Crystal problem

[sentinel040]

I was trying to think of a clever witty title but failed miserably.

I like to use old radios as well as just “fettle” them, one in particular is an FM Pye Bantam, hardly cutting edge in terms of power and sensitivity but I always wanted one as a teenager, but they illuded me then. Now I have one on 2m (and a number of others…), but the 44MHZ RX crystal for 145.5 has always sounded a little ‘rough’, almost as if it not quite on channel, even when it was, or the discriminator was not centred properly, when it was. This being the case whichever position, or radio (I have a number of Bantams now) the crystal was tried in. Then come this summer and it refused to oscillate, being as dead as the proverbial Norwegian Blue; again, it is the crystal and not the circuitry around that particular socket. The circuit of the local oscillator for the HP1FM is hopefully attached, it uses the crystal in series mode selecting the third overtone. As an experiment, trying the crystal in my Altai dip meter, which uses a Colpitts circuit, shows it oscillating on its fundamental frequency.

The question now is whether there is a way out of this as I suspect being able to find a HC25U 44MHZ crystal for 145.5 receive is rather low, verging on the non-existent and I do like to use this radio.

Ian

[Station X]

I assume the crystal is 44.933333MHz?

(145.5-10.7)/3.

[Radio Wrangler]

All crystals have all modes. A fundamental pair of modes, one series, one parallel. A third overtone pair of modes, one series, one parallel. A fifth overtone pair and so on.... you get the pattern.

The third overtone modes are near but not exactly three times the frequency of the fundamental modes. The fifth overtone ones are near but not quite exactly five times the frequency of the fundamental ones.

So when someone orders up a crystal on 44MHz series mode third overtone and usually specifies a load capacitance (series capacitor for a series mode crystal) that is the mode that they grind the crystal to plant on 44MHz. The crystal can be used on any of its other modes, but the frequencies will be a little off what you'd expect from the simple overtone numbers.

A good crystal manufacturer can do a few things to try to make the specified overtone mode a little more active than it would otherwise have been. But they usually can't make overtones more active than the fundamental. So the oscillator circuit needs an LC tank to kill its gain off at the fundamental frequency and at the other overtones. With cheap crystals they don't usually try much anyway.

So an overtone oscillator relies on a tank circuit to force correct overtone operation.

Some people see that tank and think it's a frequency trimmer. Move it too far and oscillation can jump onto other modes, often spurious modes which crystals also have.

Your crystal doesn't sound like a happy one. maybe a little more bias current on the transistor could persuade it to feel happier, but have a look for the overtone tank and have a play. It should really be a level peaking adjustment and there ought to be series capacitors with each crystal to tweak the frequency of each crystal.

One nasty effect is that the pullability (by series capacitor) is inversely proportional to the square of the overtone number compared to a fundamental on that frequency. So a third overtone has only a ninth of the pulling range of a fundamental on that frequency. This might not be enough to correct the manufacturing tolerance of a crystal. In this case, you're out of luck. Trying to pull far often stops oscillation or gives errtic operation.

Not the best of news, I'm sorry, but I've been designing crystal oscillators for a lot of years and spent time at a quartz manufacturer working to get some awkward ones manufacturable. I don't think I've experienced all the issues they can give, but I've had a fair selection dumped on me. There's a fair bit of pressure when a production line is stopped as everyone waits for you to find a solution. This usually happens when the folk on the line have already tried swapping crystals and found that all the batches in stores won't produce the goods. Oh, and decent crystals have a 12 week ageing process during manufacture...

David

[Jon_G4MDC]

I would be tempted to try that crystal in something else. It could be a case of sleeping sickness. A search for that term on here might turn up something useful. A lot has been said.

I checked my stocks. For 145.4 I could help. Sorry, sadly no 145.5

[sentinel040]

Good afternoon and thank you both for your replies.

Graham, yes you are correct, low side injection and divided by three.

David, I was aware that crystals have a variety of modes and a propensity to sometimes behave in what appear to be weird ways; however, not having a brain the size of a small planet, I unaware of all the small print! I did notice that when it oscillated in the dip oscillator the fundamental output was not exactly 1/3 of the stated overtone frequency, so that is explained, thank you.

The receive crystal trimming, as per the circuit extract, is by a separate series inductor for each 'rock', something most Pye receivers of the seemed to do, and I have tried adjusting both that and the tuned circuit(s), L9 and L10, in the transistors collector. I have also tried the crystal in another 'spare' Bantam. However, I will revisit the points raised regarding tuning and the transistors bias point over the weekend and see if I get anywhere with it. Fingers crossed....

Ian

PS David, if you need a monorail signalling system layout, let me know. ;-)

[sentinel040]

Firstly, Jon, I don't know how I missed you reply yesterday, perhaps I failed to refesh the page; regardless, thank you for looking, it is appreciated. I will try to explore further on the matter you mentioned, a vague distant memory is what prompted me to try the suspect crystal in the dip oscillator.

In between carers 'duties' I managed to sneak another look this afternoon. Whilst away from home I am rather limited in test equipment terms, as I cannot lug the workshops contents around, neither is there space here. I just pick up a couple of random projects when home and my tool box. What I do have with me is a little handheld Thandar DFM and one of those little Chinese handheld 'scopes, which as an indicating instrument has plesantly surprised me. (The old and much repaired Tek 465 does not need to worry though.)

What I have noted is that the faulty crystal (S20, or whatever these days) seems to be doing something but at a very low level and apparently on totally the wrong frequency, this is when measuring the frequency/waveform on the first local oscillator transistors emitter and nothing gets past the collector tuned circuits, no matter how they are adjusted. As David alluded to, perhaps it has failed into some odd mode. The 'scope, which has a limited frequency display as well, indicates circa 71.4MHz and the DFM circa 51MHZ, but I suspect the output is so low that neither can count properly. When switched to a good channel, you obtain the correct 44.whatever MHz frquency.

I have played with the biasing, as suggested, and can increase the ocillators output by up to about 10%, when one of the two good receive crystals (S21 or S23) is selected, but it did not help the faulty one. I have also repeatedly been through the tuned circuits in the oscillators collector to see if I could find a 'sweet spot', but again sadly to no avail.

Time to search for the fabled Unobtanium mines (given past crystal searches).

RIP Bantam.

Once again, thank you all.

Ian

[Jon_G4MDC]

In your place I would be looking to get one of the RTL SDR dongles which run from a PC on Windows or Linux. Coverage of the early ones is ~27-1000MHz or so and they work really well. Later ones had access to the internal IF and they did HF as well although funny things happen between 14 and 28MHz due to aliassing.

They make a really useful poor mans spectrum analyser for looking at what oscillators and the like are doing. They are even more useful for tuning up multiplier strips to the right harmonics. If you put one on a USB-A to USB-B extension lead they are small enough to use as a probe inside equipment you are working on. I think my first one was about £6 and now I have 4 or more. Some are used as test equipment others as band monitors. There is one permanently watching 10m for DX.

As for the crystal don't give up! You just need to find someone with a scrap TR2200 or such. They had 45MHz S20 HC25U fitted as standard. I have one here but of course don't want to wreck it - still works perfectly after 45years. My first rig.

I bought an Icom IC215 at a rally for £20. It had 11 channels fitted! If you can buy an old TR220 for that kind of money it will be a lot less than having a new crystal made. The IC215 crystals are 15Meg fundamental so would come out "off" at 3rd overtone.

I never did work out how the TR2200 managed to get away with no RX frequency pulling inductors/capacitors. There aren't any but it seems to work well enough. I suppose they controlled the crystal production well instead..

I forgot to add that there is a seller on a well known market place advertising crystals for Icom / Trio etc. I checked his list from sometime last year and sorry - S20 not found. What I downloaded and sorted by spreadsheet is quite old now I suppose - he could have added some new ones.

[sentinel040]

I was mulling this over in bed last night as I fought to get to sleep. I realised that I had missed an auction for a pair of handhelds that may have been useful. I am aware of some of the multiplicity of equipment that used the 44Mhz range, in point of fact the S21 and S23 receive crystals I have fitted are new old stock Tandy scanner crystals that I bought many years ago, in fact I bought a load of those frequencies cheap when they were closing down, but have given most of them away over the years. So, it may be time to keep an eye open on the shark infested waters of evilbay and knowing my luck find a radio stripped of its crystals.....

Back at home I have a (proper) TinySA, so it may be worth either using that when I next pop back, or bring it with me next time. I have tried Dongles in the past but ran into driver problems, an area that I am not very good at dealing with.

I am aware of the chap you mention, Paul if my memory serves me right, I do find his lists a bit of a nightmare but I understand why they are as they are; he is a decent chap who I have used many times over the years and did check with him a few weeks ago, when this problem first appeared. Vincent is another source who helped me with the crystals for the Hudson I rebuilt a couple of years ago.

I had a TR2200GX that I regret selling when I bought a Kenpro named IC2E clone and also wondered at the lack of trimming, but then again I did not have it long enough to see how crystal drift with age affected it. One of the transmit crystals in my Garex Twomobile has gone a full 12.5kHz off its original over 45 years.

This Bantam seems to have fought all the way, unstable receiver, a long list of faults, the local-ish repeater I bought crystals and fitted a small CTCSS board for going over to the dark side within a couple of months, etc, etc. So this is just the latest problem it has thrown at me.

Think positive Ian! Time to try to set up a complex evilbay search.

Thank you.

Ian

[Radio Wrangler]

I had a look in some of my boxes of crystals, but no luck. I used to have a homebrew transceiver using that multiplier-IF scheme, and any that translated into 2m FM chanels would have been used.

I don't think I want a brain the size of a planet because I understand they have the side effect of a terrible pain in all the diodes down your left side...

David

[Jon_G4MDC]

Poor old Marvin.

I'm hoping that my box of rocks might have one or two that escaped my spreadsheet. I'm not that organised but it will need the gods to smile for sure.

Ur right about Paul. He lived over the road from a good friend of mine. I recall chats in the street with him about all sorts of swag.

[sentinel040]

....no, you I suppose you don't want all those diodes to be killing you, do you, probably more than asprin could deal with.

I am touched you have both had a look, thank you. If, think positive - when I find a suitable replacement crystal I shall ask for the thread to be reopened (if it has died of natural causes) and let you know.

Take care.

Ian

[Bazz4CQJ]

If the old xtal is dead as the door-nail, is there anything to be lost by tinkering with it?

If you wanted to drill one (or two) holes in it, I could offer to bake it in a vacuum oven for you. I would have to let air back in as I don't have any dry gas, but maybe such a cycle might "do something".

If it still wasn't working, I could fill it with a "dense, powerful, cleaning solvent" and put it in an ultrasonic bath and then bake it out in the oven.

Think I'd be out of ideas at that stage.

B

[sentinel040]

Baz.

May I say sorry for the delay, but the person I care for has been in rather a bad way, which has resulted in me being somewhat preoccupied.

Thank you for your kind offer, I shall send a PM to explain the details and discuss this further.

Ian

[FrankB]

Some tricks that we used here in byegone days:
To lower the freq of a crystal, rub a lead pencil on it.
To increase the freq, put it on a piece of flat glass and use a sink cleaning compound like Comet Cleanser, etc and carefully and LIGHTLY rub it in a figure 8 pattern. Every few swipes, rinse, dry, and put back in the holder & see if it is where you want it to be.
Now I have used the pencil trick to reliably lower a crystal freq for one of my transmitters.
Now it could also be the crystal surface has become contaminated and just a careful cleaning might restore it to operation. I have also seen the spring type retainers in a holder lose their tension and cause the xtal not to oscillate.
It is a crap shoot, but if the crystal is dead, seems you have nothing to lose. I would try the cleaning method first.

[Radio Wrangler]

If the crystal oscillates in your GDO on its fundamental frequency, that's a good sign. Overtone operation needs there to be an LC tank to force an oscillator to run on an overtone mode. Overtone modes are inevitably less active than the fundamental.

Oscillation shows some normal activity and that the crystal isn't ruined.

I assume the crystal is housed in a metal holder, probably HC6 or HC25 sizes?

Is the can soldered onto the base (with a solder blob near the top sealing a little hole? Or is the can cold-welded onto the base?

Solder-sealed cans leave soldering residue inside which can move around, landing on the element and shifting the frequency. Unlike the older screwed-together crystal holders, the quartz plate inside these will have 'plated electrodes' formed on their surface, usually by vacuum deposition. Crystal grinding to raise frequency won't work because it will take the electrode off leaving a clean quartz surface.

So what could have gone wrong? Contamination by solder blobs or flux is a possibility. Corrosion of the electrodes by flux residue is another.

S20 is a frequency not assigned for commercial or military use, and I don't believe any amateurs bought new, crystalled-up Bantams from Pye, so your crystal is likely to be an aftermarket job from one of the quartz manufacturers serving the amateur radio market back in the day when the Bantams were on the surplus market. Cost was a major concern in this market, so a solder-sealed can is likely.

Cold-welded cans are better quality than solder-sealed ones. They have about 5 times better ageing rates based on averages over a lot of crystals. You paid extra for them. They are pretty much unopenable. Their benefit is in less contamination inside them. There were also glass HC6 cases where the glass can is frit-sealed onto a glass base, and these are cleaner still and have lower ageing rates. You'll only find these on special quality crystals.

So with a crystal which can at least oscillate in another circuit, but on its fundamental mode, the question is why doesn't that crystal work on its overtone mode in that circuit?

I think the crystal has started exhibiting greater losses or lower activity, however you choose to phrase it. Contamination moving around or corrosion on the plated electrodes is likely.

The oscillator circuit posted has a resistor across each crystal. It will damp the resonance of the crystal allied with its series indoctor. Trying a higher value might help persuade the circuit to start.

You can view the circuit as being a basic common-collector Colpitts circuit with the crystal inserted in series with the tank inductor. With the crystal socket shorted, it should oscillate somewhere around the 44MHz frequency with adjustment of the inductor shifting the frequency a lot. Think of the crystal as a narrow-band short circuit insisting on a more precise frequency and dominating frequency control. It won't look like a perfect short, it will look like a resistor, the ESR of that crystal of that mode. They can be engineered down to a handful of Ohms, but may be several tens of Ohms. This will set the activity of the oscillator.

You could short the crystal holder with various values of resistor and see what range the circuit will start up as an LC oscillator with.

One freedom you have is to increase that loading resistor value. Another freedom would be to play with the transistor bias current - both the gain and transition frequency, Ft, of a bipolar changes with bias current. With a battery powered or mobile set it's likely the designer was parsimonious with current and the transistor gain at the intended frequency is a bit on the low side. You may also have been affected by drift of resistor values reducing the quiescent collector current and this crystal is the first one to fail to work.

A third freedom is to change the oscillator transistor to something with a bit higher Ft at whatever bias current it is running.

Just some thoughts.

DAvid

[Jon_G4MDC]

I also think opening that crystal should be very much the last resort. On the few times I have tried it, just unsoldering the sealing solder blob sends the frequency low and nothing will bring it back.

RX for 145.5 is LO frequency of 134.8 - isn't that an airband frequency? There could be an AM TX crystal for that somewhere in the world.

I think Bantams used HC18 crystals for RX and HC6 for TX. It wouldn't be hard to adapt it for HC25.

[Radio Wrangler]

Since crystal-switched airband transceivers were in use, there have been at least three mandated replacements of aircraft COM radios (One for broadcast band attenuation, one for 25kHz channels and latest, the mandate for 8.33kHz channels) So all the aviation transceivers will have long gone on the scrap pile and will be floating around on rally stalls.

On the bad side, in the days of switched crystals it was common for designers to employ a crystal-mixer conversion in the transmit path so the same crystals worked for transmit as receive. You might be lucky and find something as a highside LO for a lower frequency aviation channel but these were 50kHz and 100kHz spacing back in the day.

David

[sentinel040]

Good afternoon.

I am sorry for the delay in responding, unfortunately my carers role has been a little demanding of late. However, over the weekend and with the limited resources I have away with me, I have had a further 'play'.

To try to tie up loose ends, the receive crystals in this Bantam are HC25U, although I recall some used wire ended ones. taking cognisance of what has been explained to me I experimented with the value of the damping resistor, the values on the resistors in the bias network and found another transistor with a bit more 'go', an AFZ12, from memory these gave a little more gain in Cambridge front ends. It would bore you to death if I explained all the ins and outs, suffice to say it was "interesting", although in the end unproductive. I mentioned earlier that as a project this radio had been a bit of what my father would have called a "right so-and -so", being unstable at one stage and then mysteriously resolving itself - a situation that always leaves me feeling ill at ease. The relevance is that back then I was doing all the work with the radio on 145.5.

Over the weekend it started to become unstable again, so I put the local oscillator back to standard, although it was only unstable when the crystal in question was switched in. By adjusting the crystals series inductor I could get it to start howling nicely, and by tweaking the collector tuned circuits I could get it to very weakly hear a signal drifting around 145.3-ish, infact very 'ish', before it leapt off somewhere else. This crystal was bought second hand and has at times behaved itself, although it has always been a bit lower in 'go' than the others channels. Its fundamental also seems to have changed as the tank circuits sort of peaks in a completely different state, although the output is very low level; unfortunately the little handheld DFM I have with me is not sensitive enough, but the dip oscillator shows it way off.

So, a wierd one alright. As David commented, crystals can do odd things. So, I think it is safe to take it off life support and minister its last rites and start hunting for something I can rob for the crystal. Why bother when I have a shelf containing SRP9000's, 8000's, PRP73's. PFX's, GP300's, early Jap stuff, etc.? Well because I like it and have a soft spot for it, and not the local swamp either, although... .

It will live again.

Once again, thank you all.

Ian

[Phil__G]

Klove (ex Quartslab) or Euroquartz will cut you a new one
Cathodeon and Websters (Chard) sadly are no more...

Maybe time to embark on a synth project?

[Jon_G4MDC]

Well done Ian for your perseverance!

If you synthesise it then it won't be a Batman any more will it? (Sp mistake intended!)

I'm thinking that I must have a 2m FM Batman somewhere. No idea what rocks might be in it. R6 for GB3PI most likely - maybe not much else. There could also be some W15 lurking somewhere with some 2m rocks in that escaped my Excel xtal listing exercise. That was just the contents of the junk box drawers.

I also once had an NR56 - SR9 monitor receiver. I think I bought it at some Leicester Rally in 197x. I got bored and so converted it for marine band. Pretty sure it is no longer with me. I think that will be 45MHz rocks too.

Would you like to try a TR2200 crystal "on borrows". PM me about a temporary swap if interested. It's only postage and a padded bag. I like to keep the TR2200 operational but it's not a daily user.

The FM version of the TR7010 SSB mobile, the TR7200G used the same crystals. I always wanted one. I still might if one came along HI!

Another possibility for custom made quartz might be here. They have some interesting RC stuff.

http://www.krystaly.cz/en/Products/Crystals