UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Components and Circuits (https://www.vintage-radio.net/forum/forumdisplay.php?f=40)
-   -   How do IF transformers work? (https://www.vintage-radio.net/forum/showthread.php?t=147578)

brightsparkey 21st Jun 2018 4:02 pm

How do IF transformers work?
 
Although I understand transformers on a basic level, it seems that they really are at the core (no pun intended..) of valve receiver design and I'm not sure I completely understand how the various types of transformer in these old radios are arranged.

I haven't had a dead enough receiver to start pulling the cans off and taking them apart, so I was wondering if there are folk here who are skilled in the art of transformer coupled amplifier design, who could explain at a practical level whats involved?

I see various arrangements of stages - my SABA IF transformers have three cores to adjust - one for the resonance at each side and one for 'koppelung' , i.e. for coupling. The Grundig , on the other hand has only 2 adjustments, one core on each end of the former. Whats the difference, internally?

How much coupling is there between the coils and how are the stages designed as a total filter?

Kevin.

PS - Transformers are out of favour these days, and I'm only really familiar with power transformers in DC-DC converters etc., hence my ignorance :-)

paulsherwin 21st Jun 2018 4:19 pm

Re: How do IF transformers work?
 
They are just tuned circuits. Superhet radios work by converting the received frequency to a fixed intermediate frequency and amplifying that through one or more stages, each with a tuned transformer. This allows high selectivity to be achieved without exotic and expensive RF design.

As you have noticed, there is quite a bit of variation in the detailed transformer construction.

brightsparkey 21st Jun 2018 4:40 pm

Re: How do IF transformers work?
 
At this level I am completely comfortable. Its the details I'm curious about. The winding arrangements, the coupling arrangements, why adjustable cores not caps, why reluctance tuning for FM, why synchronously tuned stages rather than a careful shape factor,etc.

Thanks for moving the thread, btw. - still getting used to the structure of the forum.

Kevin.

G4_Pete 21st Jun 2018 4:56 pm

Re: How do IF transformers work?
 
Hi Kevin, I think a good place to start is the history of RCA and Westinghouse as much of what you ask is buried in patents and general history from radio year dot onwards.
Pete

ms660 21st Jun 2018 5:08 pm

Re: How do IF transformers work?
 
RDH4 is always worth consulting, try chapter 11 for some IF transformer and other coil info:

http://www.paleoelectronics.com/RDH4/

Lawrence.

jjl 21st Jun 2018 5:13 pm

Re: How do IF transformers work?
 
Quote:

Originally Posted by brightsparkey (Post 1053581)
At this level I am completely comfortable. Its the details I'm curious about. The winding arrangements, the coupling arrangements, why adjustable cores not caps, why reluctance tuning for FM, why synchronously tuned stages rather than a careful shape factor,etc.

Kevin.

Earlier superhets (pre WWII) used air cored IF transformers with variable trimmer capacitors. The change to adjustable ferrite cores with fixed capacitors came about when suitable magnetic materials became available and had the advantages of better mechanical stability and less susceptibility to changes in temperature and most likely lower cost.

VHF tuning was often, but not always, performed by varying inductance due to similar mechanical / temperature stability considerations.

John

brightsparkey 21st Jun 2018 5:28 pm

Re: How do IF transformers work?
 
Hi Lawrence - yes there looks to be lots of good material there. Many thanks - bedtime reading for some time! Already spotted a few answers to my questions..

Kevin.

brightsparkey 21st Jun 2018 5:33 pm

Re: How do IF transformers work?
 
I can see that air-cored coils would be large, and would struggle for Q because of the long windings to get as much as mHs of inductance.

Also there would be a problem of the extent of the fields, so any screening can would be large compared with a ferrite or iron powder cored inductor.

Kevin.

kalee20 21st Jun 2018 6:59 pm

Re: How do IF transformers work?
 
IFT's in valve receivers are generally two tuned circuits, loosely coupled. As distinct from a power transformer (where the inductance is generally immaterial as long as it is big), the IFT has precise values of inductance so as to resonate with a parallel capacitor.

The driving valve is a pentode, so it hits the IFT with a current that is sensibly independent of the valve's load. As the IFT primary is an LC tuned circuit, at low frequencies the impedance is low (because of the L). At high frequencies it's low (because of the C). At an intermediate frequency, it is really high, so maximum voltage appears across it at this frequency.

To be sure that the peak frequency is the figure that you want it to be, either the L has to be variable, or the C. Both have been used, but trimmer capacitors are very slightly less stable than the trimmer inductance. Additionally, to keep size down, a ferrite or iron-dust core is used rather than a larger, purely air-cored thing, so why not make it adjustable and replace the expensive trimmer capacitor with a cheaper fixed one?

The secondary winding is loosely coupled to the primary. This means that the secondary winding is excited by the field produced by the primary, but the primary isn't affected massively by the presence of the secondary. (As coupling increases, this approximation becomes less valid, and weird effects happen, such as a double-peak in response. The maths is a bit heavy).

The secondary winding is also tuned by a parallel capacitor, and the same thing happens as for the primary: the current in the secondary is largely independent of the actual secondary winding load. So, same as the primary, the secondary can be tuned.

This results in essentially an asymmetric response: significantly below resonant frequency, the secondary voltage drops at the rate of 6dB/octave (at low frequencies, forget the parallel capacitors and you have a primary coil being hit by a constant AC current from the valve, so you have constant AC field. In the nearby secondary, induced voltage is proportional to rate of change of field, which is thus proportional to frequency). But significantly ABOVE resonant frequency, it drops at the rate of 18dB/octave (at high frequencies, the primary capacitor dominates: the primary circuit is hit by a constant AC current from the valve, so the voltage across the capacitor is inversely proportional to frequency. Thus the voltage that the primary winding sees is also inversely proportional to frequency and thus the primary current is inversely proportional to frequency-squared. And now the secondary voltage is proportional to rate of change of primary current, ie proportional to frequency of primary current, so you 'get back' to secondary voltage being inversely proportional to frequency. But now you have the secondary's own inductance and parallel capacitance, which form an LC 2-pole roll-off, giving an overall response inversely proportional to frequency-cubed. Phew). However, near resonance, it turns out that the response is reasonably symmetric either side of the peak frequency.

The coupling between the coils is not large. In a power transformer, it is typically 0.995, meaning that 99.5% of the magnetic flux produced by a current in the primary, links to the secondary winding. In an IFT it is much, much less, maybe 10%. This allows the primary and secondary to be independently tuned, giving better selectivity.

Adjusting the two coils by a moveable slug in each, allows each winding to be tuned, but unfortunately the coupling between them is affected by the slugs' positions. And if you alter the secondary slug, you will make a slight adjustment to primary inductance, because you are moving a piece of material in the vicinity of the primary, and vice-versa. So all adjustments are interdependent, and that is why some manuals instruct during alignment to repeat adjustments till no improvement is found.

It is possible to adjust coupling between windings by sliding the windings (with their slugs) closer together or farther apart. A more common way is to have a third winding between them, which is adjustable in some way, or by an intermediate slug, or by a shorting ring. But it gets very complicated. I can't believe that IFT's were designed by formulae; I reckon that approximate figures were calculated and then the things optimised by sliding the windings to get the required response, then the positions measured and recorded on a drawing which was used then to make winding jigs.

PJL 21st Jun 2018 7:06 pm

Re: How do IF transformers work?
 
The Decca Decola tuner has a mechanism to vary the coupling of all IF transformers simultaneously to control the bandwidth.

G6Tanuki 21st Jun 2018 7:20 pm

Re: How do IF transformers work?
 
2 Attachment(s)
Some IFTs deliberately orient the axes of the two tuned-circuits at 90 degrees to each-other to minimise magnetic coupling.

Then they provide coupling by way of a couple of turns of the secondary winding [bottom end] being wound round the primary core.

Other types do the 'minimal coupling between the cores by orientation' thing but then provide controlled coupling by way of a link-winding [a few turns] round each coil. Switchable resistors in series with the link-coupling can then provide 'loose' coupling for optimum selectivity at the cost of audio quality, or 'overcoupled' for greater bandwidth and better audio quality.


In "communications" gear you are likely to come across IFTs with the secondary winding either capacitively or inductively centre-tapped for use in crystal-filters - see the attached. These give much greater selectivity.

ms660 21st Jun 2018 7:34 pm

Re: How do IF transformers work?
 
The IFT coils in the R1155 receiver were well out of skew with each other, virtually no inductive coupling, C coupling was used.

Lawrence.

Radio Wrangler 21st Jun 2018 7:46 pm

Re: How do IF transformers work?
 
The output impedance of a valve's anode if fairly high, so this allows the tuned primary of the IFT to have quite a high Q (If the coil design is goos)

The input impedance of the next stage's grid is also high, and this allows the tuned secondary of the IFT to have quite a high Q as well.

High Q means narrow, sharp resonances.

The two resonators thus formed couple together magnetically, depending on the spacing between the separate windings.

This structure gibes an 'M' shaped frequency response. There are two bumps, but don't think one bump represents the primary resonance and the other the secondary. No, the two resonators interact through the coupling and each contributes to both bumps. If you twiddle one slug it doesn't move one bump alone, it tends to put a slope across the bumps, if you then tune the other slug to level them up again, you now have the original 'M' shape, but now shifted a bit in frequency.

If you vary the spacing of the coils, and so change their coupling, tighter coupling makes the filter wider. The bumps both move apart and the dip in the M gets more pronounced. If you want this filter to stay wide, but have a less dipped centre, you should load the two resonators with a lower total effective resistance on each and the M dip will flatten.

If you move the coils apart,the coupling gets less and the filter narrows. The M shape gets less pronounced and the loss through the filter gets worse. The filter starts to look like a single peak. If I want to put a bit of the M-isn-ness back into it to resemble a flatter top, I need to increase the Q of the two resonators.

For a simple single lump shape as narrow as possible, you just peak the output with both slugs. The designer set the coil spacing and the resonator's loads. This does for a basic superhet struggling to get all the selectivity it can.

If you do want a flat top, then you have to mess about with the resonator Qs and the coupling.

With the right maths, you can design filter blocks with several resonators, using the source and load impedances to control the Q of the end stages, then everything else is done by calculating coupling factors. The resonator tuned frequencies then have to be trimmed to correct for the influences of the coupling components.

The maths is in "Handbook of filter synthesis" by Anatol I Zverev/ Wiley, 1967 and no, it's not out of date. There are several more recent books, but largely they are about how to interpret what Zverev says. A copy of Zverev will scare the living dylights out of many professional engineers. They mostly use computer filter design applications based on Zverev. This frees them from having to understand the maths and allows them to have filters where they don't have a clue about how they work :-). Once you've got a basic understanding, you can just go straight to the pre-computed tables he handily gives for a lot of filters. Just scale them to fit what you want. Knowing how to wade into the maths and calculate values for something not in the tables qualifies you as a mage of the sixth order. Being able to go fully off-piste and create things not already in the published maths gets you 8th order an your name carved in the hall of fame. Like martial artists proud of how grey and faded their once black belt has become, RF designers like tatty copies of this book.

David

Radio Wrangler 21st Jun 2018 7:47 pm

Re: How do IF transformers work?
 
Oh, and some Eddytone receivers (like EA12) have cams and levers to move coils closer or further apart within the IFTs to give a variable bandwidth control.

David

Ed_Dinning 21st Jun 2018 8:17 pm

Re: How do IF transformers work?
 
Hi Kevin, lots of stuff in Radio Designers Handbook (Langford Smith) but not as well stated as Kalee has put it.
For experiments see if you can get a Q meter. The Advance T2 often goes cheaply and it is a great piece of kit when working with coils.

Ed

PJL 21st Jun 2018 8:21 pm

Re: How do IF transformers work?
 
I always thought:
The primary and secondary are two separate tuned circuits.
Put them close together and they are magnetically coupled and behave as a single tuned circuit
Get them just right and they are coupled largely by radiation and the primary tuned circuit and secondary tuned circuit operate independently and you have the benefit of attenuation by both.
Get them too far apart and the losses become large.

ms660 21st Jun 2018 8:37 pm

Re: How do IF transformers work?
 
For a bit more in depth on coupled tuned circuits there's chapter 9 in RDH4:

http://www.paleoelectronics.com/RDH4/CHAPTR09.PDF

Lawrence.

G6Tanuki 21st Jun 2018 8:53 pm

Re: How do IF transformers work?
 
Quote:

Originally Posted by PJL (Post 1053668)
I always thought:
The primary and secondary are two separate tuned circuits.
Put them close together and they are magnetically coupled and behave as a single tuned circuit
Get them just right and they are coupled largely by radiation and the primary tuned circuit and secondary tuned circuit operate independently and you have the benefit of attenuation by both.
Get them too far apart and the losses become large.

A pair of 'overcoupled' tuned-circuits - even if they're individually tuned to the same frequency - actually results in a setup where the pair concurrently tune to two frequencies, each slightly either side of the frequency they would tune to separately.

This explains the 'double humped' or 'M' response that some IFTs exhibit. Loved by the hifi types who want wide bandwidth, detested by us short-wave-DX types who want good adjacent-channel selectivity even if it means losing any audio above 3KHz.

'LIVEWIRE?' 21st Jun 2018 8:56 pm

Re: How do IF transformers work?
 
Many years ago now, I scrapped a pre-war Philco (wish I hadn't now!) in which the I.F.Ts were wound on lengths of wood. Adjustment was via two Compression Trimmers, one for the primary and the other for the secondary. I've no idea what the I.F. was, and have long since forgotten what model radiom, so can't look it up!

turretslug 21st Jun 2018 9:21 pm

Re: How do IF transformers work?
 
kalee20 and Radio Wrangler really ought to get a consultancy whip-round from the rest of us for this!

The link-coupling method of a few turns over the complimentary winding in an IFT mentioned by G6Tanuki in post #11 was popular in communications receivers to give a choice of several different bandwidths, but it does bring a looming problem in vintage kit with elderly insulation- one (anode-connected) coil has a couple of hundred volts or so of HT on it, the other (grid-connected) has typically a few volts negative on it from AGC action. Comms receivers frequently provide for the AGC to be switched off- with maybe 2 RF stages and 2 or 3 IF stages and a strong incoming signal, an enormous voltage at IF can appear at the final IF stage, possibly inducing flash-over in the overlaid winding, likely ruining the IFT and maybe destroying the final IF valve by applying HT to the grid with gross over-running. The epochal and brilliant-in-so-many ways RCA AR88 guarded against this with a grid-leak provision in the 3rd IF stage, giving a sort of emergency biasing-back if a large signal appeared but many other high-gain sets didn't. The GEC BRT400 receiver had a sophisticated amplified AGC system that could malfunction, and a flashed-over and ruined final IFT isn't unusual in this particular set- notorious for carbonising its innards at the best of times anyway!

So link-coupled IFTs in elderly comms receivers need nursing and care with the use of that AGC switch.... The physically-moveable IFT coupling arrangement mentioned upthread in Eddystone receivers avoids this hazard, but brings other problems in a neglected set- seizures, squeaks and scrapes.

The 1930s were a heyday of ingenuity and cleverness in the superhet radio and a few featured triple-tuned IFTs- primary, secondary and a third, electrically unconnected winding to enhance selectivity characteristics.


All times are GMT +1. The time now is 12:39 pm.

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