UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

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.

Closed Thread
 
Thread Tools
Old 11th Oct 2016, 1:43 pm   #21
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

One quick test you could try would be to wire up your existing transformer like the image below using a couple of small 25R resistors grounded as the image below.

Then connect a sig gen to the input and increase the frequency and have a look at the voltages at each 25R load using a scope and a high Z probe.

I think in theory, you need to wind a transformer with bifilar windings with Zo = 25R to get the very best bandwidth in terms of getting even power in each 25R load. 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?

Note that you will probably see an improvement in input VSWR up at high frequencies with the two 25R loads and the centre tap ground when compared to the floating 50R load. But I think it will still perform badly in the balance test above once you get up around 25-30MHz.
Attached Thumbnails
Click image for larger version

Name:	transformer2.gif
Views:	106
Size:	2.8 KB
ID:	131124  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 11th Oct 2016 at 1:49 pm.
G0HZU_JMR is offline  
Old 11th Oct 2016, 2:52 pm   #22
Guest
Guest
 
Posts: n/a
Default Re: 1:1 RF Transformer Design SWR problems

Quote:
one big bonus of winding a classic Guanella/current balun (with shorter wires and a ferrite core) is that the power handling can be very good even with a modest sized ferrite core
And if it is an unbalanced to unbalanced transformer (as it seems to be) the core only has to take the difference which would be quite small.
 
Old 11th Oct 2016, 5:37 pm   #23
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

I suppose my other comment would be that this transformer might not work 'ideally' as a (50R) 1:1 current balun either. That is, if the design goal is to achieve low loss into a balanced 50R load. Obviously, a real antenna could be quite different to a 50R load and a real transmitter won't have a 50R source impedance, so maybe the following is a bit pedantic...

But... if I assume that the characteristic impedance of the bifilar winding is 75R (and not the ideal 50R) and it is 1.5m long then it will probably act as a QW transformer at just under 50MHz. So if you put a perfect 50R load at the end of a perfect 75R QW transformer the impedance seen at the source would be (75*75)/50 =112.5R. So even though 75R is only a 1.5:1 ratio to 50R the VSWR seen at the source at about 48MHz would be 112.5/50 = 2.25:1. Not that great.

So the magR would be (2.25-1)/(2.25+1) = 0.384

The loss at about 48MHz would then be 10*log(1-(0.384*0.384)) = 0.695dB
I know that 48MHz is out of band but the loss at 30MHz will probably be in the order of 0.5dB and the VSWR would be 2:1. Get the Zo down to 65R and the loss at 30MHz would drop to about 0.2dB with a VSWR around 1.6:1. Much better.

But this highlights the design grief you face when using a low permeability core and you have to resort to a long winding to get good LF performance. The Zo of the bifilar winding needs to be close to 50R to get low VSWR and low loss in a controlled lab test. It's very punishing at the high frequency end of the band if the Zo is not close to 50R.

I think it helps to wrap tape clamps every few cm along the winding to keep the windings parallel and control the Zo better. This way you can avoid having the bulge/gap spreading effect between the bifilar wires on the outside of the core. But the other way is to use a high permeability core and use a shorter winding length and this pushes the QW transformer effect way up into the VHF band and the performance at 30MHz will be much better
Attached Thumbnails
Click image for larger version

Name:	transformer3.gif
Views:	82
Size:	3.9 KB
ID:	131137  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 11th Oct 2016 at 6:05 pm.
G0HZU_JMR is offline  
Old 11th Oct 2016, 6:33 pm   #24
Guest
Guest
 
Posts: n/a
Default Re: 1:1 RF Transformer Design SWR problems

My AC isolation for my vertical transmitting aerial is merrily a few feet of co-ax with ferrite beads on it, the co-ax doesn't know they are there, it is just co-ax. Unwanted common mode currents are simply choked away. There is no need for posh RF good cores, in fact any old ferrite will do, the lossy EMC types are fine.

In short, don't don't even try to bother with a 1:1 conventional transformer, a current balun (unun), with DC blocking caps if needed, will beat the socks off one.
 
Old 11th Oct 2016, 8:32 pm   #25
Mike. Watterson
Heptode
 
Join Date: Dec 2014
Location: Limerick, Ireland.
Posts: 901
Default Re: 1:1 RF Transformer Design SWR problems

^^^^^ What he says! ^^^^^^

I have my coax wound on an old LOPT core from a B&W "solid state" set.
If you are worried about DC, then a couple of 2nF 1kV XY rated caps in a plastic box with pair of connectors back to back. (or BNC Computer network connectors on a metal box, they have isolated outer). If desired add thick earth wire to ground spike on aerial side and MOV/gas arrestor/sparkgap etc on centre to screen of aerial side connector.

About 20m of end fed horizontal wire. Awkward to get it well tuned on 160m. Unless you have a remote ATU where aerial connects to the coax, then the coax Aerial <->ATU wants to be maybe shorter than 1/8th wave on bands hard to tune. (The aerial has a aircored transformer match to 50 Ohms only valid at 80m band). For the shorter bands I have a nearly vertical 5.5m length of wire suspended at angle from top of gable (about 4m from gable wall at end fed base).

I've wound wide band o/p transformers for FETS on binocular cores by using very many fine wires in parallel, in two colours, then suitably soldering the very fat ends to make a sort of Litz effect bifilar push-pull primary. Secondary about 20 wires twisted and in parallel, two turns. I copied a RACAL design used on a 200W 24V transistor amp.

Only make transformers when no alternative on amps. My Minicircuits 50Mhz to 1GHz IC amp uses common mode wired 1:1 chokes for input and output as that is "wider band" than the SAME part in circuit as a transformer. Gives no DC isolation at all, so there are capacitors!
Mike. Watterson is offline  
Old 11th Oct 2016, 9:59 pm   #26
ChristianFletcher
Heptode
 
ChristianFletcher's Avatar
 
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
Default Re: 1:1 RF Transformer Design SWR problems

Thank you all for the replies and helpful comments. I have ordered a couple of new T200-2 core as I cannot be completely certain regarding the cores I have bought from a radio rally stall. when I get the cores later this week I will try out some of the tests and alternative winding arrangements.

I especially want to try the voltage measurement suggested . Thanks JMR

I do have a Automatic ATU installed remotely and its specifically designed to match the end fed wire to which the tuner is directly connected. This gives a good match for all the HF bands I suspect one of the setting must be a 50 ohm resistor -hi .

I am not sure why the impedance of antenna would be relevant given that the transformer wont care what the impedance is and there won't be any transformation given 1:1 turns ratio. The radio would just see a mismatched impedance from the antenna regardless of the transformer. If I have this wrong please correct me.


I think there may be some confusion regarding what I'm trying to build and perhaps I have caused the confusion by using words like Unun and Baluns. I am trying to make a 1:1 isolation transformer that works at HF frequency. It is very likely that I won't use the transformer especially if it won't work, but a project like this is interesting to me even if the answer is failure. I will still have learned a lot about rf transformer design.

Using a Capacitive shunt to ground is likely to be the easiest solution and I mentioned this at the start. However if I can get the transformer to work I believe it would be a safer solution given the possible short circuit failure mode of capacitors. There is a commercially device available, I saw it a the radio rally in Newark last week but cannot remember who the vender was. It was also cheap. This unit was being sold especially as a ground isolator for PME installations. so I'm following a logical path in that if they were on sale it must be possible to build one. ! Assuming the units for sale actually work.

I did not wind the transformer bifilar the windings were side by side. I thought that bililar winding may have caused problems as I was thinking the capacitance between windings would be increased a that this could contribute to problems. I will certainly give it a try if you think tightening the coupling will flatten out the SWR

So advise please on how to build 1:1 rf transformer and what factors in the design I must consider. thanks also for ideas on how to test and debug the design

Again thank you for all you kind replies and you all taking time to give your time and advice so freely.

Regards Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww
ChristianFletcher is offline  
Old 11th Oct 2016, 10:48 pm   #27
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

Quote:
am not sure why the impedance of antenna would be relevant given that the transformer wont care what the impedance is and there won't be any transformation given 1:1 turns ratio.
The issue here is that your transformer won't be perfect because it has transmission line properties. I suspect that your transformer will transform the resistive part of the 50R load OK but it will add some (unwanted) series inductance for free. Probably several hundred nH.

I'm nervous to suggest this because I'm not sure how good your VNA setup/calibration is but you could try sticking a series 220pF cap inline at the VNA (source) side of the transformer in your image of the VNA test. This would help to resonate out any net series inductance produced by your transformer. Put the cap close to the transformer in series with the coax centre line as in the image below.

Does this give a very low VSWR somewhere around 18MHz? If so, try a smaller cap and see if the decent VSWR point shifts upwards. eg a series 100pF might work well at 25MHz for example. If this does happen then I think it shows that your transformer is misbehaving and adding some series inductance and this inductance is degrading the VSWR.

But there may be other issues that will make it more complicated than this. However, to model your transformer as it is now I would just use a simple model of a 1.5m long choked/balanced transmission line of Zo of approx 75R connected 'sideways' in the way a conventional transformer is wired. Even a model as crude as this ought to be able to get reasonably close to demonstrating what effects you are seeing with your VSWR test.

I don't like your spectrum analyser 'thru loss' test setup for various reasons but the model above might help explain what your VNA is displaying with the floating 50R load. Also, you really should trim the flying leads on your transformer. They look very long and this isn't helping because they add yet more (uwanted) series inductance in series with the 220pF cap.
Attached Thumbnails
Click image for larger version

Name:	miniVNAtest.jpg
Views:	107
Size:	55.9 KB
ID:	131166  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 11th Oct 2016 at 11:10 pm.
G0HZU_JMR is offline  
Old 11th Oct 2016, 11:31 pm   #28
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

Here's a quick image showing a simulation for VSWR and loss for a crude model of your existing transformer (using a 1.5m long tline with Zo = 75R) vs the prediction for a transformer wound the same length but with a lower Zo of 25R and with shorter leads to the transformer.

The biggest benefit will be from the 25R Zo.

The red traces are for the model of your current transformer and the blue ones show the change to a Zo of 25R. You can see it predicts a useful improvement in loss and VSWR at 25MHz. A VSWR reduction from about 5:1 to about 1.8:1 and a reduction in loss from just over 2.5dB down to about 0.5dB.

Sadly, the crude model assumes a lossless tline and a perfect core with close to infinite permeability so the response is good right down to VLF. But your transformer is wound on a core with low permeability so the performance will degrade at LF. Apart from this the model should be reasonable up above 10MHz for example.

The best way to simulate your transformer would be to take a full 4 port s parameter model of it with a VNA. But this would be a lot of work with a 2 port VNA and you would ideally need access to a decent lab VNA with 4 ports.
Attached Thumbnails
Click image for larger version

Name:	XFMR_COMPARE.jpg
Views:	142
Size:	89.7 KB
ID:	131167  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 11th Oct 2016 at 11:38 pm.
G0HZU_JMR is offline  
Old 12th Oct 2016, 12:37 am   #29
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

I'd probably better add a confirmation that the above simulation is for the transformer terminated as a conventional transformer rather than as a classic current balun. This explains the poor performance in the red trace. I think the Zo needs to be very low to get best performance when terminated/arranged as a conventional transformer.

I really do agree with Merlin and Mike that a current balun with ferrite is going to be a much better performer than either of the above simulations. But I guess it's interesting to dissect the conventional transformer topology to see why it performs badly in the tests etc.

I don't have a T200-2 core here but I can have a rummage around to see if I have anything similar to play with. The biggest I can find so far is a T68-2 but I think I have some bigger ones in an old high power LPF if I can find it. I'm not used to working with components this big so it will be interesting to see how tricky it is to model/analyse it all adequately

I've uploaded a better image of the simulation to photobucket here.

http://i1227.photobucket.com/albums/...MR_COMPARE.gif


You can try clicking on the download gearwheel and I think this will give a sharper image.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 12th Oct 2016 at 12:55 am.
G0HZU_JMR is offline  
Old 12th Oct 2016, 1:45 am   #30
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

I can't find any big Txxx-2 cores but I did find an old 1:1 current balun I designed and built many years ago. I assume I used ferrite for it (looks like ferrite) and I think it has a medium permeability looking at the test results.

I have a fairly modern VNA here with a decent Ecal kit and I took a quick 4 port model of the transformer as per the image below. I have a basic 4 port test fixture that uses SMA connectors and I set the reference plane to the solder pin on the PCB end of the SMA connector. The VNA can do this very accurately and the 4 port model should be very good.

See also the image for the simulation for through loss and VSWR based on the s parameters taken from the VNA.

You can see that the performance (when configured as a current balun) is very good from about 1MHz through to 50MHz. The VSWR is better than 1.05:1 across most of the range and the loss is about 0.25dB at 30MHz.

Not bad for a twisted pair transformer. It's probably only good for moderate power levels because I just used basic enamelled wire but you can see that the Zo is very close to 50R. If I change the simulator away from 50R by just a few ohms the performance suffers slightly.

I really don't want to unwind it to find out the core permeability or shorten it etc but I think I can use this transformer to validate the basic model I posted earlier. I know that the Zo of this transformer is very close to 50R so it will be interesting to configure it as a conventional transformer and compare it to the crude model in my earlier posts.
Attached Thumbnails
Click image for larger version

Name:	TLT_4port.jpg
Views:	99
Size:	107.5 KB
ID:	131175   Click image for larger version

Name:	TLT_4portSIM.jpg
Views:	81
Size:	94.0 KB
ID:	131176  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 12th Oct 2016, 9:57 am   #31
ChristianFletcher
Heptode
 
ChristianFletcher's Avatar
 
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
Default Re: 1:1 RF Transformer Design SWR problems

Hello Jeremy - This is really good.

Many thanks for your very complete reply. You have given me quite a lot to think about and also to investigate.

I did try conecting the tranformer directly to the VNA and a dummy load resistor on the secondary and it had no measureable effect.


I have the following questions.

What are you using to do the modelling is this a SPICE model? Assuming its free or low cost package your using, would you let me have your model data / net list. I have a licenced version of Proteus that has a simulation engine for mixed mode analysis which I think is spice derived. However, it would be useful to use the same model you have constructed.

Yes not sure regarding the quality of my VNA. It’s always been reliable for setting up antennas etc but how accurate it is I would not like to guess. Its limited to 150 Mhz which has always been an issue for me as I work with telemetry links at 870 Mhz.
I have a couple of spectrum analyser a very good Marconi analyser to 500 MHz and a Rigol to 1.5Ghz. I could set this up via a directional coupler and use the return loss to calculate the VSWR. However, this is not going to give me any phase information un like the cheap VNA.

Jeremy if you fancy trying a T200-2 core I can ship you one over if you PM me your details etc. I’m waiting for these to arrive.

Thanks
Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww

Last edited by ChristianFletcher; 12th Oct 2016 at 10:15 am.
ChristianFletcher is offline  
Old 12th Oct 2016, 11:59 am   #32
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

I hope my input ultimately proves useful to you. I must confess that I do like looking at stuff like this and part of the appeal is that I'm slightly out of my comfort zone here. I wouldn't normally consider connecting up a transformer in the way you have and I'm not used to modelling with components this big!

But I can add more info as below. This morning I input the 4 port data from the VNA (that models my ferrite balun) into the simulator and configured the model to be as your conventional (isolation) transformer.

I also compared it to the crude transmission line transformer (TLT) model of the same circuit that uses a library model of a (perfectly choked) transmission line. However, I'm now able to refine this model slightly because I now know how lossy it is. It has about 0.35dB/m loss at 50MHz and the model lets me enter this data. I also know it has about 1.35m of winding length and a Zo of 50R.

The plot below shows what happens when I try and configure my ferrite transformer as an isolation transformer. You can see how closely the two agree! This is because the twisted bifilar winding has a characteristic Zo (of 50R in this case) and a known length of about 1.35m.

So it is fairly easy to model. You can see that the VSWR is not that great in either case despite the transformer working extremely well as a current balun in my earlier plots. You can also see that the crude transmission line model (red trace) falls apart at lower frequencies because it uses a perfectly choked tline. So it can't capture the degradation in performance at low frequencies caused by the finite permeability of the core in the real transformer (blue trace). But apart from this the ultra crude model does really well

Note that I've also uploaded the screen capture in a word doc. This seems to preserve the quality of the image and is much sharper meaning the text is legible.
Attached Thumbnails
Click image for larger version

Name:	GreyTLT.jpg
Views:	84
Size:	103.1 KB
ID:	131179  
Attached Files
File Type: doc greyTLT.doc (338.5 KB, 51 views)
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 12th Oct 2016 at 12:10 pm.
G0HZU_JMR is offline  
Old 12th Oct 2016, 12:02 pm   #33
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

Quote:
I have the following questions.
What are you using to do the modelling is this a SPICE model?
I'm using some fairly exotic tools here to model everything. However, I'll see if I can come up with a way to make the models useful on a freebie SW platform so you can play with them yourself. RFSIM99 might be something that could be used here but I don't think it accepts 4 port models.

I'll look at some alternatives and report back.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 12th Oct 2016, 12:13 pm   #34
ChristianFletcher
Heptode
 
ChristianFletcher's Avatar
 
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
Default Re: 1:1 RF Transformer Design SWR problems

The Silvaco tools look amazing. But if you have to ask the price......
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww
ChristianFletcher is offline  
Old 12th Oct 2016, 12:22 pm   #35
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

Things get a bit more interesting if I stack two transformers in parallel. In reality, you would just wind a pair of bifilar twisted pairs on the same core to achieve a similar result but the simulator lets me predict what would happen if I stacked TWO transformers in parallel in order to get a lower effective Zo (of about 25R?)

I've also changed the Zo of the crude TLT model to 25R and you can see that the two simulations still agree quite well. You can see an improvement in VSWR and insertion loss. It gets even better with three in parallel so it might be worth experimenting with three bifilar windings in parallel on a big core. But you can also see the performance degrading at LF on the blue trace. This is because the inductance of three windings in parallel is down to a third of what it was with one transformer.

However, I'm not really sure any of this is wise in reality. I'd be concerned about how it would cope with high RF power levels or with impedances that are far removed from 50R.
Attached Files
File Type: doc greyTLTstackedx2.doc (280.5 KB, 48 views)
File Type: doc greyTLTstackedx3.doc (282.5 KB, 48 views)
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 12th Oct 2016, 1:31 pm   #36
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

I had a play with an old freebie version of the classic SPICE simulator Simetrix and this has a lossless transmission line model that can be used in a similar way to my crude model in Genesys.

See the word doc below for simulations of a transformer using windings with Zo of 75R, 50R and 25R Zo.

It's better than nothing and shows how your transformer can be expected to misbehave depending of the winding Zo and length.

4.43ns is about 1.35metres long if I assume 1ns per foot. The 100Meg bleed resistor is there because Simetrix can't accept floating nodes. So I've added an extremely large resistor here to keep the simulator happy.
You can get a permanent demo of Simetrix on their website and it will easily run this circuit. It has a circuit size limit for the demo version but the limit is very generous.

http://www.simetrix.co.uk/site/index.html

I'll keep looking for something that can cope with a 4 port s parameter model because this is the best way to model a 'real' transformer.
Attached Files
File Type: doc SimetrixTLT.doc (186.5 KB, 54 views)
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 12th Oct 2016, 3:25 pm   #37
ChristianFletcher
Heptode
 
ChristianFletcher's Avatar
 
Join Date: Dec 2003
Location: South Yorkshire, UK.
Posts: 900
Default Re: 1:1 RF Transformer Design SWR problems

Hi Jeremy


I have down loaded and had a quick go with Simetrix and RFSIM99. Is there much difference with either of these. There looks to be more tutorial etc for RFSIM99.

I have to admit you have lost me regarding the number of ports and S parameters.

Regards Chris
__________________
Every Silver Lining Has Its Cloud https://youtube.com/channel/UCvBpiuUUnErJlNBm6DWb3Ww
ChristianFletcher is offline  
Old 12th Oct 2016, 5:53 pm   #38
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

RFSIM99 is probably not going to be able to cope with the model data but it is a classic linear simulator rather than a SPICE simulator. For now, Simetrix is the only free sim I can find that can (crudely) model your transformer.

I do have some more news. I managed to find a T124-2 (might beT130-2) core here. This was used as part of a bias tee ar one point in time. However, I rewound it using bifilar wires (untwisted) and tried to make it look as your transformer in terms of how the windings spread and compress etc.

I put 30 turns on it and this used about the same amount of wire as you would have. I measured it with the VNA to create a 4 port model and then I added the stray lead inductance that you had in your VNA test setup.

You can see the result below in the word doc. It looks pretty much the same as your first VNA plot of your T200-2 transformer. i.e. the VSWR is very poor at 25MHz.

Also in the simulation is the crude model using a 70R Zo winding as this xfmr appears to have a winding Zo of about 70R. Again, you can see that the two models agree once I include the same stray inductances for your long connection wires.

So I think this is a sign of progress because the models agree with the measurements and my VNA measurement looks very similar to yours
Attached Files
File Type: doc redTLT.doc (262.0 KB, 74 views)
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 12th Oct 2016, 6:21 pm   #39
Guest
Guest
 
Posts: n/a
Default Re: 1:1 RF Transformer Design SWR problems

For 1:1 or 1 to some square of an integer (4, 9 etc.) you can't beat a current transformer (transmission line transformer or common mode choke) or two (or three). You can get a bit silly when the winding length is a significant fraction of a wave length with matched transmission lines, in practice for a 1:1 use thin co-ax, 4:1 a twisted pair (out of ethernet cable innards) near enough to 100 ohms as makes no difference for a 4:1 (two transmission line transformers, inputs parallel outputs series). After finding about them (years ago) I have not even considered a normal (voltage) transformer for anything above a MHz or so.

Have you measured the isolation performance of your voltage transformer? You will find it gets worse as the frequency goes up, quite the opposite for a transmission line one. They (current/transmission line transformers) are almost too simple to think they work so well, they do.
 
Old 12th Oct 2016, 6:28 pm   #40
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: 1:1 RF Transformer Design SWR problems

In case some viewers are uncertain about all the 4 port s parameter stuff I moved the T124-2 transformer away from the 4 port test fixture and wired it up with a floating 50R load and made a measurement of VSWR directly with the VNA. So this is very similar to Chris's first VNA test setup.

My wire connections aren't as long as the ones you used Chris, but you can see that the VSWR looks fairly familiar. It isn't as bad as your initial VNA plot but I think my bifilar windings are closer/tighter than yours so your transformer may have a higher Zo than mine.

But in the image below at least it looks to be in the same ballpark
Attached Thumbnails
Click image for larger version

Name:	Red.gif
Views:	112
Size:	17.4 KB
ID:	131198  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 11:07 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.