[SOLVED] 30-520MHz Relay switched Filter bank buzzing like humming bee???

[ahsoopk]

Dear all,

I have design a filter bank, consisting of 6 low pass filters with different cut off frequencies. Each filter is sandwiched in two relays. Relays have 50ohm contact resistance. problem is when I connect my 50Watt PA at the input of this filter bank, some filter relays start buzzing and I have lost couple of PA due to this problem. Tuning the Filter under large signal conditions doesn't help either.

Any ideas please...


Many thanks

Malik

 

[FvM]

Relays have 50ohm contact resistance

Unlikely, you mean 50 ohm transmission line matching. Are they also specified for 50 W load?

problem is when I connect my 50Watt PA at the input of this filter bank, some filter relays start buzzing and I have lost couple of PA due to this problem.

Most likely reason is that the control circuit is picking up RF and running wild. Won't happen if the circuit designed for RF power handling. Or (less funny explanation) you have huge impedance and isolation breakdown in the filters or relays.

 

[vfone]

This could happen when after the filters, the PA use a VSWR protection circuit, and could appear a feedback loop which can make the circuit oscillate (at the "buzzing" frequency of the relay).
Adjusting the RC constant of the VSWR feedback circuit would eliminate this issue.
Other possibility (very rare indeed) it happen when the power supply of the PA cannot support high peak DC currents (if the VSWR is high at mismatch output). Very easy to find using a scope on the DC line.

 

[ahsoopk]

"Most likely reason is that the control circuit is picking up RF and running wild. Won't happen if the circuit designed for RF power handling. Or (less funny explanation) you have huge impedance and isolation breakdown in the filters or relays."

Some of them work perfectly fine and others don't, all the DC control lines to relays are properly decoupled and rest of the control circuit has been decoupled as well. but still this happens, looks like that incident power and reflected power are creating a huge standing wave that is getting coupled to relay coil and then causing this contact make and break buzz even though the coil's energizing end is RF decoupled.

But how can I overcome this beats me.

"This could happen when after the filters, the PA use a VSWR protection circuit, and could appear a feedback loop which can make the circuit oscillate (at the "buzzing" frequency of the relay). "

PA doesn't have VSWR circuit.

 

[volker@muehlhaus]

creating a huge standing wave that is getting coupled to relay coil

Then it should disappear if you insert series inductance into the relais coil path, making it high impedance for RF.

 

[chuckey]

Do de-energised relays buzz? I find it incredible that an iron cored inductor will generate a low frequency varying magnetic field when subject to a few tens of volts at 30 MHZ or higher and coupled to it by 10 Pf or so. Either as said, the RF is getting into some electronic circuit and causing this effect, or perhaps you have got your earths mixed up and have a bad connection that is rectifying the RF. In any case put a 1NF disc cap across each relay coil, that will eliminate any RF.
Frank

 

[ahsoopk]

perhaps you have got your earths mixed up and have a bad connection that is rectifying the RF. Frank

i would really appreciate if you could slightly explain it more for a relatively newbie.

Secondly this happens above 250MHz, below that it doesn't happen.

Many thanks for your help guys..

 

[D.A.(Tony)Stewart]

One can make a buzzer out of small signal relays @ 300Hz using a SPDT where the energized NC contact opens the coil current.

Since relays at least 30% hysteresis, your design must have a great amount of radiated RF energy getting back into the coil being rectified to DC to de-energize and cause the buzz.

Try Ferrite donuts around coil + wire with appropriate shielding.

Where is you schematic and layout picture?

 

[chuckey]

The standard technique would be to use the chassis as the RF earth then float the relay coils with their earth connected back to the power supply or low side switch. The relay coil wiring should be decoupled to keep the RF in the right place, ferrite beads on the leads would help at these frequencies.
It is a known fact that an oxidised connection (earth tag?) can rectify RF. If you have high voltages you could be suffering from some mechanical vibration due to ceramic caps changing their shapes (or flashing across) or even a singing arc, which I have heard from an aerial tuning hut for a 30 KW MF transmitter, we reduced the power so the arc went out and called up the aerial specialist!
It could be interesting to put a small piece of paper on the blades for the contacts to see if it falls off or even if it damps the noise. Are unused filter sections disabled?
Frank

 

[biff44]

I don't think its any of that stuff. I think the rf is screwing up your DC driver cirtuit, maybe a low frequency oscillation. put an AC coupled scope probe on the driver outputs and see what is there.

 

[ahsoopk]

Dear All,

I added 15nF caps to the both ends of the energizing coil (its a single coil relay), I also added a 1.2uH RF choke to the 5V supply that energizes the relay and added a 1k resistor to ground between the control line coming from the uP and 2.7k series resistor at the base of the npn transistor that controls the switching between the two paths of the relay. Doing all this has stopped the buzzing in this particular Filter bank and shall implement it in forth coming boards to check if this is the right fix.

But I do agree with you guys that its the RF travelling back in to the DC circuit and making it fluctuate, as I have tested this with multimeter while relay is buzzing and noticed that energizing voltage drops down to 2V from 5V, I suppose it is fluctuating but too fast for multimeter to detect.

Thank you all and shall let you guys know when have done few more units with same mods.

Kind regards

Malik