Good common mode choke for 100 khz to 3 MHz?
I need a good common mode choke for this band. I made the Dipole's 1:1 balun from an FT240-43 or 61 toroid (I didn't remember, I bought it 10+ years ago) but I need an additional one, because the coax line acting like an antenna especially on the MW band.
Dipole works like a charm above 3 mhz with (almost) no noise. (2x27m on 5m high poles)
An ft140-77 with 19 turns of RG-174 can be a good choice? I can buy these materials locally, ordering from abroad is too expensive because our currency (HUF) is weaker and weaker. I need this to for my Dipole on the Kiwi and for the MiniWhip at my home (My kiwi is at the country house). (I'll explain the MiniWhip problem in an another question...)
The feed line is a professional Tri-Shield 75 ohm coax and 40+m long in both cases.
Try a MiniCircuits T1-1 located right at the Kiwi. It has about 1 pf interwinding capacitance and a low end that will almost make it to 100 kHz. The dipole (what size?) no doubt has a great deal of mismatch loss at LF but ITU noise floor rises faster toward the low end than the 20 dB/decade of mismatch drops so noise floor should be still OK.
If you have significant feedline length, CM current through the Kiwi into that can generate internal ground-plane IZ drop in the Kiwi no matter how good the balun at the antenna might be. The feedline without any antenna may look like a current sink.
I am about to re-deploy my kiwiSDR after moving to a new location and have ordered the circuit board from OSH park using the layout files posted on another thread here. Digikey has the minicircuits TC1-1 which is essentially a T1-1 with center tap on one of the windings. I'm thinking about the option of using the center tap as an earthing point at the kiwi end of the feedline. I'll arrange to have this as an option so I can try with and without earthing. In the case of a balanced twisted pair feedline I imagine it would be best to put the winding with the center tap on the feedline side so as to let the Kiwi ground float independently from the common mode earth reference for the antenna feedline. The loop antenna I will use has another broadband isolation transformer at the antenna. My home brew analog power supply has an earth isolated output and with no antenna the KiwiSDR's noise level across HF is down around -130dbm except in the DC-500khz area where it rises to abt -118dbm. I'm not sure how this compared to the experience of others. Feel free to comment.
It is a little complicated trying to envision all the potential current paths for common mode noise injection into the reciever's ground considering the network, power, antenna and GPS connections and stray capacitiance paths it is complicated. Contrary to the usual dogma of grounding everything, I'm trying to go the opposite way and keep things isolated as much as possible. Comments appreciated...
Certainly try it with and without the center tap connected to something. However do recognize that the amount of balance needed to push CM current contributions below a Kiwi's noise floor may take a lot more balance than typical hardware can provide. I tend to consider hardware balance/symmetry to perhaps improve things by a 20 or maybe 30 dB. Component tolerances and imperfections, deviations from the ideal generally keep balanced mixers, transformers and other stuctures from doing much more than this. Even/odd order suppression in balanced mixers and such generally only get one into this territory.
CM sensitivity in the Kiwi, and perhaps in many other amateur receivers, may be only 80 dB down. Against a noise floor of -157 dBm/Hz the amount of balance/symmetry/suppression needed in real situations can be 50-60 dB. It's not a simple problem and as you suggest, there are many mechanisms that can cause problems. I tried to identify some of these in Notes on Improving Receiver Noise Performance, but this isn't meant to be exhaustive.
Do try the experiment and let us hear how it goes.
no problem with the Kiwi's noise floor :) Just the coax picking up some noise ("they" still on the spectrum when I disconnect the dipole)
"coax picking up some noise" probably means that CM current into the coax is creating IZ drop across the Kiwi ground plane. Conversion from current through
the Kiwi 'ground' to equivalent differential input at the Kiwi preamp 50 ohm input is on the order of -80 dB above LF.
As little as .6 microamperes of common mode current becomes around 60 picoamperes and generates a signal at the noise floor in 1 Hz.
A T1-1 at the Kiwi SMA probably will reduce that a lot. since CM source impedance is probably on the order of a few hundred ohms.
Thanks for the encouragement. I find myself once again in a high noise environment perhaps not as bad as the suburban one I left but still disappointing considering I moved to a rural area. Previously I was able to battle the noise using a horizontal loop placed right on the ground. This antenna has low angle and high angle sensitivity but efficiency is very poor so trying to reduce feedline participation in the antenna becomes important. If well done it can be surprising how much help this arrangement can offer but it is easy to have the feedilne contributing a lot of noise if it is casually deployed. I found that isolation transformers are better than transmission line type CM chokes especially at the low end where my interests lie on 160m. I'll report back on what I find using the TC1-1 right on the SDR with whatever earthing scheme seems to work best.
Best regards...Joe ve7vxo
I find that I get the best common mode rejection using a flux coupled transformer, which provides the best results on the lower frequency bands <10MHz, and also add several ferrite sleeves over the coax to provide more effective common mode choking on the HF bands
Some experimentation is required to get the optimum results, but it doesn't harm things if you add too many ferrite sleeves.
Martin, thanks for the advice. I will definitely consider the series combo of isolation transformer and transmission line style CM chokes. On both ends of the line if necessary. My passive antennas have the isolation transformers already, active ones are based on the Wellgood amps.
Well I ordered a set of three boards from OSH park according to the gerber files left by N6GN on another thread (thank you for those) but now I have two problems. Glad I waited to order the connectors because this board has a footprint for SMA connectors which are not the usual width. I was going to put an SMA jack on one end and SMA plug on the other but I cannot find both parts in a 0.250" width to fit the board layout :(... Looks like I will have to use a jack on both ends and an SMA union to connect to the KiwiSDR unless someone here (N6GN maybe?) can point me to a source of SMA plug which fits the board.
The other issue is getting the T1-1 transformer. The TC1-1 that digikey has turns out to be way too small to fit the footprint on the board. Shipping from Mini-Circuits directly is prohibitive, and it seems to be unobtanium at other outlets, so I decided to just use a small binocular core or a pair of beads to make a DIY transformer. It looks like the minicircuits transformer is twisted wire but the ferrite material is an unknown. I was considering a -73 or -77 mix but any advice from any HB pioneers here would be welcome. What about -31 material if a suitable size core can be found? BTW I am not into NDB DXing but am interested in 160m and generally lower HF.
BR all ve7qcx
I try and avoid Mini-Circuits if at all possible, because you have to jump through hoops to order stuff, and it takes forever to arrive.
Coilcraft is pretty good for transformers and inductors and they deliver almost next day (in the UK).
The types of ferrite you can get in particular formats, usually defines what can actually be used.
73 mix twin hole beads commonly sold as BN73-202 & BN73-2402 are the most likely candidates, but Epcos N30 mix (close to 75) is pretty good if you can get hold of it.
If you can't get hold of twin hole binocular cores, then small (<50) 75, 77 or J cores are OK, but you ideally need to twist the winding pairs together, so that they are very closely coupled. Unfortunately this increases the inter-winding capacitance which degrades the HF common mode rejection.
Use as few a number of turns as possible, whilst still being able to achieve adequate low frequency performance. If you use the cores mentioned above, four turns for each winding (1:1) may be a good starting point.
Yes I have some BN-73-202 cores already, they are a little large for the PCB but I can make it work by bending the leads underneath. I was going to go with the conventional wisdom of primary Xl at least 10x the Zo of the line so two turns is enough for 160m if I was interested in going lower then 4 turns yeah, but this way I'll keep the stray capacitance down which will help the upper end. So you think it's better not to twist if using the binocular core?
Binocular cores have a much more enclosed magnetic path, which reduces flux leakage and improves coupling, so you shouldn't need to twist the wires.
If you are not interested in frequencies below 1.8MHz, then two turns may be OK, personally I'd not use less than three, but your mileage may vary. I don't think there would be much difference in the stray capacitance between using two or three turns.
Well it has been a long time since I posted on this thread but I finally got round to completing this isolation transformer. I used the layout files provided by N6GN. The minicircuits transformer was unobtanium so I made my own. I used a TDK Epcos B62152P0007X030 ferrite core. This is N30 material with a Ui of 4300 and an Al value of 2400 nH/turnsquared. This is available at digikey under their part number 495-76793-ND for $1.96 cdn. I had some kynar AWG 28 wire wrap wire and could only fit two turns per winding which resulted in about 30uH primary inductance. This turned out to make a fine transformer for HF but I would use proper magnet wire and 3 or 4 turns per side if I was interested in gowing below HF. The measurment over an HF bandwidth and a photo:
N30 material is pretty good for this application, probably slightly better than 73 mix.
I've used similar N30 cores in broadband loop antenna amplifier circuits, and they performed very well.
Are you referring to the lossless feedback amp I saw breadboarded on your website?
The same core with four turns on either side. Definitely an improvement on the low end at slight expense to the upper end of HF. I edited the previous post to add a zoomed in scan of the low end of the two turn transformer for comparison.
Below, a full HF scan and a close-up scan of the low end of the four turn variant.
Just for the sake of completeness and in case someone looking at this thread for an isolation transformer for a higher frequency range I include two more much broader sweeps. First the two turn and then the four turn transformer. Both seem to have an anomaly around 50Mhz...I suspect maybe an artifact of the measurement setup or instrument.
Yes I've been successfully using N30 cores in lossless feedback (Norton) amplifiers, mainly in Wellbrook clones.