1:1 RF Transformer Design SWR problems

[G0HZU_JMR]

Originally Posted by merlinmaxwell

Stacking two or more cores is one way to increase the al, without having much more wire. I would still go for the current balance method.

Hi Merlin, Like you I'm still not comfy with this isolation transformer approach. I started getting really scared at post #47 once I got an inkling of the system setup and so I stuck my disclaimer in about house vs outdoor ground effects. For me, that is the elephant in the room here. I don't have the practical experience in this respect because I'm a lab person.

I could go on to describe how/why the thru line testing with the analyser is flawed and how it gives a false and fairly meaningless result here. But the harder I try to explain stuff the more I realise it is just a waste of my time and I'm wasting everybody else's time filling up the thread with simulations and theory. I still don't fully know what the system looks like and I'm worried about ground interaction effects once this thing is put in line. I'm going to repeat what I said earlier in the thread that I don't this isolation transformer approach is wise when used as Chris appears to want to use it.

[Radio Wrangler]

I think having an isolating transformer so that things touchable outside are connected to outside Earth, and that things touchable inside are connected to inside earth sounds reasonable to me. The isolation has to be good enough to withstand 240v RMS in case there is a major fault in armouring continuity to the substation, without dangerous current flowing.

My remaining concern is the automatic ATU on the outside side of things. How is it powered? because this too will require to be connected to outside earth, not inside earth.

David

[ChristianFletcher]

There is a possible issue here with what we would call touch potential. There will be a difference between the ground earth and the building earth. This would obviously be serious if the neutral supply to the house ruptures.

So it would not be a good idea to touch primary and secondary shield connections at the same time. If I ever implement this I will build it in a plastic box and put a warning on it for myself. remember this equipment won't be operated by Jo blogs but by a person who is aware of the risk. better shut this forum down before someone plugs in a vintage radio or worse a 1960 TV, they never have issues do they !

However this is no worse then somebody operating mains operated equipment in the garden using a extension cable. ( bad idea with PME) I assume that there will be some regulation for this I'm only aware or the industrial limits on the railways and not domestic arrangement. It would however be of interest to try and measure the touch potential.

The supply to my remote ATU is derived from a 12 volt power supply. The power supply is industrial grade fully isolated switch mode. I did check the 0 volts connection at the ATU and it does not appear to be connected to the RF side.

Regarding the testing I have done and ground interaction the only issue we have if it actually works and passes rf without to much loss when installed. If it doesn't work I won't use it I really dont see the big problem here. There certainly no safety concerns if the material chosen can provide the isolation. I will have to check the cable spec etc. Given that I have 15kv hipot tester on the shelf I can do this.

Hope this email does not sound terse my concern is how to design a transformer not its electrical safety.

Chris

[G0HZU_JMR]

Go back and read my post #47 disclaimer again. I was referring to ground effects at RF and not electrical safety. Your preferred solution is to put this thing inline (in your house or just outside it?) and then run a coaxial feeder to an ATU.

If you do this then you are creating extra system problems at RF to try and fix your mains isolation requirements.

You are using a balun to connect unbalanced to unbalanced (house gnd to outside gnd). To me that sounds like a recipe for RFI, power suckouts, poor VSWR etc. I don't know how your house is arranged wrt the antenna location or how well 'connected' your house ground is wrt the ground outside at RF across the HF bands but there is the very real potential for some serious RF problems here.
You might be able to tame it all using ferrite clamps but I'm not sure I'd want to go down this path.

[Radio Wrangler]

I don't see that there has to be a problem, Jeremy. There is an IC7700 running by my right elbow right now and 20m is open with lots of JOTA traffic. My radio indoors is galvanically isolated from the antenna outside by transformer action, but in this case it's a resonated transformer in the form of a Z-match ATU. There is another wideband transformer in the radio providing isolation from the FET drains on 50v DC to the antenna borts which are at chassis ground.

Christian doesn't need a balun, there seems to be nothing balanced in his system, the word unun seems to have arisen for unbalanced to unbalanced transformers. All he needs is the HF equivalent of a braid-breaker.

VSWR only needs to be good enough.

David

All he needs is the HF equivalent of a braid-breaker.

Last time I a going to say this, use a current unun and isolating caps.

[ChristianFletcher]

Thanks everyone I really do appreciate your comments and time people have put in with the contributions. But my thread is about how to build the best type of wide band RF transformer with DC isolation. I think I'm making progress at this. I will keep making test and I hope people will still contribute and offer advice and correction when I’m in error.

I think I used the word UNUN. I assumed that given the input and output connections are unbalanced this would be the correct description for the transformer configuration. Perhaps my original description was incorrect if this is normally employed to prevent RF common mode voltages.

Jeremy I am sure your correct that my test are invalid. But I think they did tell me something about the transformer and the core material. I think that your comments and information contributed to what I think is a success. I just did the test measuring the voltage with two 25 Ohm resistors. I don't get any measurable imbalance now until above 30Mhz and its still tiny. The first Type 2 core failed at 14 Mhz. You models and contributions have really helped!

Merlin - I don't believe that I need a current balun, I will have to concede that it won’t do any harm but I am not aware I have a massive common mode issue although again I concede it is quite likely given the current ATU arrangement.
I am aware of how to use a capacitor RF shunt, I mention this as an alternative in my first post, but at the moment I am working on a solution that uses an dc isolation transformer. Any suggestion on how to improve this arrangement are gratefully received as you obviously have more experience than myself.

It would cost a lot of money and disruption to install a traditional earth system in my house. I don’t actually know if this transformer will work. If it does work its likely to be a compromise. I don’t see I’m doing anything wrong by exploring this and doing operation tests.

I am going to try a few power test as I’m not sure about the handling capabilities of RG174 and need to look up the spec. If these go ok Im going to put a few CQs and let’s see if you can hear me or if I can hear anyone else.

Jeremy I am not concerned about the mains voltage causing my transformer to break down, but could your model predict voltage levels within transformer. If a RF breakdown was to occur first this could cause a secondary DC breakdown. I’m certainly aware some ATU can see very high voltages etc. My VNA produces a S parameter file is this any use to you for the model ?

Best Regards Chris M0KLF PS - It would be great to have a chat about this on the bands

[G0HZU_JMR]

OK, I'll have another go at expressing my concerns wrt RF ground effects.

I think the whole point of coaxial cable is to try and keep the RF currents inside the coax and transport RF energy from the house to/from the antenna without leaking energy or absorbing external (interference) RF energy along the cable run via common mode effects. Current/choke baluns can help here on a typical ham installation as we all know. By putting this isolation transformer inline part way up the feeder this property will surely get destroyed because the integrity of the coax has been breached here.

The side effects of this could be numerous and will be layout/location sensitive but at the very least I'd expect the feeder to become leaky to RFI. This could mean that local house interference from PCs and battery chargers etc will be able to 'get in' and some of the RF from the transmitter will leak out here too. There are worse scenarios than this if a decent ground loop exists. In theory at least it could cause a huge inband suckout causing strange VSWR effects and lots of RF radiation inside the house. But I'm not an expert in house ground vs external ground effects so I can only predict the possibility of this happening and it will obviously be cable length, frequency and site specific.

Ironically, I suspect that if the isolation transformer was fitted somewhere in the coax in/near the house then it would ideally require coax based ferrite choke(s) in line to fix the RF problems the isolation transformer causes

[ChristianFletcher]

Hi Jeremy

Well I cannot argue with any of your concerns they are all valid. However I think all rf installations are a compromise. My existing arrangement with my un-earthed ATU and its ground isolated counterpoise is certainly a compromise. This is not a life and death situation the worse that is going to happen is I get some hot RF in the shack and the TV wobbles. In which case this will have been a failure. I may not like the destination but the journey is still interesting.

I have a Rf clamp somewhere, its certainly worth trying making some common mode measurements with and without the transformer in circuit.

I think you have an idea from the previous post regarding the equipment I have to hand, can you recommend anything I can do to investigate your concerns. And suggestions how to mitigate your concerns. Also any other performance test that I can do within limits of the test equipment I have.

Thanks Chris

[ChristianFletcher]

Well it appears to work fine based on a very quick and dirty tests run at 50 watts I did not get any reflected power until 28Mhz and about half of it was probably due to the ATU not matching the end fed wire very well on this band, it never does. I have to say the the ATU I have is really good (but so is a 50 ohm resistor)

I did have the power turned down and was only running 50 watts. I transmitted on FM into a dummy load for 2 minutes and there was no heating of the transformer.

The bands where in very poor condition. I did a comparative listening with and without the transformer and I could not detect any difference. I will perhaps repeat this using one of my SDR receiver as this gives an instant signal to noise read out. However given the natural fluctuation on the bands I think this is hard to do. Any Ideas ?

I did a very quick test to look for common mode issues by using the VNA to do a live scan while running my hand up and down the feeder cables and it did not have any effect.

Given that this works certainly no worse than what I have existing its worth further experimentation. The ATU earth is connected via a very dubious floating counterpoise. Perhaps the dire predictions will become true when I actually earth this end to true ground.

Video link showing operation


https://youtu.be/hur1JSQB5bk

[ChristianFletcher]

Had quite a few QSOs with mainly G station and also some foreign competition stations this afternoon on 20 and 40 and received a few real 59s and a 59+20 Sheffield to Manchester. I tried with the transformer in and out of circuit and I cannot tell any difference and it made no difference to the radio reports received. I could claim my noise level was lower than normal, but the bands were in very poor shape this afternoon so that may be stretching it little. But I think its working and I'm not detecting in SWR or RFI issues.

I can only assume that the ATU is doing a good job maybe keeping the impedance under control. Perhaps I would have problem without the remote ATU as my SWR is normally 1:1 on up to 28Mhz and then just above. On the video you can see the needle kicks as the ATU tunes. I Normally turn the power down when tuning as it only needs 5 watts to fire up the tuner.

Chris

[Radio Wrangler]

It sounds like a good result.

However, the auto atu probably isn't helping or hindering. Its reflectometer will be on its input, so it will be adjusting itself to create a 50 Ohm load at its input. If the cable and transformer is affecting anything, the tuner won't do anything differently.

I'll tell you on Saturday about the biggest mistake I ever made... an atu.

David

[ChristianFletcher]

OK David

I just thought if the antenna was not being so well matched by the ATU I may have been more likely to have RFI issues.

Looking forward to meeting you Saturday- The pie is on me.

Regards Chris

[G0HZU_JMR]

Originally Posted by ChristianFletcher

I think you have an idea from the previous post regarding the equipment I have to hand, can you recommend anything I can do to investigate your concerns. And suggestions how to mitigate your concerns. Also any other performance test that I can do within limits of the test equipment I have.

I think your existing VSWR and thru tests already show a clue that there is a system issue here wrt grounding /resonance effects. I suspect that nobody else has studied what is happening in your plots but they tell a story to me. If you look at your VSWR plot (ignore the Rs scale because I think this should stay fairly flat at 50R) you can see that the VSWR rises very gracefully up to 50MHz. It doesn't get worse than 3:1. I'm going to assume that this is a repeat of your artificial 'floating 50R load' test.

But look at your thru plot given below. There is a huge dip in thru power at 43MHz. To me, that is a clue that energy is leaving the system at that frequency in your thru plot. Some of it may be burning up in the ferrite but some of it will be radiating away via the cables. A conventional linear RF simulator won't be able to model this energy loss so it's easy to overlook what is happening here. It's able to leave the system because you have broken the screening integrity of the coaxial cable with your transformer and you have arranged the test connections such that the whole thing acts as an antenna at 43MHz. I think this suck out mechanism will be a function of the transformer and the cables.

I'm not trying to get you to stop what you are doing but I am advising you that your transformer will degrade the screening capability of your transmission line and it will form a (resonant?) antenna with your cables and the RFI symptoms it produces will be frequency, cable, site, and transformer specific. Maybe it won't ever cause a problem that is noticeable to you on Rx or Tx but I'm trying to help you understand your system better.

Look back through the thread and I'm the person telling you exactly why your initial plots were confusing, why your test setups weren't consistent, why your first transformer was misbehaving... I told you how to model and predict its performance as a transmission line. I told you to shorten the winding length and use ferrite and I told you to use a transformer with a tline that has lower Zo than the 80R twinline you were using. I also advised you to experiment with lower Zo by putting tlines in parallel. This was covered in my first two posts I also told you your long connection wires in your test setup were making the (already poor) VSWR even worse and I told you why. The thread was filling up with vague guesswork and lacked any substance so I joined to try to help you better.

I've given you a lot of detailed technical advice about the physics of the transformer itself and your test setups and I'm offering advice to you again with respect to the potential for RFI in your system. Maybe it won't ever cause you a problem but I can try and help you understand the radiation mechanism(s) that arise when you put this system together? Like I said back in post #46, I'm not an expert in house vs outside ground effects and I am definitely NOT an antenna guru but in theory at least there is the potential for significant RFI at the resonant frequency of your transformer/feeder setup and also at other frequencies too.

Do you have any ferrite clamps to hand?

[ChristianFletcher]

Thanks Jeremy

As I have said in my previous post all your comments have been really useful and all gratefully received. It is those comment that have enabled me to build something that works. Im not saying your concerns are not well founded I’m sure given your obvious experience.

I don’t think the modelling route is the way to go for me at the moment. I do a lot of electronics modelling using various programs and some of my models don’t work correctly all of the time. I can understand why they don’t work because I have experience of practically designing and building analogue and digital electronics and with constructing the models.

To use model effectively you need to have really high degree of understanding on the underlying principles. I don’t have this at the moment for RF design so my models would likely be even more inaccurate than me attempts to make practically measurements.

I would however like to see if the concerns you raise are a problem for my current situation as your model has predicted. So any idea of how I can make practical measurements and test these against the model predictions would be really interesting for me to do

I think however I have taken this thread as far as need to go. I have a better understanding or winding styles, Core materials, and wire lenghts etc. I have a working prototype to work on, explore and improve. So I am happy.

-- Jeremy did you mean clamps for common mode / RFI or measurement type clamps? Thanks


Thanks for everyone’s contributions.

[G0HZU_JMR]

I can offer advice here, but as I said earlier I'm not really an antenna/grounding guru and I don't have any practical experience when dealing with transformers like this inline. Putting something like this between two unbalanced cables does seem risky from an RFI point of view. I'm much better at component modelling and circuit or RF module design. Most of my experience is up in the GHz region so stuff way down at HF is interesting but I simply don't have much practical experience wrt antenna installation/analysis on the SW bands.

But the first thing I would do with your thru plot setup would be to confirm that energy is radiating away via the feeder braid. I think you could do this with a ferrite clamp on one or both sides of the transformer and a monitoring receiver nearby to see if the suckout shrinks and the leaked energy goes down at the test receiver. The easiest way to do this is with a ferrite clamp as per the image below. The suckout might just vanish from the analyser anyway meaning you don't need the remote receiver test.

Then maybe try some tests that are a bit more interesting. Remove the clamps to make it suckout strongly again and try increasing the length of both of your test coax cables the same amount, keeping the transformer in the middle like a dipole. Does the suckout creep down in frequency? Does it get deeper or shallower? Maybe try 2 cables a metre long each. I suspect that you have a fairly efficient antenna at 43MHz in your existing thru test jig. I think the resonant frequency is a function of the transformer and the cable dimensions.So I think the suckout will migrate down into the HF bands with longer cables. I think the efficiency will diminish as the cables get longer and also the ground loop path will be less defined in a real system. So this will help damp any resonance. I simply don't have the experience to tell if this resonant mechanism will actually cause you much grief in reality on transmit. It's more likely that the leakiness of the cable will let any strong local RFI within your house into your receiver via the cable. But I guess if the interference isn't there to start with, it doesn't matter much anyway. Or you could stick some ferrite clamps on the cables to see if this helps.

[ChristianFletcher]

Hi Jeremy that's brilliant

Thanks for that. You have confused me by the test method mentioning the "monitoring receiver" is this the spectrum analyses input.

My original plot was done with my spectrum analyses and the transformer was installed between the tracking generator and the analyses input. I just want to make sure I have understood you correctly as I know you said there was an issue with the common ground on the analyses input outputs etc.

Or are you perhaps suggesting that i transmit at very low power and then listen on a separate receiver whilst installing the clamps etc and look if i detect any change in the received signal ?

Regarding the dip in the original plot are you perhaps suggesting that the dip where the power drops off that this energy is going somewhere i.e out the feeder. I think I had not really understood the term suckout etc. I just assumed that my transformer was just not passing the RF energy more like a notch filter perhaps.

Just had another look at your sketch and another think . Perhaps I could connect the transformer as shown in your sketch and then use the monitoring receiver to listen for leaks as i frequency sweep - perhaps at resonance i will get a peak if the rf is radiating from the shield etc.


Best Regards Chris

[G0HZU_JMR]

Sorry for any confusion. Probably the easiest thing to do is replicate your existing thru test as before and get the same response as before and then try grabbing the RF cable with some of your fingers to use your fingers as ferrite clamps very close to the transformer. Just doing this might be enough to make the dip in energy go away on your analyser response No need for a monitoring receiver or any fancy ferrite clamps?

By suck out I really just mean energy that has left the system and radiated away. The notch in the response at 43MHz looks like a resonating/radiating suck out to me. I'm not back at work until next week but I can try and get some proper RG174 cable and a big ferrite core like yours and have a play. But I'm unlikely to find a core that is the same as yours. I think the ferrite and the number of turns (and cable length) will play a part here in determining the frequency and sharpness of any resonances.

[ChristianFletcher]

Thanks Jeremy

Understood regarding the test thanks. I will try and post a youtube video link so you can see how we go.

I think the material is likely to be type 43.

I was also looking at a bigger core to use the in the future. Looking at the datasheet it list RG174 as having a maximum working voltage of 300 volt. I assume that RG58 will be higher still but I don’t think I could wind RG58 on the current core without distorting the cable. I think the next core size up is listed as $78 but my pockets are not that deep at the moment.

I should probably also just try a bifilar wound arrangement using two normal wires. I have some here rated at over a 1000 volts. I am guessing using the RG174 is a quite tightly coupled arrangement and one of the reasons it appears to work well. The other advantage of single wires is I could try winding these from different end of the transformer so the connector arrangement would be better. It would be nice to have a input connector on one side and output connector on the opposite end of the enclosure.

I think I will stop buying these cores at radio rallies as without knowing the actually material used and specification you can waste a lot of time and effort.


Regards Chris

[Mike. Watterson]

Or use a plastic box with a pair of 2nF or 10nF 2kV ceramic caps between two connectors as you only want DC isolation with 1:1. Four foot earth spike via screw terminal to aerial side coax connector, with maybe arrestor and spark gap too on the aerial side.

I'd only use a transformer if I needed a balun (probably use current mode choke) or a ratio other than 1:1. If I wanted DC/Mains isolation and either of those, I'd add the capacitors too.
"RG174 as having a maximum working voltage of 300 volt." So nearly 600V with two lengths twisted together on a core.