The HRO Crystal Filter

[Bazz4CQJ]

The HRO (5) filter is, like most of the HRO, pretty much straight from 'the text book', with just the one crystal. As I understand it, the filter works by 'losing' the unwanted signal in the 'notch' which is created as a result of the series and parallel resonances of the crystal. That being so, I've never understood why is that called the "Phasing" control?.

I can see how that would work with CW reception, but I'm not interested at all in CW. For either AM or SSB, I've never found that filter to be very useful, but perhaps there might have been alignment issues in my HRO, which I'm currently rebuilding.

Pushing originality issues to one side, I wonder what else one might install in that huge space occupied by the standard filter circuitry that would be more useful for AM/SSB reception? Are other approaches like a Q-multiplier or a AF bandpass filter in contention?

[Radio Wrangler]

The quartz crystal on its own gives a low-Z dip at its series resonant frequency, and a high-Z peak at its parallel resonant frequency.

Connected in series with the signal, there will be a peak in reception at the series frequency, with a notch several kHz higher at the parallel resonance.

There is a problem. The crystal has capacitance between its electrodes, and this creates a bypass path for signals to leak around the filter.

The solution for early crystal filters was to add a winding to an IF transformer to create an opposite phase version of the signal voltage, and to use this to drive a matched capacitor to the crystal output.

This nulls the signal passed by the electrode capacitance. The skirts around the passband now fall, and the notch tends to vanish. So you off tune a bit from something and adjust this balancing capacitor "Phasing" on the front panel to minimise the offset signal, and now the crystal filter is at its most selective. The HRO has this control on the panel. The AR88 has it under the lid on the deck.

If things are a bit congested twiddling "Phasing" can give you a range of odd-shaped filter shapes and a notch which can help.

It's useful to have a mark to show the normal positionof the control.

David

[G6Tanuki]

The "single crystal" filter as used in the likes of HROs, AR88s etc. was really optimised for CW reception in most instances - you got a peak and a notch - so with skill you could either put the wanted signal into the peak, or position the notch to take out a particularly interfering signal close to the one you wanted.

The "phasing" control, as has been described, essentially neutralises the crystal's self-capacitance (which would otherwise provide a path for signals to leak past the filter).

The problem for AM/SSB reception is that in standard form the peak is extremely sharp - a few hundred Hertz - which is less-than-optimal for receiving such modes. To get round this it's possible to alter the impedance the filter is operating into - which will flatten the sharpness of the peak and make SSB reception possible - though it invariably sounds a bit wasp-in-a-matchbox strange because the flattened-filter still doesn't have a uniform frequency-response.

With the coming of SSB the professionals developed the 'ladder' crystal filter and the ceramic or mechanical filter [Collins, Kokusai] which could give a much 'squarer' response.

It would be possible to modify a HRO xtal filter to a 2-element version by getting a pair of xtals with suitably-spaced series-resonant frequencies, then putting one in the current xtal's position and the other in parallel with the 'phasing' control [which would then become a 'set-once-during-calibration-and-neber-touch-again' phase-balance control.

Alternatively - if you want something different - it's possible now to buy extremely small ceramic filters that look not much bigger than a couple of transistor-radio IF transformers. A couple of these in a suitably-screened mini diecast box could be hooked into your HRO beneath the chassis and would provide a spoectacular improvement in AM/SSB selectivity without needing any irreversible circuit-modification.

[Bazz4CQJ]

Originally Posted by G6Tanuki

Alternatively - if you want something different - it's possible now to buy extremely small ceramic filters that look not much bigger than a couple of transistor-radio IF transformers.

I'm really glad you suggested that because I do have a Murata filter which I think is 15 element; +-4kHz at 6dB and +-10kHz at 70dB. It was bought for another project (which may or may not happen) but could be diverted to the HRO. The input and output impedance are quoted at 2000R, which I assume is more directly suited to modern transformers in transistor circuits. What would be the best way to match it to the HRO?

B

[G6Tanuki]

the 2K inpedance on each end of the filter won't play well with a HRO's circuit-impedances.

The best way to handle this would be to get a couple of single-tuned miniature IF transformers [the Toko 10mm series for example].

Then wire the non-tuned secondary of the first transformer to the input of your Murata filter, and the output of the Murata filter goes to the untuned 'secondary' of the second trnasformer [essentially using this one back-to-front].

Then you use the outer ends of the transformers [ignoring any taps] to hook to the HRO's mixer-anode and first-IF-stage grid in place of the original HRO IFTs.

If you are doing this, make sure the input and output sides of the new filter are very well screened from each other - if the input and output can 'see' each other even slightly at RF without the signal needing to go through the filter, it can seriously compromise your filter's ultimate 70dB rejection. In the past while doing this sort of thing with HF double-balanced mixers followed by ceramic/mechanical filters I've made up screening-boxes by soldering together offcuts of double-sided PCB and then when everything's peaked-up properly, soldering a lid on the top to prevent 'hop-over' of stray signals.

[Bazz4CQJ]

OK - thanks for those suggestions. Could I just go back to the possible use of a Q Multiplier? Even if the addition of a filter achieves a narrow passband, then it's a symmetrical one. It's sometimes the case that you have 2 or 3 stations close together, you'd really would like to tweak the shape of the passband. I've never owned a receiver with a Q-M, often read comments about how effective they can be, so continue to wonder about that option?

B

[Herald1360]

A Q multiplier won't give a tweakable passband shape, it'll just sharpen up the selectivity/reduce the passband of the stage that it's on. It's basically adding some regeneration (positive feedback) to the stage.

[G6Tanuki]

Depending on how a Q-multiplier is used, it can either be used as a peaker or a notcher. Most often they're found as a peaking device [like regeneration] but since the effect of a Q-mult is to simulate a high-Q tuned-circuit this high-Q circuit can be used in series-mode to provide a very low-impedance signal-path to earth at a specific frequency.

The biggest problem with Q-multipliers or anything else using traditional tuned-circuits is getting a good "shape-factor".

[Radio Wrangler]

After the single crystal/phasing control type, as said above, the next step was to using a second crystal balancing the first. The electrode capacitances cancelled, and it was usual to plant the series resonance frequency of one crystal on the parallel resonance frequency of the other, so crystals with controlled series/parallel mode spacing were needed. Subsequent development concatenated more of these stages to make a recatangular shaped passband with a Chebyshev ripple passband. Because of their circuit, these were called Lattice filters.

Ladder filters came later, I think out of work at Bell labs looking into cost reductions of old style frequency division multiplex phone trunks (think of 1800 ssb signals shoved down a coax cable or over a microwave link) One key name in this area is Dishal. This structure is a long string of similar crystals with shunt capacitors at each node...hence the name.


An interesting example of DIY lattice style filtering in a homebrew receiver is by G2DAF in the first version of his RX. He went for Kokusaui mechanical filters later. The first version gives a lot of info on selecting, testing and matching classic FT243 crystals to build anIF strip.

David

[G6Tanuki]

One other thing to think about: the standard single-crystal-and-phasing-control used in the likes of the HRO gives a very peaked audio response which isn't good for receiving SSB. *but* if you introduce some audio-processing which is the mirror-image of the filter's response [if your SSB signal is tuned into the centre of the xtal filter's characteristic then your audio filter should add some treble and bass boost] it can compensate for the xtal filter's audio-response.

This is not a new idea: Dr James Robinson, the originator of the crystal filter - proposed just this approach for broadcast-band reception using 2-crystal filters to give extremely narrow IF bandwidths followed by post-detexction audio-response correction - and published several papers on what he called the "Stenode Radiostat".

[Bazz4CQJ]

OK, thanks for those comments. It looks like my original assessment that the original HRO filter is limited to a great degree to use with CW. Adding the modern ceramic resonator filter sounds relatively easy and should suit AM/SSB; do that and re-assess the performance at that stage.

I do love the idea of an HRO which has had a few refinements added to it. I'm not so much interested in restoring appearance, as restoring "capability". In the 1930's, the HRO was a very capable receiver; 80 years later, it really needs what I like to think of as a "Mid-life-update", but all done in a considered manner.

[turretslug]

Originally Posted by G6Tanuki

The best way to handle this would be to get a couple of single-tuned miniature IF transformers [the Toko 10mm series for example].
Then wire the non-tuned secondary of the first transformer to the input of your Murata filter, and the output of the Murata filter goes to the untuned 'secondary' of the second trnasformer [essentially using this one back-to-front].
Then you use the outer ends of the transformers [ignoring any taps] to hook to the HRO's mixer-anode and first-IF-stage grid in place of the original HRO IFTs.

That's a route that I've chewed over in the past- but are modern miniature 10mm IFTs really up to olde-worlde HT? I'd thought of leaving original IFT in and re-wiring its grid circuit so that it was top-coupled by a small trimmer to the first filter-matching IFT, then re-wiring any AGC control to the second filter-matching IFT in the "usual" fashion. With the unknowns re. old and new coupled tuned circuits, it would need a session on the wobbulator to tweak the top-coupling for a good overall shape but there should be mileage in it. Haven't actually got round to it though, I'd be interested to see if it would work.

[turretslug]

Oh, and Bazz4CQJ, very much go along with those sentiments- it would be difficult and uneconomic to attempt to replicate the classic gang-tuned front ends and their reduction drives, so why not build on the solid core, as it were.

[Bazz4CQJ]

I'm firmly of the opinion that the key part of what made the HRO great, and still makes it good, is the genius of National's Precision Worm Drive (PWD) and its dial. The HRO was the first decent receiver I ever owned and my (quite prejudiced) view is that no other receiver has such a wonderful tuning system. After you've got used to an HRO tuning system,..."apres moi, la deluge" applies.

In among my treasure trove, I have a National gear box which is a variation of the HRO one; it takes the standard dial, but instead of having two output shafts at right angles, it has a single output shaft (geared down at the same 50:1 ratio) in line with the input shaft. Curiously, the exact same rare thing is for sale online at the moment, but with a very tatty dial. Anyway, what about putting a high-quality four-gang pot (a bit like the one I've just bought, strangely) behind that gearbox and connecting that to 4 varicap diodes . Is that the basis for the HRO-21C (21st Century)? It will, of course, have a DFM, so the graph paper can be retired..

.

[G6Tanuki]

Originally Posted by turretslug

That's a route that I've chewed over in the past- but are modern miniature 10mm IFTs really up to olde-worlde HT?.

I've used them in such applications in the past: the ones I have in mind have the primary and secondary wound in separate sections on a small plastic bobbin so they're reasonably well isolated from each-other. The tuning-capacitor is in a slot in the bottom of the plastic former so it's also protected. The threaded ferrite 'tuning cap' covers the coils and so provided it's non-conductive there shouldn't be any risk of a coil-to-screening-can short.

Only possible difficulty I can think of is if the coil was fitted to a double-sided PCB and the tuning-cap was insufficiently recessed into the base of the coil-former it might possibly contact the upper PCB copper?

[AC/HL]

Three posts discussing an unrelated receiver deleted, please keep on topic.

[Skywave]

Originally Posted by Bazz4CQJ

. . . my (quite prejudiced) view is that no other receiver has such a wonderful tuning system. [The HRO receiver]

For its vintage, yes, agreed. But outside of that time restriction and in my equally quite prejudiced view, of the mechanical tuning arrangements ever devised, nothing beats that of the Racal RA-17 (and its derivatives).

Al.

[Herald1360]

Hi Al,

is that perseverance or persistence? Or did you miss something?

You seem to have restarted the exact digression (started by myself) that got canned already

Getting back to HROs, the mechanics of the tuning were pretty good, but the logging dial and graphs on the coilpacks were quirky to say the least, not to mention reminiscent of much earlier '20s designs.

None of which has much to do with the HRO crystal filter where it all started.

[Colinaps]

Well, I'm about to open up the 1st IF on an MX, which almost works ok after having most of the Rs and Cs under the deck replaced but whose crystal isn't doing what a crystal should, so I may be tempted by some of your suggestions.

Colin.