Re: Improving Selectivity In Vintage Receivers
 

The AR88 crystal is only active on three of the five selectivity positions. There's a switch contact shorting the crystal itself on the two widest bandwidths.

The 'phasing' trimmer C is used with the push-pull transformer winding to cancel the case capacitance of the crystal turning it into a pure single pole resonator. Switching loading gives different Q values and hence different bandwidths. So, the three narrow bandwidths are single pole shapes, not bad for ringing on CW but with very poor shape factor for adjacent signal rejection. There is no stagger tuning, only one pole so no tracking problem.

The widest bandwidth does need stagger adjustment to meet the selectivity curve in the handbook.

The HRO is a fairly similar setup, predating the AR88. The chief difference is the phasing or capacitance balance trimmer is a panel control on the HRO (and also on the AR88 follow-up set)

For SSB, the ceramic filters offer a useful improvement with their controlled, more rectangular shape.

David

Re: Improving Selectivity In Vintage Receivers
 

Found an early type HRO Crystal Filter Box with the plug in crystal. The schematic is much the same as posted previously.

The Mixer Anode Transformer and the Phase Splitting Capacitors are inside one circular can with lead outs to the external Tuning Capacitor, Xtal and Phasing control.

There is a ceramic preset between these and the 1st IF Grid Tuned Circuit which is in a second circular screening can. That contains a trimmer (accessed through the top of the crystal filter box) and a tapped inductor. It has 3 wires. Input from the "IF Gain" trimmer which goes to the tap, RF Ground (actually the AVC line decoupled) and the Grid lead output to the 1st IF Valve.

It is the tap on this tuned circuit that interests me for adding a ceramic filter.

Measuring the trimmer at what I assume was last known good tuning it is 45pF. Add perhaps 10pF of strays and it might be 55pF total. For resonance at 455kHz it suggests an inductor around 2.2mH.

Now what about the tap? Last picture shows it is about 1/3 up from the "Cold end" which seems sensible. DC Resistance to tap is 2.5 Ohms and the rest of the winding is 7 Ohms. Since the coil diameter is increasing with every turn length of wire and resistance per turn is increasing too. Is there some rule of thumb with this to estimate turns ratio from the resistance measurements? Maybe the eyeball "it's about 1/3rd up" is good enough?

Ultimately I hope to estimate the impedance at the tap to decide if it is a suitable place to connect a ceramic filter or not.

Re: Improving Selectivity In Vintage Receivers
 

Now trying to make some calculations on the coil.
The wire measures 0.25mm or ~10thou. It is cotton covered perhaps double cotton covered. The pie is wave wound on a former of 14mm dia.

So far I have failed to find any specs for this in published wire tables. Looking for expected Ohms/m if anyone might have it please.

Re: Improving Selectivity In Vintage Receivers
 

Here is a first estimate. If it's anywhere near reality my guess of the tap at 1/3rd the way up was massively out.

The physical dimensions here seem to agree with what I measure on the actual coil. The wire diameter of 0.25mm is double cotton covered which probably means the actual copper is much thinner. The Ohm/metre of 41 AWG seems to fit quite well.

What do we think? 30 or 31 turns total with the tap at 5 or 6 turns from the bottom. Say 5 or 6 times as the turns ratio?

Re: Improving Selectivity In Vintage Receivers
 

Re. post #9, the article in Wireless World by J. Fochet, F6BQP, was titled "Crystal ladder filters". It appeared in the July 1977 issue, pp. 62-63, and was translated from the French original by Pat Hawker, G3VA.

This article was included as one of 23 or so references in Joe Carr's article "IF filter options", Electronics World, May 2000, pp. 410-414.

73 John

Re: Improving Selectivity In Vintage Receivers
 

I'm sorry Jon, but I'm not sure quite sure what you are seeking in terms of that tap :dunce: ? What is the parameter you are trying to zoom in on to decide on connecting the ceramic? If it helps, I too should have that tapped coil sitting in the box of components which I took out of my HRO; are there any measurements I can take to complement your own?

@John KCOG; thanks for that info; I'll try track that article down.

B

Re: Improving Selectivity In Vintage Receivers
 

Found that coil. I'd completely forgotten that I had already assembled it in to a "first iteration" re-build of the xtal filter, but I can extract it to do measurements within a few minutes. You've not shown any numbers for inductance; would those values be helpful?

B

Re: Improving Selectivity In Vintage Receivers
 

If you can measure inductance Bazz it would be appreciated. I guessed at 2.2mH above but it was only that.

My idea with the tap is it will be a lower impedance point and might be made to look like 2k to load the ceramic output more or less correctly.

Re: Improving Selectivity In Vintage Receivers
 

Fine; will get L values to you later this evening.

B

Re: Improving Selectivity In Vintage Receivers
 

OK, have taken measurements of both R and and L on the windings of my coil.
The values of R add up within reasonable error, but I don't quite understand the L values; adding up the two segments does not give the inductance of the whole. See attached. It looks as though my allocation of green and red is wrong, wrt to real "hot and cold".

Perhaps I should add that L was measured at 10kHz, if I change to 100Hz, all the reading increase by 2-3%.
Is there some subtle reason I'm missing?

B

Re: Improving Selectivity In Vintage Receivers
 

Just for info, a pic.

Re: Improving Selectivity In Vintage Receivers
 

Thanks! I'm looking at that. I don't think it disagrees with my ideas by very much.
Cogitating some more...

The resistance values are quite close. The inductance values need some more thought to be given to them.

Reckon measuring with 100kHz is nearest to correct.

Re: Improving Selectivity In Vintage Receivers
 

Yes, I really don't know why the L values of the two segments don't add up to the whole value. The leads are open-ended, so there are no loading effects.
I cannot think what to do next to try resolve it.

B

Re: Improving Selectivity In Vintage Receivers
 

I just pulled a mains transformer out of the junk box and did measurements on the tapped primary;

0-245V L=2.94H (100kHz)
0-225V L=2.46H

225-245 L =5mH

At 100Hz, the readings are slightly higher but the issue is still the same.

Resistance values

0-245V...........R=90R
0-225................81R
225-245............ 8.8R.

B

Re: Improving Selectivity In Vintage Receivers
 

The inductances won't add up in simple form, you have to consider the mutual inductance too.

For example with a centre-tapped transformer, each half will measure at a quarter of the inductance of the whole winding - not a half.

Re: Improving Selectivity In Vintage Receivers
 

Nicely illustrated by a two-section C-core choke that I keep as a demonstrator- each individual winding measures 22mH, in series phase, 88mH.

Re: Improving Selectivity In Vintage Receivers
 

Think this comes under the "learn something new everyday" category. Thanks for your inputs on coupled inductors!

Searching Google, this site https://www.translatorscafe.com/unit...ctance-series/looks as though it could be helpful, though Jon may have a better insight in to the next step than I do.

B

Re: Improving Selectivity In Vintage Receivers
 

If you have a play with LTspice, you will find that it doesn't have a transformer as a component in its library. You just draw a series of inductors to represent each winding with inductances in ratios to represent the squares of the wanted turns ratios THEN you write a coupling factor statement and add it to your diagram. This works, but looks tedious, but it really makes you keep an eye on the amount of coupling you get in a real transformer. Less than unity coupling factor leaves you with leakage inductance and you'll see transients not being clipped etc.

The other interesting one is that their real world capacitor component has L, C and R. If you look at the model, it has the same topology as the standard model of one resonant mode of a quartz crystal! So you can make a crystal by putting a real world capacitor on your diagram and assigning appropriate values. Those values are rather extreme in order to model a crystal's huge Q. The disadvantage is that it makes your crystals unobvious to the point of becoming invisible to all but the most suspicious reader. Do this to people and they'll be looking for ways to get their own back.

David

Re: Improving Selectivity In Vintage Receivers
 

David, I think we need to know what the coupling coefficient is between the two sections of this coil. Based on my knowledge of this subject (now, some 5 hours old) I made guess at 0.8 and that came up with numbers that were 'plausible'. Would you like to offer guidance :).

B

Re: Improving Selectivity In Vintage Receivers
 

That's as good an estimate as you're going to get. To measure coupling factor you can measure the inductance of one coil while shorting the other and thus measure the leakage inductance. In theory a bit of maths will get you to what you want. In practice you'll run into stray resonances which spoil your efforts. Picking a number then twiddling it until what you see fits models under circumstances more like what it will be run under is the best you really can do.

You'll see some people with an almost legal bent wanting proof beyond any doubt for some things, but real situations can get sufficiently complex quite quickly that they become non-analytic to all intents and purposes. The three body problem hasn't been solved that I've heard of (Google it and you just get people trying to flog you books by that title... so much for google nowadays!)

There is an art in knowing when to stop trying to analyse and to go empirical. Sometimes it's the only way to make any progress.

David