Image frequency math on Tecsun PL-880

[VK-SWL6002]

Hello.. this is my first post so if I make any errors of etiquette please let me know.  

I am trying to work out the math of inter-modulation frequencies using my Tecsun PL-880 as an example.  I have been SWL on my new Tecsun PL-880 (loving it) using a 17.1m random wire antenna with 9:1 unun and getting plenty of signal.  Even setting the SW antenna gain to 'local', I seem to still be getting quite strong frequency images of the local AM radio stations.  

I have been doing a lot of research, spreadsheeting and math, but can't quite work out the exact math of why the image frequencies are where they are.  

I know the PL-880 uses an IF of 45 kHz and 10.7 MHz, but can't quite get the math to replicate why for example, I am receiving images of 720am (station 6WF here in Perth, Australia) on 6525,7675,8155,12870,15590,16070.  

One clue is the frequency difference between all images is one third of the frequency difference between the original stations.  For example, the difference between original station 720 6WF and 6RN 810 kHz is 90kHz, but the difference between images of those stations (eg 15560 for 810 and 15590 for 720 ) is one third of that .. ie 30kHz.  This one third spread is apparent right across the images, ie the 495kHz MW band spread from say 585 6NR to 1080 6IX is three times of that in the image (say 15470 (6IX image) - 15635 (6NR image) = 165kHz or 1/3 of 495kHz!

Another clue.  The spread between 720 6WF and one of its images at 16070, is 15530, half of which is 7675 which is also an image of 6WF
Another clue.  The mid point between 720 6WF and one if its images at 15590, is 8155, which is also an image (15590 - 720)/2 = 7435, and 7435 + 720 is 8155 (mid way between 720 and 15590).

If any expert with a little time on their hands can decyper the math (ie various integer multiples of IF +/- RF) I would be very interested to know!  I have spreadsheeted out all the images and still can't quite get integer multiples happening.

I hope someone can work it out, or point me to the math of an answer somewhere.  Thanks for your time!
CP  Perth, Australia

[VK4ADC]

Craig

Firstly, welcome to the forum. Your post is fine and people on here are great.

I will throw your maths even further into a tizzy.  I found a User Manual and right at the end in the specs is...

5. AM IF : 1st IF : 55.845 MHz
               2nd IF : 10.700 MHz
               3rd IF : 45 KHz

FM IF : 128 KHz

The technique of mixing up is quite conventional as it moves the first IF out of the tuning range hence less image issues. Eg for a receive frequency of 15.000 MHz, the LO will likely be at 15 + 55.845 = 70.845MHz, image at 70.845 + 55.845 = 126.770 MHz but simple filtering solves that one.  The thing about these high IF values is that there is only minor selectivity applied there.  More IF filtering occurs at 10.7 but the majority is like to happen at 45KHz via DSP in your unit. 

There are possibilities other than signal intermod. 

Straight front end overload is one - and one of the easiest ways to see if it is this is to attenuate the signals from the antenna where whey enter the radio. A variable potentiometer (/ pot) as a variable voltage divider is enough to find out.   Typically FE overload will stop sharply as the signal is attenuated (ie cut out as the pot shaft is rotated). Other desired signals will be there but weaker but the AM BC stations will disappear. 

There is a possibility of adding a High Pass Filter at 1.5-1.6MHz to attenuate BC signals but that isn't going to help if you are a BC band DXer. It is ok for HF - say 3MHz upwards.

Similarly, high Q notch filters in shunt across the antenna feed terminals can be used to attenuate the specific  BC station frequencies (one or both with a bit of finicky tuning). These are simply an LC series circuit that you tune for minimum signal at the BC stations frequency.  

You don't mention the how far physically you are away from the TXs at 720 KHz and 810 KHz and that becomes a critical factor in how you might solve it.

Try the pot idea and advise if the signal problem slowly attenuates or simply cuts out.

73 Doug

[VK-SWL6002]

Many thanks Doug for your quick and informed details.  I'll certainly experiment with the voltage divider and the tuned LC.  Thanks also for the clue on how the 3 stage IF would work... perhaps that's a clue as to why the MW band image (or sprus?) spreads are 1/3 the original.  It makes sense the 3rd stage  45kHz stage would be on chip.. the data sheet for the Si4735chip talks about its 'low IF' stage.  The 6WF/6RN Tx is 50kW and 12km away.  Further attenuating the antenna on specific frequencies will be a good experiment.  Anyway.. plenty of good math and electrical avenues to investigate thanks Doug.
cheers Craig

[VK5PJ]

Hi Craig,
  working out all the possible answers for image freq can be quite a chore, you also have to consider if there is any bleed through to your 3rd IF (455KHz in you case) as it sounds like you are in an area getting extremely strong primary signals.

I only mention this as I had a VHF receiver listening on 52.525 MHz that used to burst into life with the Radio Australia standby tones (Waltzing Matilda) when they used one of their 10 MHz frequencies I think it was. (over 30 years ago now).  This shortwave transmission found its way through the VHF front end and ended up as an image in the I.F...   this took me weeks to work out how my VHF receiver was playing Waltzing Matilda to me at certain times of the day.

Good luck with your journey but be prepared for the possibility it is caused by another external factor like another AM station.

Peter, vk5pj

[VK4ADC]

Craig

A couple of other things come to mind:

1. The FS from the BBC stations is probably in the volts/metre range at about 1KM from the transmitter and attenuation at LF is not particularly high so you might still have hundreds of mV per metre at 12KM separation.  With 17M of antenna wire that again becomes Volts applied to the receiver input. An easy target for straight front end overload.  But the story doesn't finish there because that level can itself act as a local oscillator (LO) for the first mixer and hence more mixing products are involved. Further, you have two broadcast transmitters 90KHz apart and each one can act as a LO so even more mixing opportunities exist in your front end.

This is one of the reasons that most amateur radio equipment (rather than a SW receiver) uses front end bandpass filtering purely to reduce out-of-band signal levels.

2. The BC transmitters are close enough for you to hear transmitted intermodulation products at various points due to mixing at their base frequencies and then multiples of them.  This can occur through design but also due to the 'rusty nail effect', which simply means that the signals applied across a corroded joint (which acts as a diode)  causes mixing products. That 'rusty nail' could be rusty bolt on the transmitting array or even just near to it but the field strength is so high that the metalwork attached to it radiates sufficiently for you to hear it 12KM away, particularly given 17M of wire and a sensitive receiver.
(PS 17 metres of wire operated against virtual ground is resonant at about 4.3MHz)

The simplest way to isolate whether it is receive or transmitted intermodulation product is by substituting a different receiver, preferably using the same antenna.  Transmitted products will be there on the same frequencies regardless of the receiver design.  Received products (read overload into that too) will be only on the existing one provided the replacement has better front end characteristics.

You could also try taking your receiver and a moderate (eg 10 metre) length of wire (1) closer and (2) further away from the transmitter site. Go for a drive.. That helps obviate your wire antenna at home and its mounting as a 'rusty nail' intermod generator plus helps confirm whether it is front end overload (see my previous post re the pot to vary signal levels).

I hope that all makes sense to you.

Doug

[VK-SWL6002]

Thanks Doug and Peter for taking the time to provide me with some excellent avenues for enquiry.
Looks like its over-driving producing all sorts of spurs... just using them as a tool to help me understand superhet, antenna tuning etc.
I looked for the intermodulation products and they are certainly there (and all over the place)... eg at 1926kHz (a simple addition of 720 6WF and 1206 6TAB).
720 6WF is also crystal clear at 3 x 720 = 2160 kHz. Very interesting how these could also be acting as LO's.
An interesting example Peter of the VHF image of Radio Australia... thanks for the input.. I'll be thinking through its implications for my investigations.
I'll be sure to report back if I make any progress! Thanks again.

[VU2NAN]

Hi Craig,

Your's is a 'front-end overload' problem caused by strong local signals and an extra long external antenna.

The recommended external antenna for Tecsun radios is only 10 m long with a 10:1 balun and around 7 m of coax feeder. You would do well to use the recommended length.

Were the problem to persist (due to proximity of the local stations) then Doug's notch filter suggestion should do the trick.

The problem could also be accentuated by the 1N4148 diodes D1 & D2 (connected back to back from the external antenna terminal to ground to prevent front end ESD damage) driven to conduction by the strong local signals and acting as multipliers.

Regards,

Nandu.