UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 12th Sep 2021, 11:40 am   #41
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ratio and Foster-Seeley detectors

Quote:
Originally Posted by Leon Crampin View Post
The capacitive centre tap shown above was used by Murphy amongst others. It seems to work well enough on the sets I have re-aligned.
The Thorn 1500 chassis, the collector circuit being connected directly to the secondary's capacitive tap.

Lawrence.
ms660 is offline  
Old 12th Sep 2021, 2:12 pm   #42
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ratio and Foster-Seeley detectors

A copy of RCA's original lab. report (1945) for the ratio detector here (page 79):

https://worldradiohistory.com/Archiv...ronics_120.pdf

Lawrence.
ms660 is offline  
Old 12th Sep 2021, 5:05 pm   #43
IanG3XYV
Pentode
 
Join Date: Sep 2014
Location: Bedford, Bedfordshire, UK.
Posts: 172
Default Re: Ratio and Foster-Seeley detectors

Thanks for that link Lawrence. Never read a copy of Radiotronics before. Lots of interesting articles to peruse at leisure.
IanG3XYV is offline  
Old 12th Sep 2021, 5:17 pm   #44
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ratio and Foster-Seeley detectors

No problem.

Lawrence.
ms660 is offline  
Old 12th Sep 2021, 11:59 pm   #45
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Ratio and Foster-Seeley detectors

Quote:
Originally Posted by Radio Wrangler View Post
Regarding phase shifting, The phase shift that matters, the phase shift that the ratio detector uses is that between the centre-tapped secondary and the tertiary windings.
Agree! (or between primary and secondary, if the configuration is as per Lawrence post #39)

Quote:
Originally Posted by Radio Wrangler View Post
Phase relationships with the primary are of no importance
Agree (unless as per post #39), as you say, the primary itself isn't usually connected to the secondary.

Quote:
Originally Posted by Radio Wrangler View Post
The detector function operates only from the centre-tapped secondary and the tertiary. There is no other path from the primary for any phase comparison to be made. Just so long as the centre tapped secondary and the tertiary have the right phase/frequency relationship, the primary can do what it likes as far as FM demodulation is concerned.
Agree! Though as the phase in the tertiary happens to be the same as the primary, the primary will be 90° (+/-deviation-caused shift) out of phase with secondary.

Quote:
Originally Posted by Radio Wrangler View Post
It would be a shame to have an expensive IF transformer can in the discriminator and only get one pole's worth of selectivity out of it. The accountants would get edgy.

Re-jigging the transformer so that the primary is loose coupled to both the tertiary and the centre tapped secondary gives the freedom to have more selectivity without such bad behaviour at modulation peaks
Agree – and there’s another benefit to tuning the primary, too. If you don’t tune it, the voltage that the IF amplifier can deliver to the primary is rather small (the IF amplifier is usually a pentode, thus a current source). But if you tune the primary, the load it presents to the valve is much higher, and the current in the primary is approximately Q times higher than the IF component of the anode current. You get better sensitivity.

Quote:
Originally Posted by Radio Wrangler View Post
but the centre-tapped winding and the tertiary still have to be loose-coupled between themselves, however the drive from the primary is arranged.
Agree!

It’s just a matter of how the voltage gets produced in the tertiary, you want it to be governed by the voltage in the primary and be relatively independent of what’s going on in the secondary. You don’t want the tertiary to be loosely coupled to BOTH, for instance else the phase of the voltage in the tertiary would be mid-way between the phase in the primary and the phase in the secondary. And then the transformer would be a right pain to analyse, with three coupling coefficients (pri-sec; pri-tert; sec-tert) rather than just one (pri-sec, with pri-tert being 1).

Quote:
Originally Posted by Leon Crampin View Post
The essential phase shift for these detectors is a result of the leakage inductance between primary and secondary. Loose coupling and a fairly low value of load resistance are needed to achieve this - easily demonstrated with a vector diagram.
Agree! The leakage inductance resonates with the secondary capacitance to form an LC filter.

The phase relationship of this LC filter is 0 at low frequencies, well below the LC resonance; 180° well above resonance; and transitions through 90° at resonance, with the steepness of the change, with frequency, being dependent on secondary Q. But high Q results in worse linearity, so a compromise is sought. There’s also the fact that with a high Q, the amplitude across the C will vary with frequency, whereas we want to to be relatively constant for the (small) frequency deviation. Both of these dictate a fairly low Q.

With the secondary voltage 90° shifted from the primary, because a tiny bit of the primary flux is what produces the voltage input to this LC filter, we just need to get the primary phase available for the diodes. The tertiary is the common method (close-coupled to the primary, so the tertiary is in-phase with the voltage which excites the whole thing). But it is possible, as Lawrence’s post #39, which look as it they’re not quite balanced to me, but would work!
kalee20 is online now  
Old 13th Sep 2021, 10:16 am   #46
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ratio and Foster-Seeley detectors

Re: Post#39 detector circuits, amongst others, the 1st one was used in the Alba T790 TV and the 2nd one was used in the Ferguson 3608 TV (Thorn/BRC 800 TV chassis)

Lawrence.
ms660 is offline  
Old 13th Sep 2021, 11:35 am   #47
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ratio and Foster-Seeley detectors

The enclosed scans are from my own copy of a television servicing book (Patchett) giving an explanation of the ratio detector, I had this book in the mid 60's to supplement the day release City & Guilds course that I had started earlier, it was recommended by the course teachers along with the radio servicing books, I was lucky because I was awarded several of those books by the college for my exam results so I didn't have to buy all of them.

Fuzz, distortions etc are due to the scan.

It goes on to explain the locked oscillator FM discriminator (anyone remember the EH90?) if anyone wants a scan posted up let me know.

Lawrence.
Attached Thumbnails
Click image for larger version

Name:	ratio det 1.jpg
Views:	58
Size:	58.9 KB
ID:	241182   Click image for larger version

Name:	ratio det 2.jpg
Views:	48
Size:	71.8 KB
ID:	241183   Click image for larger version

Name:	ratio det 3.jpg
Views:	45
Size:	77.2 KB
ID:	241184  

Last edited by ms660; 13th Sep 2021 at 11:46 am. Reason: extra info
ms660 is offline  
Old 19th Nov 2021, 5:02 pm   #48
AidanCroft
Tetrode
 
Join Date: Feb 2021
Location: Crewe, Cheshire, UK.
Posts: 66
Default Re: Ratio and Foster-Seeley detectors

Hiya,

Does anyone here understand and perhaps you might explain the functional differences between an FM ratio detector and a Foster Seeley detector?

I understand they are similar but from the attached photos, a Foster Seeley has a connection from primary to centre of secondary (via a capacitor) whilst a ratio has no connection, only using a tertiary winding.

Also a Foster has diodes pointing in the same direction whilst a ratio uses diodes pointing in opposite directions.

I also understand the concept of phasors adding up to produce an output. But now, this is where my understanding FALLS DOWN....

On a ratio detector, how is it that the loosely coupled secondary produces a frequency sensitive output to one diode over the other diode? Why is one preferred to another? In any case they point in opposite directions so how would that work?

On the FS detector, I can see that both diodes point in the same direction so I'm guessing one diode is preferred to the other as each half of the secondary is tuned slightly off centre from the IF frequency (and so depending on the frequency shift direction, one half of the secondary will have the greater output)?

In both types of detector, why do we need a 90° shift in phase? Can't we simply have both halves of the secondary tuned either side of the IF frequency?

Kind regards,

Aidan.
Attached Thumbnails
Click image for larger version

Name:	0S5Kk.png
Views:	35
Size:	60.7 KB
ID:	246146   Click image for larger version

Name:	Screenshot_20210906-214718_Chrome.jpg
Views:	32
Size:	27.5 KB
ID:	246147  
AidanCroft is offline  
Old 19th Nov 2021, 6:13 pm   #49
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Ratio and Foster-Seeley detectors

Before attempting to understand the Foster-Seeley and Ratio detectors, because they are rather complex, read up on the other frequency discriminators the simpler ones as they make various aspects clearer as you see the relationships.

The simple AM diode detector uses the carrier to beat up the diode, forcing it on and off. It then behaves as a synchronous rectifier, in phase with the carrier component, synchronously rectifying the carrier components, resulting in audio output.

Now move on to the 'Quadrature Detector' in this, the FM signal in the IF is hard limited and then mixed with a quadrature-shifted version of itself. The AM detector above was in phase, this FM one needs a 90 degree phase shift. This shines through in the ratio detector. The quadrature detector is simple in philosophy, but didn't get going until IC technology made 4-quadrant analogue multipliers available so it could be a synchronous rectifier without any awkward baggage.

Now read up on slope demodulation of FM and look at its poor linearity due to the shapes of filters, and it's limited deviation capability AND how it works part way down the slope of the filter, worsening S/N ratio. Follow this through into the Travis detector, which is a pair of slope detectors flying in push-pull configuration. Slope and Travis are simple things using the frequency response shape of a resonator to convert frequency wobbles to amplitude wobbles and then it hits them with an AM detector.

Now you're ready for the Foster-Seeley where phase starts to play a part. This then goes further into the ratio detector where a large capacitor is used to hold a voltage representing signal strength, and used to normalise the amplitude of the IF signal by tracking it with the diode switching thresholds.

The ratio detector is a bit of a bodge, accepting limitations to reduce the need for IF limiter stages. If you can afford the extra limiter parts, the Foster-Seeley is more hifi. Further in that direction is the quadrature detector if done with a 2-stage phase shift network.

A well-designed Travis, with a good signal to noise ratio input can demodulate large bandwidths with exceptional linearity I've measured Travis demods that can handle 20MHz deviation on a 7MHz carrier with baseband up to several MHz and giving enough linearity for approaching 70dB NPR linearity.

You've got a lot of reading ahead of you, but you need to be comfortable with it before things start to 'click'

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 19th Nov 2021, 6:22 pm   #50
Guest
Guest
 
Posts: n/a
Default Re: Ratio and Foster-Seeley detectors

Quote:
20MHz deviation on a 7MHz carrier
A typo I assume, otherwise how do you get a negative frequency? 180 degrees perhaps?
 
Old 19th Nov 2021, 10:05 pm   #51
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Ratio and Foster-Seeley detectors

Oops 70MHz, the bog standard IF for microwave radio links, the beasties carrying a few thousand phone calls over 40km hops in the days before PDH and SDH.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 23rd Nov 2021, 11:41 am   #52
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Ratio and Foster-Seeley detectors

Quote:
Originally Posted by AidanCroft View Post
Does anyone here understand and perhaps you might explain the functional differences between an FM ratio detector and a Foster Seeley detector?
There are differences, and similarities.

In both, there is a transformer which produces a pair of voltages across the centre-tapped secondary, that's 90° out of phase with the primary.

The magnitude of the voltage doesn't change (much) with frequency, but the phase relationship does.

This 90° phase-shifted voltage gets added to a portion of the primary voltage, and applied to a pair of diodes.

The total voltage applied to the diodes does then vary with frequency, because when the added voltages come a bit into-phase, the sum increases; when they go more out-of-phase, the sum decreases.

So, each diode produces a rectified voltage, usually several volts, but which varies a bit with frequency. At the centre frequency, the two voltages are equal. If frequency drops, one rectified voltage falls, one rises; and if frequency rises, it's the other way round.

This applies for both the Foster-Seeley and the Ratio Detector.

Now we come to the differences. In the Foster-Seeley, the AF output voltage is the difference of the two rectified voltages. If both voltages are 5V at the centre frequency, but one varies from 4V to 6V while the other varies from 6V to 4V over the whole frequency deviation, then the difference is 0V at the centre frequency and varies from -2V to +2V. By taking the difference, some of the non-linearities in the whole process cancel out, a bit like a push-pull amplifier.

In the ratio detector, the two voltages are added by turning one of the diodes around. So at the centre frequency, you get 5V + 5V = 10V. At one extreme of deviation, it's 4V + 6V = 10V, and at the other extreme it's 6V + 4V = 10V.

There's no AF output from the sum!

So the AF is taken from the mid-point, nominally 5V, but varying +/-1V either side (so, all things being equal, the RD gives half the output of the FS).

The clever bit about the RD is, the non-linearities mean that the sum doesn't quite stay at 10V. But if you force the sum to stay constant, by slugging the sum with a big electrolytic capacitor, then throughout the AF cycle the load on the system varies marginally, and the non-linearities are ironed-out.

It also makes the circuit relatively immune to any amplitude modulation: if the amplitude of everything tries to increase, the sum voltage can't follow it quickly, everything stays relatively constant.


Quote:
Originally Posted by AidanCroft View Post
a Foster Seeley has a connection from primary to centre of secondary (via a capacitor) whilst a ratio has no connection, only using a tertiary winding.
The business of the tertiary winding (in the RD) and a direct capacitor coupling (in the FS) is just a matter of convenience. In the FS, it's possible to apply the primary voltage, ground-referenced, so a capacitor coupling works. In the RD, it isn't - so a floating coupling winding to pick-off the primary voltage, is needed.

Hope this helps!
kalee20 is online now  
Old 12th Jan 2022, 10:57 pm   #53
AidanCroft
Tetrode
 
Join Date: Feb 2021
Location: Crewe, Cheshire, UK.
Posts: 66
Default Re: Ratio and Foster-Seeley detectors

It really helps a lot so thank you for your detailed notes.

Quote:
Originally Posted by kalee20 View Post
There are differences, and similarities.

In both, there is a transformer which produces a pair of voltages across the centre-tapped secondary, that's 90° out of phase with the primary.

The magnitude of the voltage doesn't change (much) with frequency, but the phase relationship does.

This 90° phase-shifted voltage gets added to a portion of the primary voltage, and applied to a pair of diodes.

The total voltage applied to the diodes does then vary with frequency, because when the added voltages come a bit into-phase, the sum increases; when they go more out-of-phase, the sum decreases.
I understand how the secondary output would change away from resonance because a tuned "tank" changes its response with frequency. But, how can each HALF of the secondary give a different output? Is each half of the secondary tuned either side of the IF and is that the case, if so, for both FS and RD? If so again, how is that achieved in FS and RD with only one capacitor sitting across both halves of the secondary?
AidanCroft is offline  
Old 13th Jan 2022, 12:48 am   #54
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Ratio and Foster-Seeley detectors

No, the secondary is tuned to the centre frequency, by the one capacitor. And the two halves give identical voltages.

It's because they are opposite-phase to each other, that when the phase between the CT secondary and the tertiary changes, one half becomes more in-phase and one becomes more out-of-phase.

And the voltage applied to the diodes is the sum of the half-secondary and the tertiary, so one diode thus sees more voltage and the other sees less.
kalee20 is online now  
Old 13th Jan 2022, 1:23 am   #55
Boulevardier
Octode
 
Join Date: Sep 2017
Location: Bristol, UK.
Posts: 1,642
Default Re: Ratio and Foster-Seeley detectors

Is the direction of the diodes the defining characteristic of the ratio detector? Here's an example of one (Grundig) that seems to have no tertiary winding. Is it an RD?
Mike


Click image for larger version

Name:	ratio detector.png
Views:	60
Size:	72.0 KB
ID:	249607
Boulevardier is online now  
Old 13th Jan 2022, 6:01 am   #56
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Ratio and Foster-Seeley detectors

It is a ratio detector. The directions of the diodes or the big floating electrolytic capacitor are the give-aways.

One transformer with a tertiary winding is common, but the same effect can be done using two transformers, each with only two windings (but one with the necessary centre tap). As a third alternative, a single transformer with a centre-tapped secondary can be used with a sniff of voltage brought in from the driver stage.

There are several variations you can play, so the ratio detector is a class of circuit and a modus operandi rather than one unique circuit.

Set makers played variations based on predicted cost savings, maybe two separate but more standard coils were cheaper than a specialised one? What parts did they get best discounts on?

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 13th Jan 2022, 11:02 am   #57
Boulevardier
Octode
 
Join Date: Sep 2017
Location: Bristol, UK.
Posts: 1,642
Default Re: Ratio and Foster-Seeley detectors

Thanks for that David. I feel less conflicted about RDs now
It was confusing that all the text-book accounts of RDs that I’ve come across stress the separate third winding and the nature of it’s coupling to the other windings.
Mike
Boulevardier is online now  
Old 13th Jan 2022, 2:01 pm   #58
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Ratio and Foster-Seeley detectors

You can see the little winding between pins 2 and 5 of the left-hand coil. That's effectively the tertiary winding. It's tightly coupled to the primary, so the phase is the same.

Then the primary (in the LH coil) doesn't go straight to ground, but does so through the little winding , pins 4-5 of the RH coil. So the RH coil with its centre-tapped secondary only gets a sniff of the primary - the voltage here can change its phase with what's going on in the primary as a whole (which is the same phase as the tertiary).
kalee20 is online now  
Old 13th Jan 2022, 9:30 pm   #59
AidanCroft
Tetrode
 
Join Date: Feb 2021
Location: Crewe, Cheshire, UK.
Posts: 66
Default Re: Ratio and Foster-Seeley detectors

I'm so close!

Quote:
Originally Posted by kalee20 View Post
It's because they are opposite-phase to each other, that when the phase between the CT secondary and the tertiary changes, one half becomes more in-phase and one becomes more out-of-phase.
My niggle now is, how can you have two different phases sitting on the same secondary? How can one half have a different phase to the other half in other words?

I can see how such a different adds to the so called primary "reference point", but not how how different phases arise on the secondary in the first place.

Aidan.
AidanCroft is offline  
Old 13th Jan 2022, 10:10 pm   #60
Boulevardier
Octode
 
Join Date: Sep 2017
Location: Bristol, UK.
Posts: 1,642
Default Re: Ratio and Foster-Seeley detectors

Thanks for that Kalee.

But -
Quote:
Originally Posted by kalee20 View Post
You can see the little winding between pins 2 and 5 of the left-hand coil.
I can't see a little winding between pins 2 and 5 of the left-hand transformer? Am I just being blind?

Mike
Boulevardier is online now  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 7:17 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.