Re: LF: Re: HB9ASB...

[Andy Talbot <andy.g4jnt@gmail.com>]

I'm surprised so many seem to misunderstand how high impedance E-field probes work.  There is too much hearsay and false comments being peddled.   Stop thinking of the feeder as 'part' of the antenna as if it were all one entity and consider each stage.... And forget that thing called Ground completely.

 

It starts with a high input-impedance amplifier which must have two connections - ignore any common reference for now, it just has two input pins.  A voltage is imposed across these from any antenna with two ports,  like a short dipole.   This input voltage is buffered, amplified,  and sent to the two output pins from which it travels down a feeder - balanced, coax, or whatever -  and into your receiver.   So far we haven't made any connection between input and output, and they could (and ideally would) be independent and isolated from eachother

 

However, they're not.   One input connection is usually common with one output - usually the  0V DC supply pin and the reference  (and please note, I am not referring to this as ground; ground is taboo, a dirty word,  and will not be mentioned)    Which means the other hot-side input pin now has a voltage imposed on it with respect to the reference.  This will probably come from a probe antenna which is coupling to the E-field of a radiated signal.   Now, bear in mind any antenna must have two output ports, so where is the other side?   As the reference input pin is connected to the output reference pin, any other connection to this point will form the other 'half' of the high impedance short dipole antenna.

 

Now, the feeder dropping down from the amplifier / probe assembly (or across, or up and over, or buried in the soil [see, still didn't say the taboo word] ) is connected to the input reference pin by virtue of the amplifiers internally linking them, and must therefore form the other half of the dipole.   This is bad.   We have a hi-Z dipole, with one short element in the air where it should be, and the other element being one conductor of a probably long length of feeder with its end connected to we know-not-what.

 

So what do we now have?  A long assymetric dipole plus amplifier assembly, stretching from some arbitrary shack connection up into free space.   The voltage at the not-middle of this is the stuff that is amplified and fed to the receiver.    So, as the bulk of the dipole is close to the shack expect the majority of signal received to be locally generated noise.  

 

That is the case for a completely unscreened and un-earthed (still didn't use the word :-)  system.  Picking up any locally generated E-fields at high levels onto the dipole. 

 

Now place the receiver and the rest of the local world with all its noise and QRM generators in a screened room and make the feed coaxial through a bulkhead connector into teh screened room and see what happens.   All the QRM is contained within the screened room and cannot pass outside.    The antenna is now a strange sort-of dipole with one end in free space, and the other connected to a solid mass made from the outside of the screened room.   It can't pick up any QRM from inside the screened room, and all it will now receive are signals generating an E-field between the probe and the solid mass.  Perfect!

 

The shack and the rest of the world is not in a screened room, so there lies the problem.   However much you don't want it to be, with this setup the feeder will always form the majority of the antenna because there is nothing else there that can be the other half of the dipole.   And if the bottom end of the dipole intrudes into the rest of the world, it will pick up stuff it shouldn't

 

Now, lets control the antenna structure all on its own and pretend the feeder is not there at all.  A true dipole would be nice although impractical for now, so lets go back to the classic monopole which is one half of a dipole, and the other half formed by the reflection of this in a mirror, which is usually the surface of this planet called Earth.   Mount the amplifier assembly on a conductive pole with the bottom end firmly connected to the surface of the mirror.  Now, we have a dipole formed of the complete assembly (mast plus probe) with its reflection.  It is not fed in the middle , but assymetrically near the top where the short probe forms the other side from teh feed point.   The total length of the dipole is now twice the total height of supporting mast and probe.   As we're using a high impedance input amplifier, if it nad a truely infinite input Z and zero capacitance,  the actual feed point wouldn't matter, it would always get the same voltage imposed wherever the non-symmetrical spilt occurred.  As Z and C is finite it does make a difference so longer probes help with practical amplifiers - but this is digressing.

 

The antenna so far is ideal, and is no more prone to picking up local QRM than any other antenna would be. But now we have to connect this thing called a feeder, which introduces another arm to the dipole (a tripole now perhaps ) which will ultimately go close to bad places.   If we could isolate the feeder by inserting a high common mode impedance this would do the job, but it would have to be a near infinite common-mode impedance, so is definitely not on.    A  transformer coupling would help, but even that has high (relatively speaking) capacitance across its windings, so will still leak common mode rubbish, gettign worse at higher frequencies.  And there is still the DC power issue

 

One solution is to bring the feeder down INSIDE the pole which must be firmly connected to the reflecting surface of the mirror, maintain it buried under the mirror's surface for as long as possible before it sees the Badlands and hope that burying it will decouple any local QRM.   Or make the feeder non conductive like optical fibre.  But there is still the DC power issue.

 

Which shows where the problems lie, and where to start thinking about how to stop them.    Separate dedicated mast firmly connected to the mirror's surface.  Feeder fed down inside it and buried.   If coaxial feeder is used -  connect the braid at both ends to the mirror's surface and bring in to the receiver input port coaxially and screened.   DC likewise - send it up the feeder.

 

One solution, but its a difficult bit of hardware to build well, would be a true differential input high impedance amplifier with a proper equal length short dipole.   The differential input - if ideal - will inherently provide isolation between input and output ports but it does have to be a perfect differential input, balanced antenna and so on.   Someone did mention to me once about using two identical active antennas mounted end to end to form a a dipole, with the outputs from each combined in a 180 degree hybrid combiner.  That could prove viable, but don't know if he ever tried it.

 

Phew....!

 

Andy

www.g4jnt.com

 

 

 

 

 

2011/8/10 Stefan Schäfer <Stefan.Schaefer@iup.uni-heidelberg.de>

Hi Minto,

Am 10.08.2011 13:10, schrieb Minto Witteveen:

Hi Stefan, (et al)

Well I beg to differ.. :-)
What I think happens is this: The outside of the coax picks up electromagnetic radiation like any antenna (including QRM generated by fluorescent lamps and Alinco switching power supplies). This signal travels along the coax to the Miniwhip. (also in the direction of the receiver but that is not important here as the signal is on the outside of the coax).
Upon arrival at the miniwhip this signal on the outside of the coax has nowhere to go ­but to the _inside_ of the outer mantle of the coax – it ‘rounds the corner’ at the end of the coax so to speak.

I think the mechanism is that the unwanted signal on the screen causes a potential difference between gate and source of the first (J)FET. So this causes a current flow in the output stage and so a signal at the RX input.
A common mode choke between RX and the antenna ground should form a low pass filter for unwanted signals coming from the shack. Using a common mode choke without a local ground should have little effect, except the coax is some 100m long (between choke and probe) ;-)

Ah BTW regarding the discussion "the cable to the E field probe is the actual antenna": One could just try what happens if one disconnects the power supply. If the signal is still present then the cable is the antenna, if the signal is gone: The probe must be the antenna. Isn't it?! :-)

So how to avoid the QRM that is picked up by the coax to ‘travel back’ via the inside: for the miniwhip it is indeed best (as Roelof mentioned) to short these signals to earth _outside_ the house, preferably as close to the miniwhip as possible. Grounding there would to the trick, aided by a (large enough) common mode choke between the ground point and the house.  The QRM that is picked up in the house would be – after attenuation by the choke - directed into the ground and not up into the pole and the miniwhip.

Yes yes, totally agreed.

Whatever happens in the house would then be largely irrelevant. Adding a common mode choke close to the rig will do little extra. (it would only attenuate QRM getting from the shack’s earth system to the outside of the coax).

It would almost have the same effect (when ignoring the C between cable and ground along to the choke near the antenna ground) as placing the choke near the antenna ground, both are in series and increase the current reducing impedance, yes...

Any signals picked up by the vertical coax between the earthing point and the whip will add to the received signal, but at low frequencies it will not be much.
So far for theory. Now the proof of the pudding: DCF39 is now > S9+40 dB. My old trusty QRM generator (Alinco SMPS) generates S9+25 at 135.500. When I switch off the miniwhip (cut the power) DCF39 drops down to just above the noise floor. As expected.

Ah yes, that's what i meant above (should have read your mail completely before answering ;-) ). This is the proof that Mal cannot be right when saying "the coax is the actual antenna".

But the Alinco signal only drops down some 15 dB and remains the only signal that is audible. This is exactly what I would expect: the QRM travels along the outside of the coax to the miniwhip, ‘rounds the corner’ and comes back via the inside of the coax shield. Further proof that it indeed takes this route: if I disconnect the coax in the shack the Alinco smps signal disappears also (so it is not received via any other path).

Hm, i rather expect a galvanic coupling i.e. stray currenty on the supply cable of the RX. What happens if you run the RX on batteries? The same dependency?
There could be several reasons apply here...

Last year I already bought 3 meters of copper pipe to drive into the ground in the backyard. Bet never got around to finish the job…

Today it's nice WX here! And in NL?

The main reason the signal strength is much higher with the elevated miniwhip is (I think) caused by the fact that I am surrounded by other houses, gardens, trees etc. Not comparable with an open field…

Yes.

73, Stefan /DK7FC

Regards,
Minto pa3bca






--------------------------------------------------------------------------
Ceterum censeo Carthaginem delendam esse
-----Original Message----- From: Stefan Schäfer
Sent: Wednesday, August 10, 2011 00:03
To: rsgb_lf_group@blacksheep.org
Subject: Re: LF: Re: HB9ASB...

Hi Minto,

Am 09.08.2011 22:48, schrieb Minto Witteveen:

You are right w.r.t. the cable being (a significant) part of the working of the Miniwhip antenna. [...]

I don't think so. There should be no difference between a 5m and 10m
long cable. I think about a capacitive divider. The probe has about 3
pF, that's one plate of the C. The other one is the cable and metal
connected. Once if this part of the C has say >10 * 3 pF, the difference
between longer cables become smaller and smaller.

I think it is just the S/N that rises due to lower becoming noise and
higher signal levels. On a flat field without trees and houses, you have
excellent reception even with a 2m pole :-)

73, Stefan