|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
12th Oct 2016, 7:15 pm | #41 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
But things aren't so straightforward for the conventional transformer that is trying to mimic a perfect lumped model of a transformer. In theory the impedance shouldn't change for a perfect lumped transformer but because the real versions we are winding here are acting as transmission lines the impedance at the output could easily have a worse VSWR once the transformer is put in line. I think the transformer will behave best if the windings have a very low Zo. Then it will be better at replicating the behaviour of an ideal lumped (isolation?) transformer. We can already see how badly Chris' T200-2 and my T124-2 'isolation' transformer misbehaves when given a 50R load. The VSWR really is quite poor by 25MHz.
__________________
Regards, Jeremy G0HZU |
|
12th Oct 2016, 8:25 pm | #42 |
Guest
Posts: n/a
|
Re: 1:1 RF Transformer Design SWR problems
Bloomin' awful ballpark, just try a current balun.
|
12th Oct 2016, 8:53 pm | #43 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
I also tried fitting the 220pF 'tuning' cap inline to my T124-2 transformer as I suggested back in post #27 to see if it tuned out the net inductance at the transformer input somewhere around 18MHz.
In my case with the T124-2 transformer the VSWR dip was closer to 16MHz than 18MHz but you can see that it is possible to tune out the mismatch and get good power transfer to the 50R load at this frequency with a 220pF cap inline. If there's an ATU before this transformer then this would help tune out this stray inductance anyway. So in this sense you could argue that it doesn't matter that much if the transformer transforms the 50R to something different. The ATU can tune it out (within reason of course). I've also added a word doc image to show the effect of adding this cap inline to both of the the simulated transformers. You can see that the simulation for VSWR agrees very closely and this adds further confidence that the transformer behaviour can be modelled fairly well using simple models. The VNA plot shows the response with a real 220pF cap and a real T124-2 transformer and the simulations both look to be very similar when the VSWR plots are compared
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 12th Oct 2016 at 9:10 pm. |
12th Oct 2016, 10:19 pm | #44 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Thanks for all the replies. Wow Jeremy those simulation do look really close to my measurements as does the VNA plot. I'm amazed how good the simulation is.
So are we saying that it would be possible to make the transformer work maybe for one band but never the whole hf band. I guess this is never going to work but it's still really interesting to see how the performance can be optimised and understand why it does not function as I intended. I was also surprised that the lead length on the transformer could affect the performance I thought that because the wire length is only a tiny fraction of the wave length they would be of no significance I think I need to take away everyone's comments and do some more background reading and some more performance tests and try to produce my own models. what's interesting is that there is a commercial product available. I can only assume that the dc isolation must therefore be by some other method than an isolation transformer. Thanks all
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww |
13th Oct 2016, 10:13 am | #45 |
Heptode
Join Date: Dec 2014
Location: Limerick, Ireland.
Posts: 901
|
Re: 1:1 RF Transformer Design SWR problems
It's possible to design a transformer that works 1:1 over the whole HF band (1.5MHz to 30MHz), but it would not want to be connected direct to an aerial, but a matching unit close to the aerial. They exist inside 200W transceivers.
|
13th Oct 2016, 10:43 am | #46 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Hi Mike
Thanks for your comment which I think implies hope that its possible. I have not suggested that the antenna would be connected to the transformer. I have an external automatic ATU directly driving an end fed wire. I want to install the transformer in the 50 coax feeder between the ATU and the transmitter. The transformer could be installed adjacent to the ATU. My preference would be to install it in the house close to the transceiver thus isolating the house mains EARTH from the external feed cable running down the garden. The test that I have done, show that my current design does not work this is backed up by the modelling Jeremy did and his VNA measurement on a similar transformer. I think we can see that my design efforts and calculation have clearly been a failure. So how do we design a wide band flux link transformer with flat SWR from 1 – 30 MHZ – 100 Watt Power Handling matched to 50 ohm feeders. 1) What core Size 2) What core Material 3) How many windings 4) What is the Winding arrangement? 5) Do need parallel winding as suggested by Jeremy 6) If the length of the transformer tails is critical how do we mount the core and have a enough wire for connectors. I’m happy to gather the materials and do the construction and do the modelling but I need some assistance tying down some of the above. I have done quite a lot of investigation online but have not found a reference design similar to my application perhaps this is a telling point.
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww |
13th Oct 2016, 4:05 pm | #47 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
My input to the thread was to explain why your artificial VNA test was giving a poor VSWR into its floating load on the secondary winding. I'm not sure that this test setup reflects what you will see from a given transformer when installed in the setup you describe above because of the finite isolation between your house and the outside ground. You may end up having to try various transformer designs to find the best one for your application above. But this transformer might not perform as well when you test it with your existing VNA setup with a fully floating load and you look at the VSWR into that floating 50R load. In other words we could explore the path of designing a transformer that does what you want (in terms of VSWR) when tested with your existing VNA test jig but this transformer might not work as well in your actual installation as an alternative transformer optimised for your 'real' installation.
__________________
Regards, Jeremy G0HZU |
|
13th Oct 2016, 5:12 pm | #48 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Putting the mains aspect aside I consider this to be very much a theoretic exercise. I have a specification for the design of wide band flux link transformer with flat SWR from 1 – 30 MHZ.
My original post was to show the calculations I had done and core and winding choices that I made to build such a device. My design was obviously wrong as was my test setup. However the original problem remains. Why did my original design fail, where are my calculation wrong, what am I neglecting in my design. If what Im trying to achieve is impossible than I am happy to accept that my current understanding of the theory is incomplete but I cannot determine this for myself as yet. Jeremy please could you expand on your earlier comments about putting multiple winding in parallel. I’m not quite sure I follow the mechanical winding method you’re describing. I think this is to reduce the primary inductance and again why this is a factor regarding the performance. Regards Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww |
13th Oct 2016, 5:16 pm | #49 | |
Guest
Posts: n/a
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
|
|
13th Oct 2016, 5:19 pm | #50 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Oh - ******
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww |
13th Oct 2016, 7:07 pm | #51 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Quote Jeremy in Post 21
" But in your case, if the Zo is up around 80R in your 1.5m bifilar winding then I think you will see quite a difference in voltage across each 25R load once you get up around 25-30MHz. One load will display a voltage that will probably stay fairly constant with frequency and one will lose a fair bit of amplitude. The voltage may go down by a factor of two on one of the loads as you approach 25-30MHz?" Jeremy you were right but my transformer started to get out of balance at 14 Mhz. See Attached. Please can someone explain what caused the imbalance. One thing I did note was that the 817 did not indicate any SWR reading until it hit 50Mhz when it when it went to full scale. I think the SWR indicator on the 817 is known for being deaf. I was using 500mW on FM. Yes long scope ground clips - all I had to hand x 10 antenuation Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww |
13th Oct 2016, 8:25 pm | #52 | ||
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
If so, a current balun transformer isn't appropriate. The remote ATU at the antenna has an unbalanced 50 Ohm connection to a similar transceiver in th e shack, so nothing is balanced and a current balun doesn't give LF isolation. All that seems to be wanted is a 1:1 isolating transformer good for the HF bands. My first attempt would be with a nice big toroid in something like FT43 material with a dozen turns of RG58 on it. Connect coax from the radio across the ends of the inner of the winding, connect the coax to the atu across the ends of the braid of the coax winding. It won't be perfect, but the ATU can be used to offset any errors. You get the insulation of the RG58 dielectric. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
||
13th Oct 2016, 9:01 pm | #53 | |
Guest
Posts: n/a
|
Re: 1:1 RF Transformer Design SWR problems
Back to the original question
Quote:
|
|
13th Oct 2016, 10:54 pm | #54 |
Heptode
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
|
Re: 1:1 RF Transformer Design SWR problems
Thanks merlin. For the link I was aware of this. The only prescribed option is to change the earth system and install a 16mm SQR cable. It was this article that made me want to consider alternative solution. The RSGB aren't going to advocate solutions that have any potential risk. I think it's possible to mitigate these risks. I work with 500vdc and 25kv on a daily basis designing SIL 3-4 railway signalling equipment. What i don't have is a detailed knowledge or experience of RF design.
the isolation transformer can undoubted provide safe galvanic isolation it is commonly used. If it can be made to function with acceptable rf performance well that a different question and the question I want to explore in this thread. Best regards Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww Last edited by ChristianFletcher; 13th Oct 2016 at 11:06 pm. |
14th Oct 2016, 11:31 am | #55 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
I must confess I've been confused a bit wrt the actual installation details... it wasn't until post #46 that I realised that this wasn't something outdoors up near the antenna. I assumed you wanted isolation in the braid at both DC and RF so I suggested a classic current balun/choke and a DC blocking cap. Maybe a block diagram would have helped here
However, the main reason I joined the thread at post #15 was because nobody was answering (what I thought was) the main question in the original post as below. Quote:
Maybe I've clogged the thread up too much with simulations/models to try and help explain the issues with the floating load test. Sorry about that. However, I did also try and give pointers how to improve the transformer way back in my second post. i.e. switch to a higher permeability material like ferrite allowing the use of a shorter winding and use a tline with a lower characteristic impedance than the original (80R?) transformer. This will give a big improvement in performance
__________________
Regards, Jeremy G0HZU |
|
14th Oct 2016, 11:45 am | #56 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
It's possible to predict the voltage levels in each 25 ohm resistor up at the high frequency end of the transformer's range by just applying some basic transmission line theory to the crude tline model for your transformer. This can be demonstrated with a simulator or the voltage levels at each load can also be predicted reasonably quickly with just a calculator and some tline theory. Once you see why this happens it makes sense to try to shorten the transmission line and also it makes sense that the 80R? characteristic impedance of your winding is a bit high. If you were keen you could do it by hand without the calculator as the equations aren't that complicated. The result will depend on the length of the winding and its characteristic impedance. I can show you how to do this but I'm concerned that I'm going to clog up the thread with more attachments etc.
__________________
Regards, Jeremy G0HZU |
|
14th Oct 2016, 12:36 pm | #57 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
Once you extract the 4 port (s parameter data) model from the VNA you can then use it in a simulator as a component with a connection pin for each wire in the transformer. You can connect other things to it or you can ground the pins or do pretty much anything you like with it and the model will predict its behaviour very well indeed. If you go to the Minicircuits website they sometimes supply 4 port models of their lumped transformers and these models work well up into the UHF range. http://194.75.38.69/products/transformers_sm_a.shtml Click on a device and then you can see it offers you the option to download a 4 port s parameter model of the transformer. http://194.75.38.69/MCLStore/ModelIn...88618991920614 The model lets you import the transformer into a circuit in a simulator and it should capture all its small signal (steady state) behaviour very accurately indeed. When I took the 4 port model of my ferrite transformer and my red T124-2 transformer I did effectively the same thing as Minicircuits using my VNA to produce an accurate 4 port model. This allowed me to connect it up in various configurations as either a current balun or as an isolation transformer and see how it performs in an RF simulator program on a PC. The alternative is to try and produce a crude model like my basic tline model and use this instead. The tline model should work really well but it won't be as accurate as the 4 port model from a VNA
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 14th Oct 2016 at 12:46 pm. |
|
14th Oct 2016, 5:03 pm | #58 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Quote:
If you wanted to make a transmission line winding with a characteristic impedance (Zo) of 25R using wires you could do it by twisting a pair of wires together to get a 50R transmission line. Then make an identical 50R line using twisted wires and then connect/wind them in parallel. You will then get a Zo of 25R. For example, if you wanted to match 12.5R to 50R using a quarterwave transformer then you would ideally want to use a tline transformer with Zo = (50*12.5)^0.5 = 25R. You could use two 50R lines in parallel to get 25R Zo.
__________________
Regards, Jeremy G0HZU |
|
14th Oct 2016, 5:08 pm | #59 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
Maybe this will help as well...
If you go right back to your initial VNA test with the floating 50R load I think I can try and redraw this circuit in a way that makes it much easier to understand why you need a very low Zo to get the widest bandwidth with that (artificial?) test setup. i.e. I can suggest an 'equivalent' circuit that is far easier to understand in terms of the impact of winding length and Zo. I sat down and derived an equivalent transform for this circuit and it is easier to understand. See the image below. This circuit will behave just the same with a VNA test and the power transfer to the load will be the same across frequency as well. If your existing transformer has a Zo of 80R and it is 150cm long you can simplify the analysis of your circuit to (just) your 50R load resistor in series with a shorted transmission line of half the length but twice the Zo of your actual transformer. See the image below. You can see in the simulation results for VSWR and the smith chart that the impedance for both circuits is identical across frequency as both graph traces lie directly on top of each other.. A shorted tline will look like a series inductance at low to medium frequencies and this will spoil the match. The impedance (Zo) of a transmission line is (L/C)^0.5 where L is the inductance per meter and C is the capacitance per metre. In your VNA test setup it is the inductance per metre that is spoiling your test result as it is placing an unwanted inductance is series with your 50R test load. So to 'reduce' the inductance per metre you can reduce the Zo of the transformer if you study the above equation. The other way to get less inductance is to simply shorten the transmission line length of your winding (obviously) and this is why I suggested way back in my second post that you should try a shorter winding and also a lower Zo and a core with higher permeability. Hope this all make sense
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 14th Oct 2016 at 5:21 pm. |
14th Oct 2016, 5:48 pm | #60 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: 1:1 RF Transformer Design SWR problems
For completeness, see the image below for a simulation of what happens when you shorten the winding of your transformer down to 80cm and you use a lower tline Zo of 50R and then you test it with a VNA. I assume you also use a core with higher permeability to preserve performance down at a few MHz.
The simplified equivalent will therefore be a 50R load resistor in series with a shorted tline 40cm long (i.e. half as long) with Zo = 100R (i.e. twice the Zo). The simulation below shows that the two equivalent circuits still agree and they show a useful improvement in VSWR at 30MHz
__________________
Regards, Jeremy G0HZU |