Need advice on switching bandpass filters!

[WB6VHK]

I have a single conversion hf transceiver I am trying to install bandpass filters in. I have built 4 "dual bandpass filter" boards, 80/40 30/20 17/15 12/10 and want to switch them via 2 9 position rotary switches. Is this acceptable ? Would relays be better or perhaps diode switching? This is something I have never done before and I can't seem to find much on the web, I would love as much input as you care to give me! Thanks, Darrell, WB6VHK

[VK2TIL]

Rotary switches and their associated wiring can be a terrible pain when dealing with RF; just look at the complexity of older multiband radios.

I use relays; DIP-style mostly for low-power switching (IF filters). My experimenting is at the milliwatt level so these are fine.

Larger devices are required if there is much power to handle. Here are some examples;

http://www.futurlec.com/RelDip.shtml

http://www.minikits.com.au/doc/section15.pdf

Reed relays are also useful.

For low-power applications, fast bus switches have been used by some experimenters;

http://focus.ti.com/lit/ml/scdb006a/scdb006a.pdf

I have some but haven't used them yet.

PIN diodes are useful but expensive; ordinary 1N004s have been used in their place.

But I think that relays are simplest.

[VK2HHS]

I have a single conversion hf transceiver I am trying to install bandpass filters in. I have built 4 "dual bandpass filter" boards, 80/40 30/20 17/15 12/10 and want to switch them via 2 9 position rotary switches. Is this acceptable ? Would relays be better or perhaps diode switching? This is something I have never done before and I can't seem to find much on the web, I would love as much input as you care to give me! Thanks, Darrell, WB6VHK

Not having done as much research as VK2TIL, I can't go into as much detail however due to a current project here, I've found it helpful to see how others approach the problem.

The following sites may help you.

http://www.cqham.ru/projects/100w_hf_qro/100w_hf_qro.htm

http://yb1zdx.arc.itb.ac.id/data/orari-diklat/homebrew/yd1jjj/irf510.pdf

I've currently got the following on the workbench as a longterm project. It is PIC controlled but the bandpass filters are also relay switched. You might have to do quite a bit of reading to find the relevent info though.

http://myweb.tiscali.co.uk/qrp2004/

de VK2HHS

[K9STH]

For decades manufacturers like Collins, Hallicrafters, National, etc., used rotary switches to select the filter. Those work fine! Especially in the case of Collins Radio they usually, but not always, shielded the switch sections that select the particular filter.

Usually the inputs and outputs of the various filters are switched by separate wafers on the same shaft. However, there are examples where a single wafer was used.

Glen, K9STH

[K9FV]

I use relays; DIP-style mostly for low-power switching (IF filters). My experimenting is at the milliwatt level so these are fine.

Larger devices are required if there is much power to handle. Here are some examples;

http://www.futurlec.com/RelDip.shtml

But I think that relays are simplest.

Hello Kerry, first - thank you for providing futurlec info a few months back - I've ordered 3 or 4 times from them with good success.

Will any of the relays (maybe 10 amp?) shown on the Futurlec page handle a KW up to 30 mhz (or 21mhz)? 500 Watts?

Will the 1amp and 2 amp will handle 100 watts up to 30 mhz?

How much power do you think the 5 amp latching relays shown at bottom of page will handle?

Thanks to all for any info.

73 de Ken H>

[AI3V]

Rotary switch, relay, PIN diode, They all will work ok.

The relay can be pretty small, most commercial rigs use the 1 inch guys- they are not making and breaking current, just passing it so the usual ratings do not apply. (see QSK and hot switching!)

But whatever switching method you choose, you can test it by leaving, say, the 20M filter out and measuring the isolation and feedthru in that position.

It would be a shame to have a 70DB filter, in a circuit with 40DB of leakage.

Rege

[VK2TIL]

I'm always happy to give Futurlec a plug; I have been very happy with them.

I'm not a "QRO" man; as I said, I play with milliwatts, not kilowatts.

Like this;

http://img412.imageshack.us/img412/6319/c003.jpg (http://img412.imageshack.us/my.php?image=c003.jpg)

A pair of IF crystal filters, switched by a DIP relay at each end; the red wire at each end provides switching DC via a feedthrough cap. (Still under development; there is no filter on the other side; its design has proved to be difficult and the umpteenth prototype is still on the "development" board). :)

But it's just a matter of current capability, assuming, as 3V said, that there is no "hot-switching".

In a 50-ohm environment, a kW produces 4.5 amps; 100W produces 1.4 amps.

(Although there will always be reactance, mismatching etc, simple Ohm's Law sums are good enough for most estimating; I don't think there's any need to get complicated).

The MiniKits page I linked-to earlier contains some recommendations on power handling for some of the relays he sells.

And, as 3V also said, the size & type of relay used in commercial radios is a good guide.

Leakage can be a problem with an unknown relay; it can be checked by putting some RF into the Gozinta port and seeing how much leaks from the Gozouta port with the contacts open. This would be a wise test on a single sample before ordering a heap of relays.

I imagine that leakage was a potential problem when rotary switches were used to switch RF; manufacturers probably spent some time testing their often-complex wiring layouts. And the switch was often located close to the circuits being switched and operated by a long shaft that then created mechanical layout challenges.

All-in-all I think that relays are simpler for the homebrewer these days.

[K5UOS]

Kerry,

That is some pretty "ugly"!

What are you building and what frequency is your filter designed for?
Are you using L Networks I/O?

That looks like very thick PC board. I have some that is thick as a hefty book cover. I use it for VFO's but it is tough to cut and needs a little more heat to solder things to it. I have some almost as thin as paper for shielding.

Your consruction looks real nice.

Don K5UOS

[KA8RAW]

Most modern transceivers use small relays to switch in and out bandpass filters, probably the best and easiest way.

[VK2TIL]

G'day, Don.

I'm building a Measurement Receiver; it's inspired by the EMRFD one but it's a bit different.

One use will be noise-figure measurement so I want well-defined filter bandwidths & shapes.

The filter shown is a 500-Hz Gaussian-to-6 at 5MHz; it's really good.

Sadly, I have had lots of trouble getting a wideband filter with a decent shape. The problem is caused by low crystal Q and I have trouble getting good crystals here.

Plan B is to change to 12MHz; wider bandwidth filters are easier to implement at higher frequencies. The narrow filter is so good that I'm reluctant to disturb it; different IFs for narrow and wide could be used as the "VFO" is a signal generator.

The PCB is just ordinary stuff; the photo does make it look thick but perhaps it's just the angle.

There are L-networks at the ends; you can see the toroid Ls and the ceramic trimmers for peaking.

The project is on the backburner at present but it may be revived some day. :)

[KA5S]

FWIW, the Ameritron ARF-1000 (www.ameritron.com/Product.php?productid=ARF-1000) uses what looks like the relay-switched filter board from the ALS-600 (http://www.eham.net/reviews/detail/603) amplifier, switched by 12 VDC DC and a 6-position switch.

(Off topic: Probably a good spur tamer for various amplifiers we don't talk about. )

Cortland
KA5S