A Bias-T and active antenna performance

KA7UKA7U
edited February 2018 in Installation, Antennas and Interference
Working with the Home Brew Shielded Magnetic Loop has led me to want a Bias T that will work from 15KHz to 32MHz. I have the input side from another commercial loop antenna and just as a guess I placed a 1 mH inductor from the coax input to the +vdc on the Hiletgo LNA board and that works to power the amp and push signal down the coax. The circuit from the LNA is a 10pF capacitor to the 1mH choke and the coax center conductor. The self resonance of the 1mH choke is stated at 1MHz and the calculated resonance of the capacitor/inductor circuit is 1591.549KHz.  

I wonder what affect this will have on the antenna. By observation I do not see the affect on the waterfall, but I have not cross checked with providing power from a VDC source not carried by the feedline. 
Will a Bias-T notch frequencies at the resonant points?
Will feed back from the Bias-T cause oscillations and harmonic generation through the LNA?

I'm not seeing this, but just wonder what to consider when setting up a Bias-T to operate from 15KHz through 32MHz.
This Bias-T circuit is currently in service at http://ka7u.no-ip.org:8073
Ron
KA7U

Comments

  • G8JNJG8JNJ
    Hi Ron,

    Bias tee's are not as simple as they seem at first sight.

    The trick is to use a lossy ferrite so that you don't get a very high value of Q which could lead to unwanted resonances. The core also has to be big enough so that it doesn't saturate (and produce unwanted intermodulation products) at the required current draw.

    12 turns through a BN73-202 binocular core (Fair-Rite 2873000202) is a good starting point if you want to get down to just a few KHz.

    The DC blocking capacitor needs to have a low value of XC at a few KHz so 1uF is probably the largest value (XC @ 10KHz = 15 Ohms)  you can easily obtain in a non-polarised version. 

    For the DC power bias tee. Use two chokes made from 12 turns through a BN73-202 binocular core. The first is to help remove RF from the power line. Connect one end to the DC input and the other end to the second choke. Also connect this point to a shunt capacitor connected to ground 100uF and 0.1uF in parallel should be OK. The other end of the second choke connects to your signal line and the receiver is DC blocked with the 1uF cap.

    I'd also suggest adding a shunt connected polarity protection diode across the 100uF capacitor and fit a fuse or Polyfuse (resetable fuse) in series with the DC input supply.

    Something like the right hand side of this diagram. Ignore the less than optimum component values, as they were what I had to hand at the time I built it, but I now know better :-)


    Regards,

    Martin - G8JNJ









  • KA7UKA7U
    Martin,
    I've kept your Bias T directions for future use. For now I'm using the Bias T that George Smart provided with his Wellgood board, but if I build from scratch I'll fall back to your BN-73-202 cores. I do want to add voltage adjustment to the Bias T units and have ordered parts for the scheme in Figure 7 on this data sheet:

    I've been adjusting voltage on the AC line with a variac, but that isn't practical in all circumstances and the input voltage to these active antennas does make a difference in performance. Do you have observations or comments about the LM317T voltage regulator. Or other preferences?
    Ron
    KA7U
  • G8JNJG8JNJ
    Hi Ron,

    You raise an interesting point regarding the noise contribution from three terminal voltage regulators.

    This topic was covered in some detail in the March 2018 issue of RADCOM (p76-77), the member's magazine pubilshed by the Radio Society of Great Britain.

    The bottom line is that most three terminal regulators generate noise (on both the input and output pins), particularly at lower frequencies <1MHz and that larger values of decoupling capacitors need to be added from the 0v / ground to the input, output and adjustment pins to minimise this as much as possible.

    Normally in other circuits, the input and output noise wouldn't be too problematic, as there would be plenty of another decoupling capacitors scattered around the circuits that were being powered by the regulator. However, in this case, where it is being used in conjunction with a Bias Tee and it shares the RF signal path, the noise contribution becomes much more important.

    Based on the article I'd suggest adding 0.1uF's from the input and output pins directly to ground with as short a lead as possible in order to provide some HF / VHF decoupling with additional 2,200uF capacitors across them to kill the VLF noise. If you are using a variable voltage regulator, also add a 1uF from the adjustment pin to ground.

    Incidentally, LF noise can also be a problem when using some broadband MMIC amplifiers. These devices benefit from having low frequency decoupling from the input pin to ground. The assumption is that this helps to remove LF noise from the MMIC's internal biasing network, so it needs to be treated like an AF amplifier. The difficult part is figuring out how to do this without shunting the wanted RF signal to ground. If you are coupling directly into a low impedance loop it becomes a lot easier. All you have to do is shunt any existing input RF coupling capacitors (typically 10 or 100nF) with a much larger value in parallel. I've used 10uF polarised SMD capacitors for this purpose, but even larger values may be required in some extreme cases.

    SM5BSZ has some videos on the subject. Although he is concentrating on sideband noise it is also applicable to the noise floor at VLF frequencies too.







    Regards,

    Martin - G8JNJ





                                                                                                                                          


  • KA7UKA7U
    Martin,
    I placed Mr. Smart's Bias-T into service @ http://ka7u.no-ip.org:8074 today. This Bias-T is provided 3.09vdc with the LM317 voltage control circuit. See attached schematic. The LM317 is powered with a 14vdc linear Walwart with the DC input cable wrapped several turns through a local mode choke (43 type core). This LM317 circuit is hand wired on perf-board and the Bias-T and LM317 are enclosed in an aluminum box. I've tested it with both VHF and UHF transmissions close by and have not noticed any added noise through the antenna system to the KiwiSDR. The noise floor through this power source is the same as the noise floor using a linear power supply with the voltage set with a lab quality variac. So I'm happy with it and plan to make a couple more. George Smart provides a nice Bias-T board with the Wellgood boards. The Wellgood boards are working fine business too. So no complaints with Mr. Smart.

    I've had lots of fun with this antenna project. I'm still a fan of the Hiletgo LNA. :)
    Ron
    KA7U


    Attachments:
    https://forum.kiwisdr.com/uploads/Uploader/3e/d753c1c9c20322af7eec94c6f95757.png
  • G8JNJG8JNJ
    Hi Ron,

    Good stuff.

    I'm assuming that you made the PSU noise floor comparisons with the bias tee in circuit, but without the antenna connected ?

    Regards,

    Martin - G8JNJ


  • KA7UKA7U
    Martin,
    I did test without the antenna. Both power sources add noise below 25KHz. Both power supplies are reasonably clean above 25KHz. 
    Ron
    KA7U
  • G8JNJG8JNJ
    Hi Ron,

    OK - if that's a problem you need to use much bigger value capacitors :-)

    Regards,

    Martin - G8JNJ
    KA7U
  • KA7UKA7U
    edited March 2018
    I've attached a couple photos of the Bias-T builds. Now that these are made I did a google search for DIY printed circuit boards and find the kits are available at very low prices. Shipping times are 30-60 days though.
    http://aliexpress.com and search for DIY LM317 circuit boards.  Lots cheaper than buying parts from Mouser and easier to use a PCB than hand wiring. They sell nice cast aluminum boxes too.

    Ron
    KA7U


    Attachments:
    https://forum.kiwisdr.com/uploads/Uploader/37/566da1847b8b35a626f749d4145fa1.jpg
    https://forum.kiwisdr.com/uploads/Uploader/1a/178ab2100e34b425a38b5079dc3320.jpg
Sign In or Register to comment.