Philips PM5326 RF Signal Generator - reed relay problems

[di_colourist]

I'm working on this Philips RF Signal Generator which has problems with the Reed Relays used for band switching. They're the sort with a visible reed switch in the coil.

Some of these quite literally fell apart at the slightest touch, and others seem to be borderline contact wise - air inside?? Not sure.

I tried the obvious thing and replaced several of the reed switches within the original coils as a test, but with a somewhat unreliable result. It worked, but not reliably so.

Are these relays still available from anywhere? if not, is anything available with the same form factor perhaps? I've never had to deal with these before.

They have a 5vdc coil and are driven directly from a BCD to decimal decoder.

Thanks!

Trevor.

[CambridgeWorks]

A picture with all dimensions would help.
I have some reeds I can sort through.
Rob

[di_colourist]

Quite right Rob! Here's a quick sketch to go with the picture attached to my original post. The pin spacing would appear to be 7.5 x 11mm connected as drawn. I could of course make a small adapter board and use different relays, but I'd like to at least try to keep things original if possible. It's a nice sig gen with built in frequency counter and it's PLL so its reasonably stable.I guess it was the last of its kind before everything went synthesised? It does have other issues too, but this is the biggest.

[Terry_VK5TM]

A quick squizz round some of the usual suppliers suggests you will be hard pushed to find anything in the style of the originals.

The majority seem to be in the style of an IC package - fully sealed.

I would just replace them with an adapter board and newer component, however, I would make sure that the supply rail and BCD decoder are OK before doing so.

[di_colourist]

Hi Terry. I'm sure you're right, and that's the avenue that I may well end up having to go down.

The BCD decoder tests good, and the supply is also good.

Jumpering across the affected reed switches (most of them!) brings the ranges back to life, so this, unfortunately, is definitely the problem here.

Trevor.

[CambridgeWorks]

Found these this morning, but far too big I imagine?
Rob

[eaglework]

Hi Trevor,
I went round this loop with one of these sig gens 2 or 3 years ago and failed to find suitable replacement reed relays.
In fact the ones I bought, which just about fitted, had really fragile glass to wire seals and broke with almost any attempt to fit them within the coils.
Still sat under the bench in bits.
If you decide to go down the piggy back board route I'd be interested in your approach.
Good luck.
Jon

[di_colourist]

Hi Rob. I think you're right - they might be a bit on the long side. Thanks for looking though - I really appreciate it!

[di_colourist]

Hi Jon. I did manage to replace three of the reeds, just as a test, and they did work after a fashion. I guess all reed switches aren't created equal.
I think I'll end up making a little sub panel at 90 degrees to the main board, and just replace the whole lot. There'e a few more one the board underneath, and I'd put money on at least one of them having gone south too. Looks like this might be a common problem on these gennys. I'll let you know what I come up with!

[Radio Wrangler]

In the original part photo, you can see the reed connections bent right up to the glass seal. This is a right no-no and pretty much ensures eventual failure. The wires have to be supported while being bent so no stress reaches the glass seal, you then get several mm of straight wire up to the seal.

Given the big loops as made, this isn't VHF stuff, so there's nothing wrong with 5v DIL packaged relays. In years gone by we did a lot of reliability testing with Panasonic and Takemasawa parts. Using milled housings to envelop the relay case it was possible to make a precision attenuator set good to 90MHz. What this did was create distributed capacitance to the internal loop from the common pin up to the contacts and back down, turning it into a matched transmission line. This is good for an attenuator, but not wanted for switching coils in a sig gen.

David

[Terry_VK5TM]

Those reed relays switch various coils in the VCO up 125MHz, so some of them are at VHF.

[Radio Wrangler]

I guess they just treat the inductance of the reed and those long leads as part of the inductor being switched.

David

[Terry_VK5TM]

They must be OK, so that raises the possible problem of what happens with a non-original replacement - will probably need a fairly intensive re-adjustment and calibration of all the frequency ranges.

[Trigon.]

If the problem with the replacement reeds was simply a lack of ampere turns, perhaps upping the relay supply to 12V (as per earlier units) would do the trick. (The 7445 BCD decoder outputs are good to 30v/80mA.)

Cheers

[eaglework]

Hi Trevor,
Good luck with your idea of a sub panel with modern DIL reeds - do let us know how you get on.
Jon

[Philips210]

Hello.

I have a small stock of reed switches but I'm not sure if I have anything that small. I will check later.

Incidentally, CPC stock a range of reed switches so may be worth a look.
https://cpc.farnell.com/search?st=reed%20switch&gs=true

Regards,
Symon

[di_colourist]

Hello all. The relays as originally fitted do indeed seem to have the (gentle) curve of the reed switch leads beginng a little close to the glass / metal pinch for comfort. I wonder if that contributed to the eventual failure of the part buy compromising the seal?
The reed switches I fitted to the original coils as a test almost certainly demand more magnetic flux than the original parts. I kind of proved this to myself by adding a small external bias in the form of a permanent magnet near the coils. This restored normal, reliable operation, so the suggestion of increasing the supply voltage a little could well work in practice. I'm undecided as to where to go next. I may try to source more "sensitive" reed switches (requiring less ampere turns) than the ones I fitted as a test, in fact the ones I put in were certainly of unknown parentage shall we say, and have no specs with them whatsoever. The only thing I was able to measure was the contact resistance which came in at 0.3 ohms - comparable with the originals that still actually functioned. I think I'll attempt to source some better (more sensitive) reeds, and in the meantime I'll lash up a daughter board with some more modern parts that I have to hand. I'll try it on the highest range first and see what happens. (I'm a bit cocerened about stray capacitance) Depending on the result of that test, I'll either lash up the others (10 in total), or fit new reeds throughout. I wasn't brave enough to put the curves in as per the original equipment, and I soldered on very short pig tails instead. I'll let you all know what eventually happens. Trevor.

[Philips210]

Trevor,

I've found some small reed switches that may be suitable. What is the diameter and length of the glass envelope for your switches?

Regards,
Symon

[di_colourist]

Hi Symon.

The originals are (approx) 12mm x 1.5mm

Thanks for looking for me - I really appreciate the help.

Even the ones that currently work are somewhat borderline and most likely heading South.

I double checked the +5v rail and it's bang on the money - I haven't tried tweaking it up yet as I need to look at the circuit first to see where else it goes. I might try an external supply when I get a moment, but the day job keeps getting in the way as is always the case!

Trevor.

[Philips210]

Hello Trevor,

The smallest reeds that I have may be suitable. Please see attached pic of the three versions I have:

On the left of the pic the smallest are 11.25mm long and 1.75mm diameter

In the middle, they are 14.75mm long and 2.2mm in diameter and

on the right, 18.85mm long and 3.0mm in diameter.

Unfortunately, I don't have any more than what's in the pic. I have several larger reeds but they won't be suitable.
If any in the pic will be suitable then I can put them in the post. PM me if required.

Regards,
Symon