Low Pass filtering top range of a receiver?

[rwgast_lowlevellogicdesin]

This may be the stupedist question ever, but this is one of those times I would rather ask than build a test... So lets say you have an RTL SDR as a receiver and its tested to cut off at 1733mhz. I was watching some videos for a guy who makes band pass filters for the L band, hydrogen line etc etc and all his filters cut off way past 1700Mhz more like 2.4/3ghz and hes talking about noise in those bands, yet his gear is targeted at DVB Stick users for those most part.

So would adding a 1700mhz LP assuming no IL do anything for noise? I remember when I use to have my router near an SDR dongle I would get all kinds of crap but I always though it came from PSU harmonics or something not the actual 2.4ghz signal.

[donmr]

So would adding a 1700mhz LP assuming no IL do anything for noise? I remember when I use to have my router near an SDR dongle I would get all kinds of crap but I always though it came from PSU harmonics or something not the actual 2.4ghz signal.

It all depends on where your noise is coming from.

The "best" filter to reduce noise is a bandpass filter for the frequency range you are using.  This is because interference/noise can be at any frequency above or below that.

Simple lowpass filters are usually put on transmitters to suppress harmonics from the transmitter which are "usually" above the carrier.

[rwgast_lowlevellogicdesin]

No no I understand that, im just wondering in theory if a lowpass filter at the top frequency of a receivers range can help lower noise even though the reciver can not receive signal above its top frequency.

[T3sl4co1l]

No no I understand that, im just wondering in theory if a lowpass filter at the top frequency of a receivers range can help lower noise even though the reciver can not receive signal above its top frequency.

Well, by definition you're saying it can...

If you're getting aliasing, then that's receiving.  A legitimate method, after all -- the time-domain version is called equivalent time sampling.

Aliasing is only useful with an image rejection filter.  In which case, the available bandwidth of the receiver is that of the filter, and the sampled bandwidth is half the sample rate (whichever one is lower).

If it does reject such frequencies, then it necessarily is filtered, at least to some extent (maybe the gentle RC rolloff of the ADC sample-and-hold, rather than a practical high-order one).

If it's susceptible rather than sensitive to such frequencies, then it may be that the linear signal path rejects those frequencies well, but nonlinearities cause IMD that mixes the interfering signal into the passband anyway, at least when sufficiently strong.

Dealing with strong sources of interference often needs steep filters (are you filtering out 1V/m or 100V/m?), so you're back to the same problem, except it doesn't show up in the linear range because of filtering and processing.

An extreme example is a photodiode: it acts as a diode-diode for frequencies up to some MHz (depending on recovery time), but it acts as a photo-diode only at 300-600THz.  That's a huge bandwidth, at a huge center frequency.  It's not quite fair to say it's rectifying all the same, but it is a square-law sort of detector, in either case.

Only use the bandwidth you need! Filter everything else!

Tim

[rwgast_lowlevellogicdesin]

Well i would like to keep the RTL broadband for browsing, at least this one. BUT I would like to filter out FM and All TV Broadcast along with a 1700mhz Low pass. I just finished building a 5 pole LP filter for the LO and it helped a lot.

The issue is all these filters on the front end will cause a ton of insertion loss. My homemade planar EM 1090 filter has a 2.5DB loss alone... is there some way to filter out all this crap at once without out to much insertion lost?

[uncle_bob]

Well i would like to keep the RTL broadband for browsing, at least this one. BUT I would like to filter out FM and All TV Broadcast along with a 1700mhz Low pass. I just finished building a 5 pole LP filter for the LO and it helped a lot.

The issue is all these filters on the front end will cause a ton of insertion loss. My homemade planar EM 1090 filter has a 2.5DB loss alone... is there some way to filter out all this crap at once without out to much insertion lost?

Hi

It is quite possible to design a rejection filter with < 0.5 db of loss. It's dependent on the quality of parts used and the layout of the filter.

Bob

[rwgast_lowlevellogicdesin]

So where would you start making a low loss 1700mhz LP filter? Obviously lumped element is going to be out of the question maybe a planar on some kind of exotic PCB material?

[uncle_bob]

So where would you start making a low loss 1700mhz LP filter? Obviously lumped element is going to be out of the question maybe a planar on some kind of exotic PCB material?

Hi

Why would lumped element be out of the question? You need high Q coils and caps, but they both exist. You also should be careful about pcb layout issues.

Bob