Image frequency ratio.

[ukcol]

I think I may be missing some here.

It is in the nature of the superhet receiver that 2 different radio frequencies can produce a valid IF frequency. One of those frequencies is, of course, the desired station and the other is the so called “image” frequency. To give an example, if the desired station is at 1Mhz in the medium wave band and the receiver’s IF is 470kHz then using a “high side” local oscillator (1.47Mhz) this would give an image frequency of 1.94Mhz. The rejection of the image frequency is a function of the RF circuit(s) that precede the mixer and the “image rejection ratio” is the measure of the effectiveness of this rejection. This ratio is usually expressed in decibels (dB) but can (as in the BBC RD report below) be expressed as a straight numerical ratio and is stated at the radio frequency concerned.


In BBC Research Department report No. G033 on the Murphy A40C (May 1945):

"2.1.

Image or Second Channel Ratios.

These are poor as the intermediate frequency is only 119kc/s. The short wave image ratios are, however, fairly good owing to the double superheterodyne principle involved."

Fair enough so far, but later in the report:

"IMAGE RATIO

200 kc/s......................................160/1
1 Mc/s.........................................750/1
12Mc/s........................................30/1"

The 30/1 figure for the short wave performance seems to me to be at odds with the statement above.


How the BBC RD took image performance figures at the time is given in their report No. G.024 (September 1944):

"2.3.

SECOND CHANNEL OR IMAGE FREQUENCY RATIO.

The image or second channel ratio is the discrimination of a superheterodyne receiver between a wanted signal to which the receiver is tuned and an unwanted signal at a frequency twice the intermediate frequency away. It is evident that this discrimination is supplied by the radio frequency tuned circuits only. This image ratio is simply the selectivity of these circuits at a certain definite off tune frequency, namely resonance plus or minus twice the intermediate frequency: “plus” if the [local] oscillator frequency is greater than signal frequency, “minus” if the other case. This measurement is done at the mid frequency of each tuning band.
The receiver is tuned to the signal generator output, which is modulated to 30% by 400 c/s. The input to the receiver is adjusted to give the standard output. The signal generator is then set to the image frequency, and the output from it increased until the standard receiver output is again obtained. The ration of the two signal generator outputs is the image ratio. The receiver is fed through the standard dummy aerial………….."


The statement in G033 above that the short wave image ratios are "fairly good" seems to me inconsistent with the 30/1 figure. If I understand these ratios correctly the short wave ratios are the poorest of the three given or perhaps they simply mean that without the double conversion the ratios for SW would have been even worse?


Image ratios figures are not something I have had to work with a lot and perhaps part of my problem is that is that I don't have a feel for the norm; perhaps an image ratio of 30/1 at 12 Mc/s is quite good?

[G6Tanuki]

What's the first-IF when the radio's in use on shortwave?

If this is known, and you can make an educated guess at the likely "Q" of the signal-frequency tuned-circuit, a bit of curve-plotting will give the likely image-rejection. I suspect a typical broadcast-receiver with wax-coated enamelled wire wound on a paxolin former would have struggled to get a front-end "Q" much above 50!

I feel that the review is saying that 30:1 is better than a typical equivalent receiver with a similar very-low-IF would have achieved at 12MHz.

Even with a 450-470KHz IF, SW receivers with just a single signal-frequency tuned-circuit often had major issues with image-rejection above 10MHz; tracking/alignment on such receivers was never that good, and absent a separate front-panel antenna-circuit-peaking control the loading effect of the antenna on its tuned-circuit guaranteed a low Q, making images a big problem.

I suppose images did help a receiver to sound 'lively' though - receiving every station at two places on the dial!

[ms660]

The lowest SW 1st IF that can be tuned is approx. 0.546 MHz and the highest is approx. 1.5 MHz, the main SW tuning is the band set, the MW tuning is the bandspread, so far as I can make out.

Lawrence.

[ukcol]

The A40C (the set I am particularly interested in this context) uses the MW broadcast band as the first IF. It takes the middle of each of the 13, 16, 19, 21, 25, 31, 42, & 49 metre bands and down converts it to 1Mhz. Thus each of these bands when converted are tuned by the MW broadcast band range of the set. Obviously because this first IF is not a narrow bandwidth IF the image rejection has to be as least as bad as the MW band proper. The main merit of this particular double conversion design is that is enables a very effective electrical (as distinct from mechanical) short wave bandspread.

Thanks for the replies. When I saw that 30/1 image figure for 12Mhz I thought perhaps I had misunderstood image ratio notation completely; I am suitably reassured.

[ukcol]

Another question if I may. Would those image ratios above be power or voltage ratios? It makes a big difference when converting them to decibels.

[Radio Wrangler]

This set is a double superhet, so it has multiple images (and even image pairs of the image from each mixer!)

The need for a superhet down-conversion process comes from the difficulty of getting sufficient selectivity at higher frequencies. LC filters can be made appropriate for SSB midulation up to around 100kHz, and for CW to somewhat less. Filters for AM can be made to about 600kHz. You can go a bit higher on all these but it starts to cost you in money and insertion loss. Crystal filters can do CW bandwidths from 400kHz up to 20MHz, they can do SSB over a similar range. AM is wide enough to become difficult but can do 5MHz to 30MHz. These figures are just rough guides. Outside them the cost and difficulty accelerate upwards.

So however many stages of superhet conversion you choose, your final IF has to land on a frequency where there are filter techniques for your chosen modes.

Having chosen a final IF, the mixer before it will give an image pair spaced twice the IF apart. If the LO is topside of your wanted frequency, the image is topside of that. If the LO is below, then the image is below that. If the image is far enough away that it is feasible to builda filter ath the incoming RF frequency which will attenuate signals on the image enough, then a single superhet will do for you. If the filter demanded is too difficult to make tuneable on that band, then you are in double superhet territory... or higher!

Your second IF has the ultimate selectivity for the mode. The first IF is high enough that the images at the RF incoming frequency can be suppressed, and yet low enough that the image of the second mixer can be suppressed. I've worked on a receiver with six conversions. You either add IFs to get what you need, or else you need a wondrous filtering technique to break out of the compromises. That 6-fold superhet used LC filters because quartz couldn't be got linear enough and an all L-C structure ensued.

So to consider the Murphy double superhet n terms of image rejection ratios, you need to do it twice. Consider the rejection of the first mixer's image, then that of the second mixer's image.

David

[Radio Wrangler]

It is important if those ratios are voltage or power ratios, and they rarely say.

If stated in dB, then those should be power ratios.

When someone starts talking about "voltage decibels" walk away The decibel is defined as a ratio of two powers. Some people talk of gain calculated from a voltage ratio and forget that impedance levels may have changed.

Used properly, these things are very helpful, but not everyone uses them properly.

David

[turretslug]

There were a few car radio SW converters that used a similar concept to the Murphy set discussed of translating a given chunk of SW spectrum to the MW band- it's a straightforward and appealing idea for getting good SW coverage and bandspread, but must be pretty demanding of interconnect screening integrity and front-end bandpass characteristics to avoid the "too many stations...." syndrome. They generally had a series of pushbuttons for the various SW BC bands, again like the Murphy with RF tuning pre-set on the middle of the band- one manual tuning knob on a car radio is enough of a distraction without having a pre-selector to play with on the convertor....

In a way, the pre-war Murphy's use of the MW range as a tuneable IF anticipated the architecture of many post-war comms receivers that adopted a tuneable IF over a span somewhere in the MF region following a front-end that translated switchable (arguably tunable in the case of Wadley sets!) chunks of HF spectrum down to MF. So it perhaps merits an honourable mention in receiver evolution.

[ms660]

Those ratios might be voltage ratios.

Lawrence.

[ms660]

There's another BBC report for the A40c (1938) the results are in table 1 towards the end, it gives image frequency results in decibels, one frequency for LW and two for MW:

http://downloads.bbc.co.uk/rd/pubs/reports/1938-12.pdf

Lawrence.

[ukcol]

I have a copy of that report here Lawrence but as you say they only gave figures for MW and LW. Unfortunately the BBC weren't interested in the 40's SW performance for the particular application the report addressed.

However, comparing the dB figures for the unmodified A40C in the G015 table with the numerical ratios given in G033 indicates that the latter must indeed have been voltage ratios. Thanks for drawing my attention to that.



BTW. The very fact that the BBC were even considering a domestic receiver (albeit one modified with an extra RF stage) for the roll of a re-broadcast receiver, is a testament to the 40's technical reputation.

[Hartley118]

Originally Posted by Radio Wrangler

It is important if those ratios are voltage or power ratios, and they rarely say.

If stated in dB, then those should be power ratios.

When someone starts talking about "voltage decibels" walk away

David

Is it any worse than quoting dB sound pressure level I wonder?

Martin