Nooelect Flamingo AM BC Filter results

[KA1GJU]

In an effort to keep a daytime, 5KW AM radio station on 1540kHz 5.1miles (8.2km) away from me appearing in numerous locations in the upper bands, I thought I would give the Nooelec Flamingo AM BC filter a try. Thinking all my SDR's at three entirely different locations and antennas were being subjected to overload suddenly, starting a few months ago. The filter is here:

https://www.nooelec.com/store/flamingo-am.html?srsltid=AfmBOorQBl3XhX_EfBwc1SL6IRCxGcpGrzZSCdKZjMy_C5IvWI_oIgmr#product-tabs

Their description states the following:
We designed Flamingo AM to provide sufficient attenuation for broadcast AM frequencies (>40dB typical) while ensuring adjacent bands, such as 160m, are minimally affected. The -3dB rolloff of the filter is 350kHz and 1900kHz. Minimal out-of-band insertion loss means the filter can stay in place for most any application, though we do recommend removing Distill:AM from your setup when not listening to HF frequencies. As a true bandstop filter, you are able to pass-through DC (bias power) when it is required.

In the 'Downloads' section I found the following which looked promising:
https://www.nooelec.com/store/downloads/dl/file/id/78/product/290/distill_am_dc_to_5mhz_sweep.png

Using the SDR-IP, I took screen shots of the entire AM band and shots of 3080kHz where AM1540's second harmonic has a strong presence. There's also a shot of the third harmonic on 4620kHz with the filter in line. (No image taken w/out, sorry!) Since the 3080kHz signal is still there and on the third harmonic of 4620kHz with the filter inline, wouldn't that tell me that the radio station has some issues and not my equipment?

Apparently the order of the images gets messed up once submitted, but the comments are underneath the image!

[RadioSDR1]

I might be missing something here. With that disclaimer out of the way , if the filter is said to provide a typical attenuation in excess of 40db, it does not appear you are getting that. For instance on 590 the signal is -24.7 dBm without the filter but only down to -32.7 dBm with the filter. That's only 8 db down. With 40db of attenuation, I would expect the meter to show in excess of -60dBm. Right?

I use the RTL-SDR bandstop filter and it does better way better than 8 dB. Between that and the AM filter in my current radio RSPDX-R2, those in tandem completely wipe out my entire AM band including a 50 KW AM station only 11 miles away. Well, some carriers survive all that but not the modulation.

[KA1GJU]

That's what I was thinking, but I'm not an RF Engineer nor do I play one on TV. Per the specs, 590kHz should have ~10dB attenuation. Stupid me put the frequency on 590kHz instead of the problematic station on 1540kHz. I will take screen shots this AM once they jump to high power with the tuned freq of 1540kHz for better reference.
The attached image is the published filter specs with my edits of 590 and 1540kHz arrows.

One would think 1540kHz would have a sizeable drop. From what I see in the carrier signal strength, it goes off the chart (-30db?) to ~-41dB.
If I can only find my W1VLF filters I made a few years ago that I used for the RSP1A that needed them. Those worked great! The schematic is attached too.

[RadioSDR1]

I totally get it. I'm no RF engineer either. Just for reference, I'm including screen shots of my filters. I will show my AM band with no filters, with the built in RSPDX-R2 filter with the RTL-SDR.com "Broadcast AM Reject Filter" and with both filters running in tandem. With mine I can wipe out the entire AM band including several 50 KW transmitters 10-25 miles away with no meaningful affect on HF though the RTL-SDR.com filter does attenuate 160m too much.

One way or the other, I think you should eventually be able to have your cake and eat it too with respect to nailing down the appropriate filtering cocktail. Though if the filter you have is really giving you the stated attenuation on the broadcast frequency of trouble, then you may have in fact discovered harmonic problems from the broadcast transmitter. Though if the frequency of trouble is not completely eliminated, you may be observing high order mixing products. My usual test for that is to change the bandwidth and/or the center VFO frequency on the bottom slide bar of SDRC. If the suspect signal changes frequency, then it's a high order mixing product. If the frequency stays the same, then it's a real signal or sometimes an image. Or I suppose possibly a transmitter harmonic too, depending.

[jdow]

Hm, I thought I put a reply in there - a LONG reply. I noted such things as 4620 is not a harmonic of 590 so this squirrel chase is off on wrong feet to begin with.

Odds are you get WAY more than 8dB of BCB reduction from the Flamingo filter. I have one and it works as expected. When your system mentions "overload" forget about readings having any relationship to reality.

1) Put in an honest attenuator. Adjust it until the 590 kHz signal is at least 12 dB from the top of the scale for starters. What do you see on the various frequencies you tested? For this test set visual gain to 0 dB.

2) Add the BCB attenuator and check what happens to BCB signal levels when the system has a chance of being right for relative measurements. And, what do you see on the various frequencies you tested if you saw anything there in test 1.

3) If you still see the spurs, sit down with a calculator and some paper (or modern equivalents are allowed of course.) Note the frequencies of the highest level signals. Pick two at a time, F1 and F2. Calculate N*F1 + M*F2 for M and N integers from -3 to 3 such as 2*F1 -1*F2. (N = 2 M = -1 in this example.) Look for a match to the spurs you found. Then go looking for the rusty metal rubbing up against corroded metal that is generating your spur.

4) With the attenuator setting as in 1 with or without BCB filter, disconnect the antenna. If the noise goes down likely you are done - optimized for at least 2.5 MHz and lower. Check that a weak signal's SNR does not go down with 10 dB more attenuation. The system should be as good as it gets for SNR up until disconnecting the antenna makes less than a 6 dB difference in SNR on a weak signal.

BTW - the problem is overloading the A/D converter with the voltage sum of all the signals it can see. That is why setting to about 16 dB below the top line is a good idea.

So that's the homework for today.

{^_^}

[gm8arv]

Mine is as bad as the OP mentioned. Very little attenuation at the low end.

[KA1GJU]

Unfortunately I was tuned to 590 during the photo shoot, which is another strong signal in the AM band. Didn't know about it until the photos were taken. My issue is with 1540, 3080, and 4620kHz, which is leading me to issues with the 1540 kHz TXR. I had an issue about a year ago, but it went away... I think. But this one I can hear on my rice box radios too on 3080 and to a less extent on 4620kHz.

On VC in KP2 land and will worry about it when back from VC

73 Kriss KP2/KA1GJU

[jdow]

Ah, I have a Kenwood TH-F6 handheld I use for finding where the harmonics are coming from by driving around. It MAY be, though, that 1540 has a bad joint on their tower. They might want to know you are fighting harmonics on their signal if multiple radios get it. Heck, you might be able to teach them how to use your large collection of radios with a copy of SDRC to check for harmonics, recreationally of course. That would help you and help them.

{^_^}