Second build has shipped. Message will appear here when store is ready for third build ordering.

Sensitivity below 700khz

I have two Kiwi's with the antenna jacks paralleled. My antenna is a W6PLVP loop being used to couple signal from a 240' tower. Antenna works great but the two SDR's have dismal sensitivity below about 700khz. I added a third SDR to the antenna and it is receiving stations very well on each frequency from 530 on up yet the Kiwi's hear few signals. Was not always this way. When I relocated the receivers, the coaxial connectors were torn from the boards and the screw terminals are now being used. Was there a software update that could have affected the sensitivity ? The S meter stays S9 across the AM band when the antenna is removed. S meter calibration is on 0db. That seems unusual. Anyone been down this road before ?

«1

Comments

  • did you try them one at a time, with no paralleling?

  • Not yet. But I did power each one down and it made no difference. I will try that next and report back.


  • Hmmm,

    >When I relocated the receivers, the coaxial connectors were torn from the boards and the screw terminals are now being used.

    I wonder if there has been other physical damage ?

    It seems to me that the input coupling capacitors have somehow become broken, and you are only receiving properly on the higher frequency bands due to the residual capacitance.

    I'd have a close look at the PCB under a magnifier to see if you can spot anything wrong.

    Regards,

    Martin

  • Thanks for the reply Martin. I wondered that as well but discounted it as the connectors were torn from the top of the board. Let me explain this event : I had some rather large RG6 cables into a T connector feeding each SDR. Been in the same place for years. When I wanted to move them to escape some new electrical interference, it didn't dawn on my that they were very small connectors until both were snapped off. As far as the matter of only higher frequencies passing, both Kiwi's get WWVB 60khz at 1000 miles fine all day. It is amazing how quickly the reception drops off. 700 WLW at 80 miles is about S9+40db and crystal clear while 650 WSM at 240 miles is inaudible yet very clear on the portable SDR fed from a T at the same point. It would be hard to build a high pass filter that would do what this is doing.

  • jksjks
    edited December 2023

    S9 with no antenna connected? That's not good.

    If you zoom all the way out, and enable the RF spectrum display ('s' key), what level is the noise floor? Is it constant from MW all the way to 30 MHz? Can you post a screenshot here?

    I wonder if an ADC output has become shorted high. Although the audio would be very distorted if that were true.

    I'm so glad we decided to go with thru-hole SMAs with Kiwi-2..

  • edited December 2023

    Hi John,

    With antenna and without. I am assuming that the SMA connector was not performing any internal ground function between pins 2 and 3 ?

  • jksjks
    edited December 2023

    Could there be a really bad SMPS that's been introduced somewhere recently?

    Here's a Kiwi with S8 noise all across the MW band. This is classic badly behaved SMPS behavior where the unstable SM carrier gets wider at each harmonic (see 200 - 500 kHz for example). And the harmonics get stronger too for some reason. Then they die down and come roaring back above 1.6M to wipe out 160m.

    I see this kind of behavior all the time on public Kiwis. Try and setup your spectrum to make a capture like this. Use the same frequency span, put the aperture setting into manual mode and use the 'S' key (capital S) to autoscale the waterfall/spectrum.


  • What is SMPS ? I assumed Power Supply and that seems to be where most of the noise originated. I bought a nice linear supply numerous years ago and watched my sensitivity slowly suffer. The supply was apparently going bad. Will take screenshots during the day and post after I try a few power supplies.

  • Sorry, yes, SMPS = switch mode power supply.


  • The power supply was the majority of it. The S Meter on this one is still above S9 with no antenna but here is that one's noise floor. The other one drops to an S5 without antenna. The one that drops to an S5 is more sensitive. I need to play with it some more and I am not sure that the good performance I had on the low end is still there now that I have both receivers running. Both Kiwis are set to 0 on S Meter Calibrate.

  • That doesn't look good.

    The S-Meter calibrate offset value should typically be around -13dB, rather than 0dB to provide a more accurate S-Meter reading, and a near calibrated dBm digital readout.

    However, that doesn't account for the residual high level of noise.

    Is the KiWi in a metal case or is it still in the original perspex housing ?

    This may sound obvious, but have you got the coax screen connected to the correct screw terminal ?

    The KiWi RF input is very sensitive to local noise sources. When you say that you have no antenna connected, do you mean that nothing is connected to the RF input terminal block, or do you still have a short length of coax connected ?

    Even a very, very short length of unscreened wire will pickup lots of local noise.

    Poor quality coax, with very little screen coverage, can also introduce lots of noise, especially if it runs alongside, or is bunched with, mains and data cables.

    Have you got any other items such as power supplies or electronic equipment in very close proximity to the KiWi ?

    Regards,

    Martin

  • I believe I am on the correct terminals with about 4-5" of wire (ground and +) from each Kiwi into the coax.

    Metal cases on both Kiwi's (both Kiwis are on top of each other). Fact is, when I had the bad power supply into one Kiwi, it raised the noise floor audibly on the other. I have quad shield RG6 from outdoors but the last 15" is RG58 onto a terminal block and then to the 4-5" wires to the Kiwi terminals.

    When I say no antenna, the RG58 and 4 inch wires are still connected so 'no antenna' is not technically accurate.

    The only thing close by is an IP based power off/on switcher at about 6 inches.

    Bob

  • Looking at the input terminals, I found that if I shorted the input (it had a 30 ohm resistor across it already to control signal levels), the signal was a lot cleaner. So I had overload issues as well. Now have 5.1 ohms across the terminals and here is the new spectrum with the 5 ohms to ground. Hearing Cuba on 530 from Indiana...that works.


  • Hi Bob,

    This is still odd.

    Get rid of those 4-5" pigtails if you can, as they will definitely be picking up noise.

    The overload problem is something different, but it's difficult to tell from the previous screen grabs.

    Is it not affecting your other 'good' KiWi too ?

    Any chance of another screen grab so that we can see the whole 0-30MHz spectrum with the antenna connected ?

    Regards,

    Martin

  • Hi Martin,

    This is this good SDR....want to make sure it is at full potential before introducing the other one back in the mix.


    19.8khz from Australia is in now so it is passing VLF well.


    Here is a 0-30mhz with antenna connected with 5.1 ohms to across the input to control overload. This was just a handy value I had. Apparently the W6LVP "loop" I am using as a coupler has a lot of signal output.

    Bob

  • You should consider using a simple T-attentuator rather than a parallel resister. The feedline VSWR and loop LNA would be happier.

  • Hi Bob,

    If you have your S-meter and waterfall calibration offset o a value of -13dB, then you shouldn't have any overload problems as long as all the signals in the spectrum display are under -30dB.

    I think you can safely decrease the amount of attenuation you have in place, and also consider using a splitter rather than a simple Tee, as suggested.

    This will provide more isolation between KiWi's and may further help to reduce noise ingress.

    Regards,

    Martin

  • Hi Martin,

    I have thought about a splitter but the lowest frequency they are rated for is 5mhz and I need one that will pass down to 20khz. I have not tried one to see if ii would work.

    I have the 'good' SDR working very well now but when I check the version of it, I get this response.

    The poorer performer has 1.646 which is current. Before moving on to the second one, do you happen to know how to get around this issue ?

    On the suggestion of a more suitable attenuator, I have a commercial BNC 0-60db attenuator in 10db steps. Just saw it the other day when cleaning out a cabinet. Now to figure out where I put it !

    Bob

  • these pop up on eBay from time to time and aren't outrageous new


  • Yes, the two, three and four port versions are all useful.

    You can often obtain the 75 ohm versions a lot cheaper on auction sites, and they are still good.

    https://www.minicircuits.com/WebStore/dashboard.html?model=ZSC-2-2-75%2B

    https://www.minicircuits.com/WebStore/dashboard.html?model=ZSC-4-3%2B

    Regards,

    Martin

  • Hi Bob,

    Some splitters still work well below their defined lower frequency limit, a lot depends upon their construction.

    The loss increases and the port isolation degrades, but they are very often still usable.

    If you have a KiWi error like that, then personally I'd back up all of your settings to a PC and a KiWi image to an SD card, then rebuild the KiWi from scratch.

    I'm not an expert on this, but a search on the forum should provide suitable instructions and where new SD image files can be obtained from.

    Regards,

    Martin

  • Re: Git clone damaged!

    On the admin page, console tab, try using the "re-clone sources" button. I think that is present in v1.627

  • That fixed it John...thanks ! Updated to current version.


    Martin, found a good splitter on Ebay....thanks !


    It is working very well on MW now as I was aiming for. 207.135.223.250:8073


    Will start on the less sensitive one after the splitter arrives.


    Bob

  • Hi Bob,

    What antenna are you using at the moment ?

    There are quite noticeable intermodulation products from Medium Wave broadcast stations.

    I don't think it is occurring in the KiWi itself, but the antenna feeding it.

    Have you got any filters in line between the antenna and KiWi RF input?

    Regards,

    Martin.

  • Hi Martin,


    I suspected the IM by the every 10khz line on the waterfall. Plus I have heard audio mixed with 19.8khz.

    Reception at the guy anchors on my 240' tower is exceptional in the MW band. My long term goal been to successfully couple it into a Kiwi at the other end of an RG6 line. I am close. I took a W6LVP loop and changed the loop to 20' of RG59 and looped it through the guys near the anchor. The W6LVP preamp is where the overload is. I was going to reduce the length of the 'loop' but W6LVP took an interest and has designed and built a device which does all his standard preamp box does except it has no preamp....a passive device. My next step will be to try it and see how it works. If it doesn't, I'll go back to the present setup and reduce the 'loop' size.


    Bob

  • Hi Bob,

    Just a random thought...

    If you have room for more antennas, is there a reason you have chosen to use the broadband active loop ?

    Would a wire, passive antenna be an option ?

    Regards,

    Martin

  • Hi Martin,


    I have long used a several hundred foot wire with good success but have always been fascinated by the superb MW reception near the tower anchors. I have now gotten that reception into the SDR on MW but it is not good on the LW band where my wire antenna has regularly heard Morocco in the past. I reduced the size of the loop and improved the result. More to play with once the winter cold takes a break.


    Bob

  • Hi Bob,

    Are the tower anchors conductive steel or insulated kevlar type stuff ?

    Regards,

    Martin

  • Martin,

    Back from vacation and had a nice weather day yesterday to play. The anchors are steel. And my wire antenna is intact and working but if it fails, I will have to pay someone to climb 170' to repair it. If I can equal or exceed the wire performance with the 'loop coupler', I would like to relegate the wire to backup duty. Here is where I am now. W6LVP built me a passive device which is effectively a loop interface without the amplifier. I made a 20' 'loop' of RG59 and optimized it's pickup using a portable TEF6686 SDR and the results were superior to the wire antenna and noise was low. I then took the TEF6686 into the shack and had the same good result there. But....when I fed it to the Kiwi, the results were dismal. The TEF6686 had loads of NDB signals but the Kiwi had almost none. This applies to both Kiwi's on this antenna but the massive signal from the 800' of wire antenna forces it's way through and works better but not spectacularly. My guess is that despite the coaxial and terminal strip appearing to be in parallel, the area of the board where I tore off the coaxial connector may have completed some part of the circuit. I ordered a Kiwi2 some time ago and as soon as it arrives, I will have the answer to that question. Meanwhile, I have a great signal from the VLF band through MW and likely through HF but just can't effectively get it into a Kiwi. As always, any thoughts welcomed. This is the Kiwi fed by the passive loop coupler at the end of 125' of RG6.....same feed that performs exceptionally well on the TEF6686.


    Bob


  • edited February 4

    Hi Bob,

    OK all good.

    I was thinking that you could perhaps make a coupler, to fit over the steel guy rope.

    Something like a big Fair-Rite split ferrite core.

    https://fair-rite.com/product/round-cable-snap-its-431177081/

    https://fair-rite.com/product/round-cable-snap-its-2631181381/

    Pass the wire rope though the core and then add a secondary winding, with whatever number of turns gives the best results with your receiver.

    This is the same principle as used in the Shared Apex Loop receive antenna.

    https://dxsupply.com/produktfiler/AS-SAL-Manual_R1.10.pdf

    You may need to add an attenuator to obtain the best tradeoff between the Signal to Noise performance and Dynamic Range.

    Updated suggestion,

    If you fitted a coupler to each guy wire, you may be able to make use of its directional properties, just like the SAL loops.

    Regards,

    Martin

Sign In or Register to comment.