sure the cable has twisted conductors. This will help minimize radiated emissions.
The new board looks much better, still not perfect but the physical constraints limit the possibilities.
I think you can do without L1, I don't believe it provides any benefit and may actually cause problems.
Thanks Charles, I will leave the PCB design with L1 in place (just in case, for now), but will call it Link1.

On the existing PCB I've got the motor leads down to 3", and put some twist in them. Assembly was not so bad that I can't live with it.

I wasn't able to put Caps on the motor to frame, can't solder to the motor body, but have put some on the PCB, directly on the motor wire ends.

I have also grounded the motor casing, which in turn grounds the aluminium tube.

The remote range has trebled! From 6" to about 18", still not good enough though.

I have posted on the RF forum, concerned that the ANT input to the RX480-E receiver is grounded, I'm measuring 0.3 Ohms. Surely that can't work ...

2019-05-20 10.58.56.jpg
 
Have you tried to source some mu metal? It is magnetic shielding metal and should be easy to get. Place it around the sensitive low voltage and low current circuitry.

The DC resistance of the antenna depends on the structure. A short circuit will appear as an open circuit at 1/4 wavelength. You have now crossed over into the magical area of Radio Frequency! I don't know the structure of your particular antenna but if you are not certain try a replacement.
 

Thread Starter

daba1955

Joined Apr 27, 2019
84
Have you tried to source some mu metal? It is magnetic shielding metal and should be easy to get. Place it around the sensitive low voltage and low current circuitry.

The DC resistance of the antenna depends on the structure. A short circuit will appear as an open circuit at 1/4 wavelength. You have now crossed over into the magical area of Radio Frequency! I don't know the structure of your particular antenna but if you are not certain try a replacement.
That's exactly the sort of information that I lack the knowledge of, Thanks Charles.

A replacement should be here today. I'm concerned that the antenna I'm trying to work with is broken inside, as the hex insert can be turned in the rubber moulding. I believe it to have become detached from the helical coil. I have no way of testing this hypothesis, other than to carefully cut the rubber away. Needless to say, having the antenna screwed onto the stud makes little difference to the range I get.

2019-05-20 15.09.43.jpg
 
Top