1. Home
  2. Electronics Design
  3. EMI leaking down shaft of rotary switch

EMI leaking down shaft of rotary switch

I=92m using Grayhill series 50 rotary switches for low level signals.

formatting link

Every once in a while, the rotating shaft loses contact with the grounded switch body. When this happens I get electro-static fields in the room leaking into the circuit. (The circuit lives inside a metallic box.) It is only the high impedance part of the circuit that appears susceptible. I was able to measure shaft to ground resistances and correlate large resistance (1k ohm to 10 k ohm) with EM leakage. Typical resistance values were less than 1 ohm. So I=92ve contacted Grayhill and they are preparing a quote on switches with a different shaft to body sealing option. But these are special order mil-spec parts and I=92m afraid the price will match. There is a blurb on the first page about positive shaft grounding for emi shielding, but I found that in a random sample of new switches on hand about 1/2 showed 1 k ohm shaft to body resistance with a bit of wiggling of the shaft. And I worry that the rest may =91fail=92 with a bit of wear. These switches will get a lot of work.

I started thinking about making my own grounding contact to the shaft. There are probably hundreds of ways to do this. My =91best idea=92 so far is a conductive knob, perhaps nickel plated, and a conductive washer (phosphor bronze?) that sits under the knob and has some sort of fingers that make contact with the bottom of the knob.

Searching the web proved fruitless, but as always you have to know what to search for. It seems like someone must have already had this problem and perhaps has a solution to sell me. Any thoughts, ideas, links or clues will be greatly appreciated.

Thanks, George Herold

Reply to
ggherold
Loading thread data ...

Wavy washer maybe. Or EMI grounding finger stock.

Maybe wrap a pallpoint-pen sort of spring around the shaft, and anchor it to a screw nearby? All stock parts.

That will side-load the shaft, which could be a problem if you don't have a panel bushing.

Does the shaft rotate continuously? If not, ground it with a wire! Wrap it a few turns around the shaft to reduce wear.

John

Reply to
John Larkin

How 'bout a spring-loaded graphite brush, as stolen from a motor? Those are cheap, low resistance, and should last forever.

Cheers, James Arthur

Reply to
James Arthur

=92ve

b
,

"Wavy washer maybe. " Yup I thought about those. I don't know if they are conductive, but I'll check.

I also thought about a simple wire, but this is not just for personal use but for a product and I'm not sure the wire will make it by the Aesthetic Czar (my boss) it not only has to work it has to look good too.

George

Reply to
ggherold

t

=92ve

rb

s,

Graphite brush onto the outside of a conductive knob? (not too much room left on the shaft once the knob is in place.) It will certainly work, but it's got more of the uglies than the wire solution.

But thanks,

George

Reply to
ggherold

"Wavy washer maybe. " Yup I thought about those. I don't know if they are conductive, but I'll check.

BeCu wave washers. Simple & cheap:

formatting link

Reply to
Giblets

I meant on the shaft, bearing on the non-flatted portion.

A conductive knob might make an unfortunate antenna; an insulated knob and/or shaft extension might fix the whole problem.

Or, if the knob is metal and close to the bushing, John's wavy washer, by all means.

Cheers, James Arthur

Reply to
James Arthur

The EMI leakage must be rather high in frequency, as the (semi)-coaxial waveguide created by the sounded slot has a high cutoff frequency. Try a (thin?) brass tube, the longer the better - to maky your own high cutoff waveguide, as well as a few pf SMT bypass caps on affected inputs.

Reply to
Robert Baer

If they usually work, how about running a dc current through the ground and gating the measurement off when you see a voltage pulse? Sort of like an IF noise blanker in a radio.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net
Reply to
Phil Hobbs

Thanks for the link, I'll see if I can find any in the McMaster- Carr.

George H.

Reply to
ggherold

s

hat

I=92ve

r

lurb

.

st

is

eas,

"A conductive knob might make an unfortunate antenna; an insulated knob and/or shaft extension might fix the whole problem."

James, I think the shaft is already acting as an antenna when it intermittently losses contact with ground. A conductive knob will make this worse... unless I can ground the knob. I guess the only reason I was thinking of a conductive knob was that would give me more area to make contact with.

George H.

Reply to
ggherold

How about a plastic knob with an internal metal bushing; between the bushing and the panel, put a coil spring or a wavy washer or a donut of that conductive EMI brillo pad sort of stuff.

Or go to cold switching.

John

Reply to
John Larkin

t

=92ve

rb

s,

Robert, Most of the interference is coming from the overhead fluorescent lighting. Dominant frequency is about 25kHz. I think the only reason I see the interference is because the fields in the room are so large compared to the tiny Johnson noise voltages I=92m observing. The interfering signal is only perhaps 10% of the Johnson noise. But of course this number will depend on the environment.

George H

Reply to
ggherold

t

=92ve

rb

s,

t -

Hmmm, yeah I could just turn on an LED when the shaft looses contact with the ground. But the only way I can sense that is to attach a wire. In which case.....

I was thinking about John=92s wire solution last night. I should be able to design a simple knob to hide a few turns of wire inside of it. I don=92t mind the wire solution, if there are problems the user can install a new wire. I was really hoping that there already existed some type of grounded knob that I could just buy.

George H.

Reply to
ggherold

ls.

elds

t that

th

=A0So I=92ve

a

rder

a blurb

,
1/2

the

ear.

=91best

has

.

this

ideas,

"> Or go to cold switching."

Yeah, that's the obvious solution. I really makes life a lot more difficult. I've got a ten position switch. I'll need ten little SPST relays, I've got to power them, and then there is always the added capacitance. (Hmm, well the capacitance may even out since I'll be losing the C in the switch.)

George.

Reply to
ggherold

How far along is the design? Can you use the switch to operate a transmission gate so the switch is isolated from the signals?

The spring idea might not be so far off. Back at Motorola we had a cheap radio that would misbehave on the MIL 810C vibration tests because the volume and squelch knobs would turn on their own. One of the mechanical engineers patented a "anti-rotational friction device" which was nothing more than a spring that went around the shaft a few turns and put side loading on the shafts.

Can you get a similar switch with plastic shaft?

--
Joe Leikhim K4SAT
"The RFI-EMI-GUY"©

"Use only Genuine Interocitor Parts" Tom Servo  ;-P
Reply to
RFI-EMI-GUY

Right, obviously. Clearly I've not been communicating well; I guess I've been too terse.

Since you suspect fluorescent lighting as the EMI source, the pickup's likely electrostatic. That's why I thought replacing that metallic-shaft-antenna with a non-conductive extension might be fruitful. You can't easily replace the original shaft, so I suggested extending it with a non-conductive extender.

You could, in the extreme, recess the switch behind the front panel, and actuate it via a fiberglass extension.

Maybe pictures will help.

FIG. 1 ======

fiber .-. shaft | | | \\ | __ .--. | -------------' '--| |- 3a. (threads) | | | | | | | '-------. | | '--- |

Reply to
James Arthur

Ahh Excellent. Push the shaft to the side so that it always makes contact. That might work. (Perhaps this is what John L. was suggesting earlier and I was just too dumb to get it.) A plastic shaft would work also. I've always disliked plastic shafts, (They spell 'cheap' to me.) but maybe I could be persuaded.

George H.

Reply to
ggherold

e

Wow, Thanks James. I get your idea now. I just got off the phone with the sales guy at Grayhill. (I talked with the engineer yesterday.) And they are proposing exactly what you drew. (I think) There is a bushing on the rotating shaft that 'should' make contact with the threaded portion of the switch. The bushing is press fit into place and this task sometimes must leave a bit of a gap. They are going to add a conductive wave washer between the bushing and the threaded part to make positive electrical contact all the time. (At least this is my understanding of what they propose.) So now it has to go through their channels and get back to me with a cost estimate. This sounds like the ideal solution and if they are cheap enough we can use them in other locations also.

Thanks to all of S.E.D. for the wonderful ideas. I'm going to put this problem aside until I hear back from Grayhill.

George Herold

Reply to
ggherold

I figured that the spring itself, wrapped 180 degrees around the shaft, would be the ground. I wish we could do sketches here.

anchor////////////// ( )/ (shaft)/ ( )/ anchor//////////////

The anchor could be a screw or a ground lug.

You could do two, in opposite directions, if you don't like the side loading.

A plastic

That would sure be easier.

John

Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.