FM and TV output amplifier combiner

[Leiser]

Hi,

I would like to know your opinion about the feasibility to combine the outputs of a FM and a DAB antenna LNAs. The idea is to use a single cable for both signals, using a combiner and a splitter at the receiver with lump components. I´ve never done it before.

I have some doubts about intermodulation between the two frequencies.

Any thoughts?

Thanks a lot

 

[logbook]

What specific frequency bands are you talking about? To separate widely separated frequencies with a diplexer at each end of the cable is not a big deal in itself, especially if they are off the shelf components.

 

[Leiser]

Hi,

The frequency bands are:

FM 88-108MHz
DAB 174-240MHz

The idea is to connect the amplifiers to the radio in a car using just a single cable instead of one for FM and another for DAB. I have done that for AM and FM and it works, using as a combiner a cap and an inductor. In this case we are thinking using the same cable for all the amplifiers, AM, FM, DAB and having a sppliter box at the receiver.

 

[logbook]

My first thought was a Mini-Circuits diplexer

http://www.mini-circuits.com/products/filters_sm_diplexer.html

but their surface mount offering only goes to a cross-over frequency of 60MHz and you need 137MHz.

I tried the free diplexer software suggested in an earlier posting, but the isolation is only around 15dB if 1dB in-band ripple is required for a 3 element filter (6 components in total).

 

[logbook]

http://www.breeeng.com/800509.html

looks interesting. Price is unknown and I have never dealt with this company.

 

[logbook]

http://www.satcure.co.uk/accs/page15.htm

"V24-310 FM+DAB+UHF triplexer. Combine all your aerial feeds into one cable. This can be fed into the DIP2 (above) to add the LNB feed as well. So all feeds can go down one cable to a decombiner wall plate."

Now this one is probably more suitable due to its cost (£9.95). It is pretty short on specs like most domestic TV kit.

 

[Leiser]

Logbook,

Thanks for the information!. What do you mean by cross-over frequency? how did you work out the 137MHz?

I run a simulation using 3rd order butterworth filters. So I was thinking building the duplexer with caps and inductors. In the simulation I get quite good isolation between bands.

Regards,

 

[logbook]

We want the the low pass filter response to be similar to the high pass filter response (only mirrored). We have a lower frequency that we need FL and an upper frequency that we need FU. We want geometric separation so the crossing point is sqrt(FL*FU).

The cross-over frequency is "half way" between the upper and lower bands but as a factor not as linear separation.

If the centre is say 50MHz you don't go 10MHz either way you go say 10% higher and 10% lower (geometric not arithmetic). By using the geometric centre between the two bands you get the highest percentage separation between the bands which gives the best isolation.

If you get acceptable separation and input VSWR with 3rd order filters then go for it.

 

[biff44]

Yeah, I would also suggest the diplexer. However, since the LNAs have a lot of gain, you could brute force the diplexer by just having a lowpass + 3 dB attenuator connect in a tee to the highpass + a 3 dB attenuator.

 

[biff44]

The main reason you need the diplexer arrangement is to not take a 3 dB hit on noise figure. A secondary reason is frequency band flatness.

 

[logbook]

biff44, we aren't going to get a 3dB hit on noise figure in this case because we are picking off after the LNA's. The noise figure will be essentially unchanged if the LNAs have more than say 13dB gain. However if we "brute force" the diplexer with a 3dB pad and a straight T we stand a good chance of back driving the LNA's badly enough to make them intermodulate in an unpleasant manner.