ad: k1jek

QRM Eliminator investigations - the phase shifter .....

Discussion in 'Radio Circuits, Repair & Performance' started by G1OJS, Jun 23, 2019.

ad: L-HROutlet
ad: l-rl
ad: Left-3
ad: Subscribe
ad: Left-2
ad: abrind-2
ad: L-MFJ
  1. G1OJS

    G1OJS Premium Subscriber QRZ Page

    I've been obsessing recently over the noise level on 80m (well who hasn't at some point?) and looking at dedicated receiving antennas, reducing RFI from my own house, and QRM eliminator circuits.

    QRM eliminators are simple in concept; take a signal from a second antenna, provide adjustment of the signal's amplitude and phase, and add it to the signal from the main antenna. This effectively steers a spatial null in the main antenna pattern, and if you're lucky this null will cover a significant source of QRM. The tricky bit is the phase shifter circuit that ideally causes no amplitude changes, works across all frequencies of interest, and is usable (smooth shift from 0 to 360 degrees).

    One phase shifter circuit that caught my eye was the circuit below, mainly due to its relative simplicity and use of familiar components. This is just the phase shifter - there's obviously variable gain amplification and summing with the main antenna to do too. I wondered immediately why VR1 and VR2 are independent controls; how does this bit of circuit work I wondered? Looking at the symmetry of it, why wouldn't anyone simply gang the two so that as as the tap on VR1 goes up, the tap on VR2 goes down?


    So, I did a bit of digging and circuit modelling and found that (unless I've done something wrong):
    • Ganging VR1 and VR2 indeed doesn't lose anything out of the available phase shifts, and hence should be done as twiddling two knobs (Phase & Amplitude) to achieve a null is far easier than twiddling three (2 x Phase + 1 x Amplitude).
    • The circuit can produce phase shifts between 0 and 180 only (again pretty obvious from looking at the circuit) so an additional 180 shift (e.g. DPDT switch) could very usefully be added between VR1 and T1.
    • Ganged or not, the phase shifter indeed causes amplitude changes as many have mentioned.
    • A bit surprised but only because I've not heard anyone mention it; the circuit is frequency dependent (obvious really as it's an R-C network) & will give a much reduced range of phase shifts as frequency increases. As frequency decreases, the phase shifts concentrate to a rapid shift at the middle of the adjustment range. So a multi-band version would really need a band switch and carefully selected pairs of capacitors for each band.
    Anyway, here's what I think would be an improved single-band version, if you want to try it (and I might try it myself ...)


    By the way, also wondering whether it would be possible to produce an absolute knockout QRM eliminator by working with DSP in I&Q? All things are possible there? Could even be worked into the architecture of a modern DSP rig - if done at IF, frequency dependency is less of an issue and it just needs an additional receive chain......
    Last edited: Jun 23, 2019
  2. G1OJS

    G1OJS Premium Subscriber QRZ Page

    Argh! I *did* make errors!

    I'm still thinking! Here is the best I have so far for the way the amplitude and phase varies at the output of the original circuit above as VR1 and VR2 is changed from 0% to 100% (bold numbers). So I was wrong about the ganging idea, and about the limitation to 0-180 phase, I think!


    Wish we could delete posts here.............
    Last edited: Jun 23, 2019
    KA0HCP likes this.
  3. KA0HCP

    KA0HCP XML Subscriber QRZ Page

    No deletions necessary. There is something to be learned from the process! Your post is the first I recall of anyone attempting to tackle a phase cancellation circuit. Thanks, Bill.
  4. G1OJS

    G1OJS Premium Subscriber QRZ Page

    Thanks - that's genuine encouragement. I've been known to recall emails to correct typos in the past so you can see how I like to get things right :)

    I'll post my next set of thoughts soon.
  5. KA0HCP

    KA0HCP XML Subscriber QRZ Page

    Tom W8JI designed the MFJ-Noise Eliminators -1025/1026. He comments on his web page about them having varying response at different frequencies, and mentions that the design could be improved for around $1,000. A few years later he designed the DXE NCC-1 and NCC-2 Noise Eliminators. The NCC-2 ($795) uses swappable filters modules for each band.

    W8JI, MFJ-1025/1026

    Last edited: Jun 23, 2019
    G1OJS likes this.
  6. G3YRO

    G3YRO Ham Member QRZ Page

    The problem with ANY of these phasing type "Noise Eliminators" is that are GREAT at nulling out ONE source of noise. (like a bad switch mode psu in a neighbour's house) . . . I have found I can get a 40dB null with my MFJ 1025 (I bought mine very cheap secondhand, but modified it to a 1026, as you really need the preamp for the noise antenna).

    However, what they CAN'T do is remove multiple sources of noise.

    The biggest HF noise problem here in Britain currently comes from the fact the most people still receive Broadband Internet via overhead copper phone cables. This creates a terrible noise level, which is really bad on 160 & 80m, and gradually dies off on the higher bands. (if you have overhead cables in your street, it can be S9 +20dB. In my case, the nearest overhead cables are 500 yards away, but that still causes an S6 noise level.)

    Unfortunately, because the noise is radiated from multiple cables, and from along all of them, a Phasing type device won't have any effect trying to cancel it out.

    Roger G3YRO
    G1OJS likes this.
  7. G1OJS

    G1OJS Premium Subscriber QRZ Page

    Thanks Roger, too true. I ordered a goniometer off ebay last week with the intention of making a "two knob" QRM canceller (amplitude and phase) thinking it would be easier to set, but in the meantime I've confirmed exactly what you said. I was too easily intrigued by the videos on the web that seem to show broadband noise being reduced! So now I'm playing with separate receive antennas (as far from houses as I can get, which isn't far unfortunately). Had fun building a loop preamp and PTT-controlled antenna switch, but so far it seems to offer a slight improvement in signal to noise on maybe 10% of signals on 80m. My noise level is about S5-S7 - so I might end up just living with it!
  8. KC3SWL

    KC3SWL XML Subscriber QRZ Page

    The Timewave/JPS ANC-4 is what I have. It works rather well on my QRM. A s-9 can be knocked down to an s-6 or better which, if you look at a signal meter, might not seem that much . However the reduction helps to unmask those weaker stations that wouldn't otherwise be heard. My average noise reduction is about 3 s units across the bands.
    Sometimes the reduction is greater. It all depends on how the antenna is set up .The very best I got was reducing from S-8 to S-3 one time. That was a good day.
  9. AA3EE

    AA3EE Ham Member QRZ Page

    Flat phase and gain are extremely important in digital TV, at least over a 6 MHz channel (which is only a 1% bandwidth at UHF). Our circuits used PIN diodes since they were frequency independent. Control was a 10 turn pot. I wish I would have asked the engineer how it worked when I was there.
  10. WL7PM

    WL7PM Ham Member QRZ Page

    . I have obtained the ANC-4 as well, and can observe the same " S9 noise reduces to S6" which is exactly the insertion loss for the device... It does, indeed lower the S meter reading.... both noise AND signal.
    I have tried long, medium and short single wires, and a 50 foot dipole ' noise sense antenna' but have not actually improved my signal to noise level as of yet.
    Not giving up on the device, but would sure like to hear more input from users.

Share This Page