Relay Switching at Signal Level: Experiences and Ideas Please.

[Snappa1]

I am starting the conceptual design of a communications reciver. It will have a valve front end and IF, with semiconductors after that.
In the olden days, I would have used ceramic switches for bandswitching, IF Filter switching, mode switching, etc., etc. However, today, such switches are not readily available (and expensive), and the way seems to be to use small relays.
By my rough count, I might need 30+ relays, which at, perhaps £1.50 each, adds up to a lot of money. Some examples use exotic, expensive relays, others seem to use 'industrial' relays. Before I dive in and spend, I thought I'd ask for people's experiences with signal level switching with relays.
Good ? Bad ? Type of relay ? Contact material ? Wetting current ? Good and bad model numbers ? Where to buy ? etc., etc.
It's a long term project, so I won't be buying immediately, but once I know what I'm after, I'll keep an eye on the market.
Thanks for all your ideas.

[swordholder]

Have you thought about using switching diodes? I used them successfully on a small receiver I built covering the amateur bands using the Plessey/G4CLF board.

Mike

[ms660]

Switch wafers run out around £4 + vat each.
Switching loads of coils etc designed around valves with relays might be a bit of a headache, stray C, feed back, birdies, reliability etc.

[Tim]

Cmos switches? Used them once in a mixer.

[M0FYA Andy]

One concern I would have with relays is that the contacts generally don't have a wiping action to keep the surfaces free of oxide. In the aircraft industry I have recently retired from, it was always a design aim to have at least 5 mA 'wetting current' through a relay contact that was otherwise just switching a logic level, to ensure the contacts didn't oxidise.
Andy

[GMB]

Transceivers from the 70's seem to make much use of reed relays packaged to look like a DIL IC.

Reeds have the advantage of being sealed so don't suffer from contact oxidisation.

[Snappa1]

Keep the ideas coming ! Thanks.
Switching diodes; I've read about IMD introduced by these. Other homebrew receivers often use relays, and I am trying to get a better fix on which types to use.
Reed relays. Sealed, yes, great. I haven't looked to see if they are still available. What contact material should I be looking for ?
Switch wafers at £4. Are these phenolic or ceramic ? Do you have a source ?
Small relays allow good IO separation which should help prevent stray coupling, etc.

Is there anyone with actual experience, either good or bad ?
Thanks.

[M0FYA Andy]

I have bought from this company before - http://www.blore-ed.com/
Not their switches, I was buyng potentiometers, but they were very helpful and more than happy to sell a small order to an individual. It might be worth an enquiry about their switches.
Andy

[neon indicator]

FT102 is example of wrong kind or relays. Rubbish compared to FT101ZD.

Use mix of multipole switches (Cheap and good NOS from Russia/Ukraine, also new available), diodes (1N4148, PIN, Schottky etc), CMOS switches and sealed relays. For battery powered gear you can get bistable type that don't need power except to change state.

Current FT817ND uses relays, CMOS switches and diode switches.

[kalee20]

I wouldn't use switching diodes, CMOS etc in a valved front end. Apart from irrational ideas like not wanting these silly holes anywhere in my equipment - or impure sand extract - there is the (very remote) possibility of occasional flashover, and the (rather more likely) possibility of test prods slipping and putting momentary high voltages on the switching. Bye-bye CMOS!

Relays, to me, are the way to go - and at that, sealed cross-point contact types or reed relays. You have the advantage that the relays can be sited where you want them, so there should be less chance of stray coupling than if you'd used a wafer switch where you have to bring all your signal wires up to the switch.

[Snappa1]

Thanks Kalee20.
Those are the arguments I have been reading. Can you give me any part numbers of suitable relays ? From what I can see on the 'net, people seem to be using fairly standard relays. Or are they ?

[Skywave]

Use of F.E.T.s as switches? I once used this approach in an audio mixer.

Al.

[trh01uk]

Relays are fine for small signal switching, but you definitely need a wetting current to keep the contacts clean.

Don't assume a hermetically sealed relay will get around the problem - it may for some years - but having had to redesign the Larkspur C11 transmitter aerial switching relay circuit to introduce a wetting current I can definitely say that the hermetically sealed ones fail just like all the others. Notably on that circuit, the transmit contacts were just fine (because of the high current they were passing), but the receive side went very lossy. All it needed was a few mA from a local supply to sort this out.

Whether you use diodes or relays is really a matter of personal choice. Bear in mind the need to consider the loss you can stand when ON, and the isolation you need when OFF. Bias networks can easily compromise these figures if not very carefully designed over the full frequency of operation.


Richard

[ms660]

Switch wafers at £4. Are these phenolic or ceramic ? Do you have a source ?

Not sure, but Farnell do them, made by NSF. I think RS probably do them.
To find them quick at Farnell just search...wafer switch...

Cheers.

[Herald1360]

ISTR mercury wetted miniature reed relays are available- that should sort out any contact wetting problems........

[Snappa1]

Thanks for all your further comments. Not so keen on mercury; banned for commercial use, and very expensive. FETs; yes, where signal levels are known, but at the front end of a receiver, strong signals could cause overload/IMD.
Is there no one who has actually used relays the way I want to ? It seems that all of you who have replied above (thanks) have good ideas, but none of you have actually used relays like this, and can give me a part number / supplier.
Keep the ideas coming, please. Or am I in the wrong forum ?
Thanks again.

[neon indicator]

I have used relays, sparingly as you suggest. Sealed type about 8mm approx high and DIL pitch of pins. Single, dual and quad pole change-over depending on application. (2 x 1 pole and 1 x 2pole may suit layout better than 1 x 4 pole, though uses about 3x power). I use reed relays (N.O. type) occasionally. They need no wetting current. I avoid too many relays in battery powered gear apart from latching change over type.

A hybrid approach of multipole switch and then relays for signals ideally not brought to switch. Of course an R/T always uses at least one relay and often two as the there is the single PTT line from microphone or the morse keyer. Often one relay for power lines and a different one for aerial/signal lines.

There are so many relays depending on your relay supply voltage and the nature of signal to be switched. Switching power (i.e. the supply line at 500mA to 50A) doesn't really need a sealed relay, and will need larger type than signal (not usually DIL footprint).

[GMB]

DIL reed relay units are available from Farnell.

I have only fixed transceivers that use them rather than design from scratch.

[Ian - G4JQT]

The Elecraft K2 is a top-of-the-range amateur band transceiver kit from the USA. It uses relays. The manual is available here (.pdf) for downloading.

I'm sure the manual will name the types of relays used that are suitable for use in the signal path.

Regards,

Ian

[dseymo1]

How about a hybrid approach, using reed switches arranged on a rotary switch frame, actuated by a rotating magnet?