Probable RFI issue between servos and locos

[videobruce]

I have 5 N scale Atlas diesels (out of a total of 30) that are causing a chattering to my turnout servo motors. Speed dependent, the faster I run the engines, the louder the chattering gets.
This seems to just affect servos closest to the loco, not the entire layout (9x19' layout, DC operation, 60+ servos total in operation). Usually when I'm running the engine over the turnouts, but not always. There is one spot that is just flex track when the chattering occurs.

I did come across this thread that seems to describe what I'm getting;
http://blog.model-train-help.com/2017/08/problem-with-chattering-and-oscillating-servo-point-motors.html

But, other than possibly adding a 'cap' across the motor leads, that was the supposed 'fix'. Anyone else heard of this happening??

Again, it's just 6 out of 30 locomotives that are causing the problem. Life Like GP20, 2 Atlas GP40 & 2 Atlas GP7.

BTW, all those units run smooth & quiet by themselves, no stalling,  sputtering or finger needed to get them to move. 

Your thoughts please.

(updated post)

[peteski]

If there is any RFI, it has very low power.

The RFI is usually generated in DC locos by the arcing of the motor brushes and also any arcing between the track and wheels.  Maybe shielding the circuitry inside the servo with metal (aluminum foil) which is connected to the negative (ground wire) feeding the servo would help.  Not quite sure jow to accomplish this (maybe wrap the entire servo in aluminum foil connected to negative wire?

Yes, a  possibility (if not already installed, would be tin install RFI suppression components (couple of chokes and caps) at the loco's motor like it has been done in European models from the start.  Some recent American prototype models (made in China)  have the RFI supression circuit installed, but most do not.  Just installing a cap across the loco's motor might work, but it will not be as effective as the choke/caps combo.

I suppose the RFI could also be picked up by the wires connected to the servo, then getting through to the low power servo driver circuit.  Placing a 1uF ceramic cap across the power and negative wires feeding the servo might also be worth a try.

[videobruce]

Wrapping the servos would be a tall order. Getting a good ground would be tougher. Can't solder to foil can one?
Would it matter if that cap was across the leads at the servo end or the controller end?

[peteski]

Yes soldering anything to aluminum is very difficult. But I suppose wrapping some bare copper wire over the aluminum wrap might work.

If adding a cap on the power feed to the servo I would install it close to the servo itself.

While the thread you pointed to earlier mentions that adding just a capacitor across the motor leads did not take care of the problem, it might still be worth a try (at least as a quick test).  Even a 0.1 micro Farad ceramic cap should do a good job shunting RFI.  That might be the simplest thing to try (even if you have to leave the shell off for the test).

[mmagliaro]

So the servo controller moves the servos by the controller sending pulses to them,
where the duration of the pulses controls the direction of the servo (I did not realize that's how these worked, so I learned something today!)  And the pulse widths are 1, 1.5 or 2 ms, typically.  That's a really bad number for susceptibility to relatively low-frequency motor RF noise.

Did you use twisted-pair wire from your servo controllers out to the servos themselves?  That can help reject induced noise on the wires (and I suspect that's the problem - that those wires are like big ol' antennas picking up the motor noise).

The other thing you can try, already discussed, is to put a 0.1 uF cap between the signal line and ground at the servos.

--
While you have just a few engines that are generating enough RF noise to cause this, my instinct would be to make the servos resiliant to this type of noise, rather than just fix the few locos that are causing it by putting a choke/cap circuit on them, even though that is probably less work because you only have a handful of engines causing the problem and you have 60+ servos.  The reason is that your servos are clearly susceptible to RF noise, and another engine, or even some other noise source, could always come along and start causing this again if you just fix the engines.

I'd start by changing the wiring to one servo to twisted-pair if you didn't do them that way, and see if the problem goes away on that one servo.  Secondarily, if that doesn't work, try the caps on the servo.

Here's an interesting discussion:
https://electronics.stackexchange.com/questions/275441/what-caused-this-noise-in-these-servo-signals-and-how-do-i-properly-describe-the
This guy's problem is from two servos interfering with each other, but the principle is the same - induced noise on the line.

[videobruce]

Thanks for the link, but most of that is over my head. I do get the "big ol' antenna" deal. Dealing with RFI is a PITA, but doesn't it make more sense to deal with the source?
The feeds are 3 wire (flat leads).

[peteski]

Thanks for the link, but most of that is over my head. I do get the "big ol' antenna" deal. Dealing with RFI is a PITA, but doesn't it make more sense to deal with the source?

Sure.  Install caps and chokes in your locos and see what happens.

[CRL]

I cannot vouch for this, but an old audio guy once told me that making super neat wiring with squared off corners in wiring audio components created RFI antennas, and the way to minimize RFI was to let your wiring run in loops & curves. Ham radio operators will install devices in their neighbors homes to hold antenna connective wires in tight loops on TV’s to break up the RFI from their transmitters.

[alhoop]

I cannot vouch for this, but an old audio guy once told me that making super neat wiring with squared off corners in wiring audio components created RFI antennas, and the way to minimize RFI was to let your wiring run in loops & curves. Ham radio operators will install devices in their neighbors homes to hold antenna connective wires in tight loops on TV’s to break up the RFI from their transmitters.

The old audio guy is correct - pretty wiring, ,ie, squared off corners, parallel lines and the like are trouble. my experience predates when the ET's gave us the transistor.TIC
Al

[mmagliaro]

Thanks for the link, but most of that is over my head. I do get the "big ol' antenna" deal. Dealing with RFI is a PITA, but doesn't it make more sense to deal with the source?
The feeds are 3 wire (flat leads).

You can certainly try the chokes and caps on the engines.  And dealing with the "source" makes sense, except that in this case, you have a potentially unlimited number of sources.  You may not get rid of this problem by fixing the handful of engines that are currently exhibiting it if you happen to pick up other engines that do the same thing.  I would argue that a more permanent fix is to make the servo wiring less susceptible to the interference.

As for the 3 wires, I would have to see some sort of wiring diagram or other info, but I would guess that only one of them can be a ground, so try wiring a cap from each of the other two to the ground and see what happens.
The flat cable is trouble for noise like this.  Any chance you can temporarily rig a 3-wire twisted cable up to just one of the servos to see if the problem goes away?

[peteski]

As for the 3 wires, I would have to see some sort of wiring diagram or other info, but I would guess that only one of them can be a ground, so try wiring a cap from each of the other two to the ground and see what happens.

Not sure if connecting a cap between the control line and ground is a good idea.  Control signal for a servo consists of short pulses (basically a low duty-cycle square wave). Square wave consists of many higher frequency harmonics.  A cap can attenuate those harmonics, and distort the signal. Shielding the signal wire would be am more reasonable solution.

While Bruce does own many locos, ony very few exhibit the problem.  I suspect that there is something in the motors of those few examples (brush composition or condition of the commutator) that are causing the noise. INstalling the RFI suppression in those few mostly should nto be too time consuming.

Actually Bruce, this gives me an idea.  You mentioned that you have 2 of the same loco model, and only one exhibits the problem.  If you are handy enough, try swapping the motor between those models and see if the problem follow the motor (I believe it should).  If yes, replace the brushes and try to clean the gaps between the commutator segments to see if that solves the RFI problem.

[nickelplate759]

A reminder for folks who don't deal with such things regularly - Ethernet cable (Cat5, Cat6, etc.) is made up of four twisted pairs (inside the cable bundle).  So if you use it to hook up servos, etc. it should help with RFI rejection.  You can buy it at  hardware stores by the foot, or you can get 100m boxes of the stuff if you want a lot of it.

[videobruce]

Thanks for all the replies, but I'm not getting email notifications of those, so I haven't been here any sooner.
I had the suggestion of running modifying the line to the servo at the servo end, but I see that might cause other problems.

What I was able to do 12 years ago when I started this layout, I find I can't do now which I'm not happy about. Age has crept up to me unfortunately.  It seems the easiest solution is to sell the units that exhibit the problem. Like I stated they run fine, no finger operation needed to get them to move, slow or fast which makes me mad since most are my favorite locos.

I updated my OP, all three of the GP40's exhibit the problem. I must of missed testing one initially.

[mmagliaro]

I cannot vouch for this, but an old audio guy once told me that making super neat wiring with squared off corners in wiring audio components created RFI antennas, and the way to minimize RFI was to let your wiring run in loops & curves. Ham radio operators will install devices in their neighbors homes to hold antenna connective wires in tight loops on TV’s to break up the RFI from their transmitters.

Yep, and the same is true of old tube amps (which I use for guitar).  You really don't want to get inside those hand point-to-point-wired high voltage tube chassis and start moving wires around so everything is parallel and neatly aligned along the chassis edges.  It doesn't have to be a "rat's nest", but you don't want a city-block-style "grid" either.

[Maletrain]

A related question in my mind is what makes a group of wires "twisted"?

Would a flat cable (with 3 conductors) that is twisted along its path between components be any different than three individual wires that are twisted together along the same path?

I had been told that the purpose of twisting the wires for things like the track power bus was to cut down on the space between the wires along the path by holding then close together.  Wouldn't bonding them into a flat cable do that, maybe even better.

And, if the twist is actually the functional aspect with respect to the issue(s?), why wouldn't twisting a flat cable be as helpful as twisting individual wires?