Great video again, Manuel. I never done KN5L method on FT82-43, only on FT114-43 and FT140-43. I noticed that efficiency on single FT114-43 was 86-89% 28:4turns, but the same KN5L method on 2xFT114-43 was 92-93% on 80-15m and 89% on 10m band. I think it is worth trying stacked FT82-43.
It's easy to end up pulling the widings tight only in one direction! Either away from you, or towards you! Remember to pull it tight in BOTH directions so you get a nice tight coil.
Merci Manuel! I’m anxious to see you measure the real losses at 5 watts. I am starting to think that the losses on the transfo for efhw are enormous, and nobody realizes it. Good work Manuel, Best!
I realized, I should have added that to the Video, but it was already uploading. So: The back to back results were way worse compared to the original measurement of the same sized core. I will probably do a real world test, to make sure it has nothing to do with my test setup / enviroment. 73 Manuel; DL2MAN
Thanks, I wonder about that. Losses are expected, but I would have expected 1 %. I am really interested in this, I have had no success with efhw, compared to dipoles.
The way this works, is that the primary winding needs to have enough turns to achieve an impedance of 150-200 ohms. This is why there is 2 turns on the typical FT240-43 and 3 turns on an FT140-43. A smaller core needs more turns. The more secondary turns, the more the high frequency response is adversely affected, with increasing stray L and components. So a trade off. The capacitor deals with stray inductance at the higher frequency end and is used to tune for best performance. It isn't necessary, but useful to tweak better performance. The other point is that binocular cores are naturally more efficient (this is why they are used in rf amplifiers and not toroids). I don't understand why toroids continue to have the focus of attention. Efficiency is improved by stacking - two FT140-43 are about 5-6% more efficient than a single FT140-43, only require a turn primary and therefore have a better HF response.
Hello, the main measure is missing: the insertion loss! Assemble a second identical transformer and connect the two high impedance secondaries together. The primaries will go to the two ports of the VNA. Measure the transfer loss (Gain) and divide by two. Those who do not have a VNA can do the same test with an rtx, a power meter and a dummy load. Let us know. 73s.
While I'm sure you can add sufficient capacitance by increasing the inter-winding capacitance this way, what I am curious about is any efficiency differences between the 2 different coil making methods, the bulk capacitor vs distributed capacitance. I have lots of capacitors on hand to use, but I'd like to use whichever method works best. Would you do an efficiency test for us in order to compare the 2 coils when you have an opportunity? Thanks.
I realized, I should have added that to the Video, but it was already uploading. So: The back to back results were way worse compared to the original measurement of the same sized core. I will probably do a real world test, to make sure it has nothing to do with my test setup / enviroment. 73 Manuel; DL2MAN
Excellent work winding the core, Manuel!👍 I know your fingers must have been approaching ‘cramp-threshold’ after all those previous attempts!👀 Mine would have been. Hope the TX results will be fruitful. These videos really give a boost to real-world QRP ops! 73
Could the ohmic loss in the primary be a reason for low efficiency when tested with 5W input? It's half the thickness of a traditional bifilar winding and will be carrying 7x the current of the secondary.
I think I'm feeling stupid at the moment. I've watched you wind that core at least five or six times now and it seems to me like you are winding up with effectively a 3:18 turn transformer. Winding the secondary the way that you are seems to me like the first 3 turns of the secondary are canceled out by the primary and subtracted and you wind up with a 3:18 Turn or 1:36 impedance ratio transformer instead of 3:21 and 1:49 respectively. What is the value of your load resistance. If you were figuring on 1:49 but only got 1:36 impedance ratio you might get an even lower SWR if you added three more turns to the secondary. Someone else mentioned stacking two 82-43 cores. That might be worth trying. Also have you tried 3:24 Turns ratio for 1:64 impedance ratio? TNX
I have used your alternate whining method multiple times and it has worked very well. Is there an alternate winding for using it for traps for a multi-band antenna?
👍Thanks for sharing. Is that x5 15K parallel resistors (3K) load you have? I have made some UNUNs for EFHW where I ended up doing a similar thing to get a good match. In my case, I wound the primary turns over the secondary turns.
Yes, and did that today. But before someone starts winding, read further till the part where Manual says: "I turned back to the original design with cap, beceause of the loss" Test 1. First tried this method, with the 100Pf capacitor, removing the cap was improvement, adding 47 Pf even more. Test 2. So, rewound it with 3 Primairy + 3 and then wound back, 47Pf was still there and removing made it even better. (adding 100Pf did not) What i always do: add 1 and remove 1 winding from the secundairy side, adding 1 gave improvement, possible i miscounted. What i also discovered: it got wider, even on 50 Mhz under 1:2 results: 3.65 Mhz 1:1.6 14.14 Mhz 1:1.47 21.5 Mhz 1:1.29 28.9 Mhz 1:1.05 Again, nice experiment. Not suitable for real life due to losses. 73, imre
Thanks for the video Manuel! I tested the KN5L windig after your hint as well this afternoon. And compared it with the "normal" 3to21 and 4to28 winding on a Würth core (similar size as a ft82). With the KN5L winding and 4to28 I got good results and even better with a 33pF capaciator. The würth core has a lower AL than the Amidon. I will test tomorrow also your Suggestionen of 3to21. But according AF7NX, there is maybe one general downside with this winding pattern. You will physically cross the low site (GND connection) wire with approx. 1/3 of the high site wire on the core. So this wires are very close and can have voltage difference of >210Vpp @10W and more. What do you think? 73 de Jan
I'm not Manuel, but I'd suggest a bit of electrical or Kaptan tape, or position the cross-over point a bit up and squeeze in a touch of epoxy adhesive to keep them apart.
Hi Jan, that´s a valid point ! if the enamel coating was damaged, this will give very ugly effects in deed. I found anyway, that this pattern was not as efficient as expected in my back to back measurement. Actually pretty dissapointing and worse compared to the more "traditional" winding pattern with C. Practice test will follow. With your great PCB´s. 73 Manuel; DL2MAN
@@dl2man Hi Manuel, sounds interesting! one other thing I was Thinking about: As you changed the winding directing after the 3 "primary" windings, maybe you need to add 3 additional windings on the "secondary" coil to get the 1:49 because it is a auto transformer and he flux is 180° shifted?
This is really exciting - this week I bought ten of these toroids and 10M of that wire to experiment on. In another video you mentioned using an FT140-43 as the 1:49 UnUn for the 100W version of the U01. Have you tried this winding pattern on the bigger toroid? Cheers! -Rob VK4SSE
No, I didn´t test this Pattern at all in practice. It made me suspecious, as the windings are very close together where we have higher Voltages. So I went with classical 2:14 + cap Design on my 100W Transformers.
@@dl2manI have stacked two 140-43s and made the usual 14:2 with the crossover, and first 2 turns twisted together. The 100 pF capacitors are in the mail. Will this work for 100 watts? I am a (relatively) new HAM and have purchased my first 100 watter.. Please help!! 73 de VU2EHC
@@drankurbaruahYes it will. 80 and 40m are fine with swr around 1.5:1 but I find it best to use an atu on 20 to 10m. You will go nuts trying for low vswr on all bands. Cheers from VK5LB
Great video again, Manuel. I never done KN5L method on FT82-43, only on FT114-43 and FT140-43. I noticed that efficiency on single FT114-43 was 86-89% 28:4turns, but the same KN5L method on 2xFT114-43 was 92-93% on 80-15m and 89% on 10m band. I think it is worth trying stacked FT82-43.
It's easy to end up pulling the widings tight only in one direction! Either away from you, or towards you! Remember to pull it tight in BOTH directions so you get a nice tight coil.
Merci Manuel! I’m anxious to see you measure the real losses at 5 watts. I am starting to think that the losses on the transfo for efhw are enormous, and nobody realizes it. Good work Manuel, Best!
I realized, I should have added that to the Video, but it was already uploading. So: The back to back results were way worse compared to the original measurement of the same sized core. I will probably do a real world test, to make sure it has nothing to do with my test setup / enviroment. 73 Manuel; DL2MAN
Thanks, I wonder about that. Losses are expected, but I would have expected 1 %. I am really interested in this, I have had no success with efhw, compared to dipoles.
Thanks!
The way this works, is that the primary winding needs to have enough turns to achieve an impedance of 150-200 ohms. This is why there is 2 turns on the typical FT240-43 and 3 turns on an FT140-43. A smaller core needs more turns. The more secondary turns, the more the high frequency response is adversely affected, with increasing stray L and components. So a trade off. The capacitor deals with stray inductance at the higher frequency end and is used to tune for best performance. It isn't necessary, but useful to tweak better performance.
The other point is that binocular cores are naturally more efficient (this is why they are used in rf amplifiers and not toroids). I don't understand why toroids continue to have the focus of attention. Efficiency is improved by stacking - two FT140-43 are about 5-6% more efficient than a single FT140-43, only require a turn primary and therefore have a better HF response.
Binocular are better? I'll try and build mine like that in future! Thanks.
Hello, the main measure is missing: the insertion loss!
Assemble a second identical transformer and connect the two high impedance secondaries together. The primaries will go to the two ports of the VNA. Measure the transfer loss (Gain) and divide by two. Those who do not have a VNA can do the same test with an rtx, a power meter and a dummy load. Let us know. 73s.
While I'm sure you can add sufficient capacitance by increasing the inter-winding capacitance this way, what I am curious about is any efficiency differences between the 2 different coil making methods, the bulk capacitor vs distributed capacitance. I have lots of capacitors on hand to use, but I'd like to use whichever method works best. Would you do an efficiency test for us in order to compare the 2 coils when you have an opportunity? Thanks.
I realized, I should have added that to the Video, but it was already uploading. So: The back to back results were way worse compared to the original measurement of the same sized core. I will probably do a real world test, to make sure it has nothing to do with my test setup / enviroment. 73 Manuel; DL2MAN
Excellent work winding the core, Manuel!👍
I know your fingers must have been approaching ‘cramp-threshold’ after all those previous attempts!👀
Mine would have been.
Hope the TX results will be fruitful.
These videos really give a boost to real-world QRP ops!
73
For sure my fingers have hurt ;) But that´s the price when you want to find out something. As you said: this is real world ;) 73 Manuel; DL2MAN
Awesome video, thank you for the idea! I used a FT114-43 I am on a peak today with contacts Arizona to Florida. It works great!
I'm loving these! Thank you for sharing!!
That is something new to me hehe. In the notebook now for more experiments to make 😁. Thanks for sharing. 73
what value for your dummy load over antenna and earth terminal please ? the resistor ?
Could the ohmic loss in the primary be a reason for low efficiency when tested with 5W input? It's half the thickness of a traditional bifilar winding and will be carrying 7x the current of the secondary.
Ohmic loss ? I don´t think so. On thermal images you usually see the core heating up, not the wire.
I think I'm feeling stupid at the moment. I've watched you wind that core at least five or six times now and it seems to me like you are winding up with effectively a 3:18 turn transformer. Winding the secondary the way that you are seems to me like the first 3 turns of the secondary are canceled out by the primary and subtracted and you wind up with a 3:18 Turn or 1:36 impedance ratio transformer instead of 3:21 and 1:49 respectively.
What is the value of your load resistance. If you were figuring on 1:49 but only got 1:36 impedance ratio you might get an even lower SWR if you added three more turns to the secondary.
Someone else mentioned stacking two 82-43 cores. That might be worth trying. Also have you tried 3:24 Turns ratio for 1:64 impedance ratio? TNX
Well, now to dig out my spare cores and . . . Thanks for the hints. 73 de W8IJN
I have used your alternate whining method multiple times and it has worked very well. Is there an alternate winding for using it for traps for a multi-band antenna?
👍Thanks for sharing.
Is that x5 15K parallel resistors (3K) load you have?
I have made some UNUNs for EFHW where I ended up doing a similar thing to get a good match. In my case, I wound the primary turns over the secondary turns.
Thanks! Great wristwatch, by the way 😉👍 73!
Can i do with 140/43 toroid? Thank's
Yes, you can do that of course. However, I´d suggest to go 2:14 Turns in case of FT140-43.
Yes, and did that today. But before someone starts winding, read further till the part where
Manual says: "I turned back to the original design with cap, beceause of the loss"
Test 1. First tried this method, with the 100Pf capacitor, removing the cap was improvement, adding 47 Pf even more.
Test 2. So, rewound it with 3 Primairy + 3 and then wound back, 47Pf was still there and removing made it even better. (adding 100Pf did not)
What i always do: add 1 and remove 1 winding from the secundairy side, adding 1 gave improvement, possible i miscounted.
What i also discovered: it got wider, even on 50 Mhz under 1:2
results:
3.65 Mhz 1:1.6
14.14 Mhz 1:1.47
21.5 Mhz 1:1.29
28.9 Mhz 1:1.05
Again, nice experiment. Not suitable for real life due to losses.
73, imre
Thanks for the video!
Nice experiment, why is there only one connector (Antenna) ? There should be another connector for the receiver, right? How does this work?
Hi, please watch Part1 to cover the basics and you will understand how it works. th-cam.com/video/2-4J8ECkoe4/w-d-xo.html 73 Manuel; DL2MAN
Thanks for the video Manuel! I tested the KN5L windig after your hint as well this afternoon. And compared it with the "normal" 3to21 and 4to28 winding on a Würth core (similar size as a ft82). With the KN5L winding and 4to28 I got good results and even better with a 33pF capaciator. The würth core has a lower AL than the Amidon. I will test tomorrow also your Suggestionen of 3to21.
But according AF7NX, there is maybe one general downside with this winding pattern.
You will physically cross the low site (GND connection) wire with approx. 1/3 of the high site wire on the core. So this wires are very close and can have voltage difference of >210Vpp @10W and more. What do you think?
73 de Jan
I'm not Manuel, but I'd suggest a bit of electrical or Kaptan tape, or position the cross-over point a bit up and squeeze in a touch of epoxy adhesive to keep them apart.
Hi Jan, that´s a valid point ! if the enamel coating was damaged, this will give very ugly effects in deed. I found anyway, that this pattern was not as efficient as expected in my back to back measurement. Actually pretty dissapointing and worse compared to the more "traditional" winding pattern with C. Practice test will follow. With your great PCB´s. 73 Manuel; DL2MAN
@@dl2man Hi Manuel, sounds interesting! one other thing I was Thinking about: As you changed the winding directing after the 3 "primary" windings, maybe you need to add 3 additional windings on the "secondary" coil to get the 1:49 because it is a auto transformer and he flux is 180° shifted?
This is really exciting - this week I bought ten of these toroids and 10M of that wire to experiment on. In another video you mentioned using an FT140-43 as the 1:49 UnUn for the 100W version of the U01. Have you tried this winding pattern on the bigger toroid? Cheers! -Rob VK4SSE
No, I didn´t test this Pattern at all in practice. It made me suspecious, as the windings are very close together where we have higher Voltages. So I went with classical 2:14 + cap Design on my 100W Transformers.
@@dl2manI have stacked two 140-43s and made the usual 14:2 with the crossover, and first 2 turns twisted together. The 100 pF capacitors are in the mail.
Will this work for 100 watts?
I am a (relatively) new HAM and have purchased my first 100 watter..
Please help!!
73 de VU2EHC
@@drankurbaruahYes it will. 80 and 40m are fine with swr around 1.5:1 but I find it best to use an atu on 20 to 10m. You will go nuts trying for low vswr on all bands. Cheers from VK5LB
@@Yosemite_Sam Thanks for the reply. I built it and it worked well.. made contacts with Cape Town from India!!!
😊😊 Super happy!!
73 de VU2EHC
Danke
I enjoy this series! vy 73 de Mike
Nice video. Please, AWG number? Thanks, Friend. 73 PU2AOO
I wound that one with AWG22 / 0,65mm on FT82-43 core.
@@dl2man Many thanks my friend. 73