Hello Everyone

I did a search for this but couldn't really find anything specific.

A little background:
I have a PCB with 4 Teledyne SPDT High Frequency RF coax relays (click here for data sheet PDF).
The outputs of the relay are connected to various RF test equipment like a GPS tester.
The relay switching from closed to open is controlled by I2C ports and serial bus interface from a PC. So I can basically control the switches with my computer by sending highs, lows through the serial cable to the relay PCB

My dilemma:
The switches tend to get very hot after 2-3 days of continuous use. This causes the relays to burn out or become very unreliable at high temps and causes all kinds of problems in my entire system.

My question:
How can I test these switches on the PCB? I do not have the luxury of replacing them. An ideal solution would be to build an external circuit, connect it to the relay's using coax cable and do an automated test where it checks each relay if it is switching open/close. The final result can simply be a green LED that says everything is OK. The test system is going to be used by others less tech savvy, not just by me, that's why the result should be simple. Green or red LED.

Any help on this would be much appreciated!

 

The relays are latching types and are likely are not designed for the coils to be continuously energized (although this doesn't seem to be mentioned in the data sheet). That's why they are getting hot. You need to modify the control circuit so it only applies a short pulse (100ms minimum) to switch the relays.

You can build a test circuit that has an LED (with suitable current limit resistor) in series with each contact and a power supply. The LED will light when the contacts are closed.

__________________
Carl

 

hi Carl,
The datasheet suggests that the latching '#' is an option.?

Even if the relays are continously powered they shouldnt burn out, I would reduce the +V supply to the relay coils.

EDIT: Carl this is what I am referring to.

Attached Images

esp03 Oct. 27.gif (5.7 KB, 5 views)

__________________
Eric
"Good enough is Perfect"

PIC tutorials:
Gramo's: www.digital-diy.net/
Bill's: www.blueroomelectronics.com/

 

Hi crutschow

Thanks for replying.

They can be constantly energized coils as long as they are driven by the correct voltage in my case (24V). However, the reliability of the relays becomes really bad after prolonged use, not as much burning out but it just disrupts the measurements from the rest of the test system.

The way the PCB is currently etched already has LED's to show that the relay is switching between NC and NO, however this is controlled by I2C (serial connector) interface which is not user friendly at all. See schematic below:



The final solution that will test all four relays has to have a simple connection procedure and simple result so that anyone will be able to test the relays on the board.

What I had in mind was:
- Create a timer circuit in an external box that will connect to each of the four relays in parallel and then open and close each of the relays one at a time.
- Each time one of the relays has been able to toggle between an open and close, a green LED would light up for that one relay indicating a pass, red indicating a failed open/close.
- Repeat the process for the 3 remaining relays.
- Stop after all four relays have been tested.

This would mean I would need to use a 555 timer, four flip flops to latch a pass/fail, 4 green and 4 red LED's and a gate to store the results?

 

Don't see that latching is an option. The datasheet describes it as a "Latching" relay.

If you reduce the voltage, it may not be enough to reliabily operate the relay, especially if he's using them at high temperatures. Perhaps the relays should be able to tolerate continuous voltage application but, since he's having trouble with the relays overheating, I believe it's better to pulse them and then reduce the voltage to zero. Shouldn't be that hard to do if they are being controlled from a PC.

__________________
Carl

 

Sorry I had the wrong data sheet linked:
http://www.teledynecoax.com/pdf/coax...s/CCR-33FS.pdf

It's actually the NON-LATCHING one that I am using.

Also, they need to be energized for periods of time depending on which RF test equipment is running. For example a GPS tester requires a longer delivery time than just a burst.

 

is your application something that could benefit from forced air cooling? that may help the reliablility side a little?

__________________
There will come a day when PICs will rule the world!

 

The relays already have heat shields between them and are lying on thermal grease cooling pads. There are open and not enclosed so there is good airflow.

But anyway, if it gets too hot (stuff starts melting ) we just shut down the system but that defeats the purpose.
That's why it's better for us to test if the relays are actually functional before 'de-energizing' them.
If the test fails then there is nothing to do but replace the individual relay.

 

Are you running these test in a temp chamber? Op spec of the part is 65C, are you keeping the temp under that? The data sheet for the part sucks, and at $73.00 per relay, I would contact Teledyne apps engineers for help.

__________________
Pay it forward.

 

Agree. With two coils inside a single relay it is probably of a latching type design.

Drive them properly with a pulse and they will live long and will never get hot. All your reliability problems will disappear automatically.

__________________
L.Chung

 

A test chamber would be nice

But for the high RF frequencies and the small variations I get in power measurements, these relays provided the best results for the situation I am in.

The relays are of failsafe type, (which was my mistake for not mentioning ) not latching, so pulses would not work because some of the test equipment that the relays are switching current to need to be on for a length of time (3mins - 5 mins).

Also, the the test equipment don't turn on once and then the test is over. It does it a number of times every day (50 - 100 tests or more)

BTW, Thanks for all the help I'm getting, very useful stuff here.