Towertalk
[Top] [All Lists]

[TowerTalk] FW: question about a matching transformer

To: <towertalk@contesting.com>
Subject: [TowerTalk] FW: question about a matching transformer
From: "Matt" <maflukey@gmail.com>
Date: Fri, 15 Jul 2016 01:24:30 -0500
List-post: <towertalk@contesting.com">mailto:towertalk@contesting.com>
Hi Gary,

I know you are looking at the quintifilar design but thought I would share this 
report.   I once built W2FMI 2:1 un-un based on his trifilar coaxial design 
using homemade 14 ohm coax on an FT240K - reference Figures 7-12(c) & 7-13 in 
the 4th edition.  I installed this un-un in a prototype stackmatch box similar 
to CN2R design except with etched circuit board.   I tested it behind an AL1500 
at legal limit.   The transformer ran very cool on power soak test and SWR was 
flat up to 15m indicating very good transformer efficiency.   At 10m the SWR 
was up just a bit and the transformer ran slightly warm - but not hot.   
Constructing the 14 ohm coax was a real pain, as was winding the un-un because 
it is so stiff.  As I was needing 3 sets of stack boxes I ultimately decided 
that the home-brew approach was more trouble that it was worth and abandoned 
the prototype for a set of WX0B boxes.  2.25:1 ratio is not an issue.   
Incidentally the applications are all for double stacks - not triples.

Hope this info helps & good luck on your project!

Matt
KM5VI



-----Original Message-----
From: TowerTalk [mailto:towertalk-bounces@contesting.com] On Behalf Of 
StellarCAT
Sent: Tuesday, July 12, 2016 7:21 PM
To: tower <towertalk@contesting.com>
Subject: [TowerTalk] question about a matching transformer

Hello, 

this is to those experts out there familiar with toroidal transformers. 

The common one used in boxes like the “StackMatch” is a 2.25:1 transformer. it 
is tri-filiar wound on a 2.4” core. So it provides, when seeing 2 50 ohm loads 
(2 monobanders for example) an SWR of about 1.14:1. 

So in the WX0B version, shown in the ARRL AB, there are 4 turns. In the W2FMI 
version there are 6 turns. In either case the transformation is clear: 3 
‘coils’ divided down to 2 coils or a turns ratio of 1.5:1 so the impedance 
ratio would be 2.25:1 (1.5^2). 

But there is a version of it – shown in W2FMI’s book, that provides for a 2:1 
match. And my initial thoughts are – if I am using 2 monobanders and I can do 
2:1 why would I not do it? 

W2FMI shows again a tri-filiar wound coil – with 6 turns and the ‘top’ or end 
coil, that which is going to the source, is tapped down 1 turn from the source 
end. Thus you have essentially (5/6 + 2)/2 = 1.416:1 or an impedance ratio of 
2:1! 

That all is straight forward. I’ve never seen someone do this – but it seems 
again straight forward. 

So I have been TRYING to get it to work without any success. 

I start with a 240 core (tested both K and 61 material) ... and I have attached 
to it – short leads, 2 50 ohm resistors with a measured resistance of 25.3. And 
an SO239. I’ve checked it using a SARK110. This is not on a PCB – no relays, no 
board traces – just short leads on a transformer, 2 resistors and comp cap. 

So wound with 5 turns I can establish, using the K material, a relatively good 
response for the 2.25:1 version. (The 61 material is not quite as good). I have 
to use a compensation capacitor at the load input and have tested various 
values. The best ‘fit’, using the 2.25:1 ratio (trying to duplicate its results 
first) is to use a value of 75pf. That gives me an SWR of 1.14:1 on 20 and 15 
and 1.15:1 on 20. 

So I then tap down ~1 turn (for 5 turns tapped down 1 turn it should still be, 
in theory, 1.96:1 or an SWR of 1.02:1) ... and I cannot for the live of me get 
any kind of what seems like proper results. It is quite high on 10 meters – 
approaching 2:1 and typically 1.2:1 or worse on 20 depending on the comp cap. 
I’ve tested anywhere from no comp cap to 200pf in small increments ... it just 
doesn’t work. [note any comment about the choice of the resistor load, which 
are metal film, has to be considered under the light of – it works using the 
2.25:1 version!]

Can an expert perhaps write to me directly, not burdening others here, and 
hopefully point me in the correct direction please? This SHOULD work! 

Gary
K9RX at arrl dot net

[ref: The SARK has been OSL calibrated with the cable in use – and the resistor 
load has been checked at the end of that independent of the transformer and 
reads 2.00:1 from 14 through 29Mhz with 0 ohms reactance[ 
_______________________________________________



_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

_______________________________________________



_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
<Prev in Thread] Current Thread [Next in Thread>