EMRFD Message Archive 536

Message Date From Subject
536 2007-03-18 01:55:00 Ed Almos Matching Transformer With Dual Outputs
Having worked my way through EMRFD a few times I found this board. I
have never seen so much RF design talent gathered together in one
place, so I hope you do not mind me joining your little group.

I have a small problem which I would like to solve. This is not urgent
but any assistance that can be provided would be appreciated. I need
to wind a matching transformer with the following specifications:

Working frequency 45 MHz
Input (Primary) 50 ohms
Output # 1 (Secondary 1) 50 ohms
Output # 2 (Secondary 2) 200 ohms

The formula to construct a 4:1 balun is already known but do I need to
take into account the effects of secondary # 1? Does the primary see
an impedance of 40 ohms (200 ohms and 50 ohms in parallel)?

73

Edward Almos HA6SST
537 2007-03-18 17:23:18 Jim Kortge Re: Matching Transformer With Dual Outputs
Ed Almos wrote:
> Having worked my way through EMRFD a few times I found this board. I
> have never seen so much RF design talent gathered together in one
> place, so I hope you do not mind me joining your little group.
>
> I have a small problem which I would like to solve. This is not urgent
> but any assistance that can be provided would be appreciated. I need
> to wind a matching transformer with the following specifications:
>
> Working frequency 45 MHz
> Input (Primary) 50 ohms
> Output # 1 (Secondary 1) 50 ohms
> Output # 2 (Secondary 2) 200 ohms
>
> The formula to construct a 4:1 balun is already known but do I need to
> take into account the effects of secondary # 1? Does the primary see
> an impedance of 40 ohms (200 ohms and 50 ohms in parallel)?

Ed,

I think what you get is a transformer that looking into the primary you would
see 25 OHms. I modeled it in CircuitMaker and that indeed is what it is
showing. It also seems intuitive as I think about it that the equivalent is two
transformers, the first a 1:1 and the second a 1:2. Each primary would look like
50 Ohms with the respective secondary loads of 50 and 200 Ohms, and if you
connected the primaries in parallel, you would get 25 Ohms.

Anybody else have a comment. Its an interesting problem.....

72,

Jim, K8IQY
538 2007-03-18 18:31:33 Allison Parent Re: Matching Transformer With Dual Outputs
539 2007-03-18 18:43:50 kerrypwr Re: Matching Transformer With Dual Outputs
Sometimes we confuse "transformer" with "balun"; they are different
things, although a balun can transform and a transformer can convert
bal to unbal.

I think Edward just needs a transformer to convert the impedances; the
usual squaring of the turns ratio will apply to the 100:50 (2:1
impedance ratio/1.4:1 turns ratio) & 100:200 (1:2 impedance
ratio/1:1.4 turns ratio).

If a balun function is required, it may be implemented with a separate
1:1 balun.
540 2007-03-19 08:38:10 Ed Almos Re: Matching Transformer With Dual Outputs
Sometimes in RF work I do get the terms balun and transformer
confused, my apologies. Maybe we can look at this a different way, if
I wind (say) 8 turns on the primary and 8 turns on secondary #1 for
the 50 ohm output how many turns would I need to wind
541 2007-03-19 13:47:21 kerrypwr Re: Matching Transformer With Dual Outputs
I visualise it in this way;

Transformer 1 - Primary 100 ohms : Secondary 50 ohms and

Transformer 2 - Primary 100 ohms : Secondary 200 ohms.

When the two primaries are paralleled, the impedance will be 50 ohms.

So the turns ratios will be;

Transformer 1 - Primary 1.4 : Secondary 1 and

Transformer 2 - Primary 1 : Secondary 1.4.

Suppose we have just one primary winding of, say, 10 turns; secondary
1 will require 7 turns & secondary 2 will require 14 turns.

An alternative would be to wind two transformers and parallel the
primaries; I don't think that differing numbers of turns on each
primary would matter, it's the impedances transferred through the
transformers that matter.

Well, that's my shot at it. A very interesting question, Edward.
542 2007-03-20 08:55:50 Ed Almos Re: Matching Transformer With Dual Outputs
I agree with your theory however an experimental circuit has come up
with some different results.

I have some Minicircuits 2:1 impedance ratio transformers so I made up
a small circuit using two of them plus a 50 ohm resistor and a 200 ohm
resistor. After carefully making sure that everything was the right
way round I measured the impedance at 45MHz using my VNA. It showed
that I had a 42 ohm impedance at the input.

Some of this error may be due to my breadboarding technique but the
source of the 8 ohm difference is unknown.

Edward Almos HA6SST

543 2007-03-20 14:00:24 kerrypwr Re: Matching Transformer With Dual Outputs
I think that's a good result; SWR of 1.2 on the input.

It might be a product of the transformers.

An "ideal" 100:50 transformer will be slightly-different to an ideal
100:200 one; the Zo of the line (the GM of the two impedances) will be
different.

The MCL units are pretty broadband and MCL says that they can be used
at input impedances as low as 12.5 ohms but they can only be "perfect"
for one pair of impedances.

Unless you have a real need for 1:1 SWR on the input, I wouldn't be
too worried.
544 2007-03-20 14:20:37 Ed Almos Re: Matching Transformer With Dual Outputs
Well, when you express it as a 1.2:1 SWR it does not sound too bad.
There are of course variations in the windings of the transformers but
also the 50 ohm resistor is really 49.81 ohms and the 200 ohm resistor
was measured at 199.7 ohms. All of these will contribute to my error.

The easiest cure is an 8 ohm resistor in series with the input,
instant 50 ohm match.

Thank you for your help.

Edward Almos HA6SST

-- In emrfd@yahoogroups.com, "kerrypwr" wrote:
>
> I think that's a good result; SWR of 1.2 on the input.
>
> It might be a product of the transformers.
>
> An "ideal" 100:50 transformer will be slightly-different to an ideal
> 100:200 one; the Zo of the line (the GM of the two impedances) will be
> different.
>
> The MCL units are pretty broadband and MCL says that they can be used
> at input impedances as low as 12.5 ohms but they can only be "perfect"
> for one pair of impedances.
>
> Unless you have a real need for 1:1 SWR on the input, I wouldn't be
> too worried.
>
545 2007-03-20 15:18:24 kerrypwr Re: Matching Transformer With Dual Outputs
Because of the voltage-divider effect of the extra resistor you will
be giving-away a lot of your hard-won power.

And will the 50 & 200-ohm secondary loads be truly resistive and
exactly those values?
546 2007-03-20 20:33:13 Andy Re: Matching Transformer With Dual Outputs
> also the 50 ohm resistor is really 49.81 ohms and the 200 ohm resistor
> was measured at 199.7 ohms.

Measured at DC or at 45 MHz?

Use your VNA and see what impedance it gets separately for each of the two
transformers (with their loads).

There are lots of things that might make your readings differ from "theory."

Andy
547 2007-03-21 06:28:23 Ed Almos Re: Matching Transformer With Dual Outputs
You might as well have all of the tale in case you can suggest a
better solution.

Input to the transformer is from the 45 MHz first IF amplifier of a
prototype transceiver, a MAR-1 amplifier chip. Output impedance is
assumed to be 50 ohms

Secondary #1 is connected to a matching pad and then a TUF-3 mixer.
Input impedance is assumed to be 50 ohms.

Secondary #2 is connected to the input pin of a MC3371 FM detector
chip. Input impedance is stated as 200 -j13 ohms which is close enough
to 200 ohms.

I am of course open to suggestions for a better way of splitting the
signal between the two stages.

Edward Almos HA6SST


548 2007-03-21 06:47:04 Kenneth E. String... Re: Matching Transformer With Dual Outputs
Have you considered that one or both of the turns ratios is not
perfect? I believe the specified input VSWR of these transformers is
< 1.5:1 in all cases. At 100 ohms, this would be between:

~67 and ~150 ohms

If you parallel these transformers and assume one transfomer
reflects 67 ohms and the other 150 ohms (worst case). The net result
is ~46 ohms. This would make your 42 ohms seem reasonable.

Ken

550 2007-03-21 14:00:12 kerrypwr Re: Matching Transformer With Dual Outputs
I think your method of splitting is most ingenious.

The problem lies in your quest for perfection!!

No "50-ohm" device is truly so. AE1X has pointed-out that the specs
of the MCL transformers indicate that they are not "perfect". Another
example is the MMIC; whilst they are nominally 50-ohms they also are
not "perfect".

The other devices you are using will have similar imperfections.

Experienced designers know how much slack to cut. As a "mug" designer
myself I am often tricked into seeking a level of excellence that's
usually unnecessary and often impossible to achieve.

I'm still learning how much slack to cut!! :)
552 2007-03-22 05:55:35 Kenneth Stringham Re: Matching Transformer With Dual Outputs
Another way to achieve the desired separate between
paths is to use a splitter combiner. Are we talking
about high or low power here? You could use a two (2)
way splitter and use transformers that match
impedances to 50 instead of 100. You will get two (2)
equal signals with minimum loses using this approach.

Ken



____________________________________________________________________________________
We won't tell. Get more on shows you hate to love
(and love to hate): Yahoo! TV's Guilty Pleasures list.
http://tv.yahoo.com/collections/265
553 2007-03-22 05:58:23 Kenneth Stringham Re: Matching Transformer With Dual Outputs
Another advantage to the splitter topology is that
there will be isolation between ports. The arrangement
with the two primaries in parallel will permit signals
to couple between ports and possibly generate unwanted
mixing products or other spurious signals.

Ken




____________________________________________________________________________________
It's here! Your new message!
Get new email alerts with the free Yahoo! Toolbar.
http://tools.search.yahoo.com/toolbar/features/mail/
554 2007-03-22 07:05:24 John Marshall Re: Matching Transformer With Dual Outputs
Perhaps you don't really need a transformer to do this. You don't say
what value matching pad you will use ahead of the TUF-3 mixer but for
the sake of this suggestion let's say it's 3dB. A pi attenuator
designed for 50 ohms in and out calculates to 292 ohms shunt, 17.6
series, 292 shunt. Instead of winding a transformer, why not let the
input impedance of the MC3371 replace part of first attenuator shunt
resistor?

MAR-1 output goes through 92 ohms to MC3371 input and through 17.6
ohms to TUF-3 mixer in shunt with 292 ohms. The 92 ohm resistor in
series with the MC3371 input adds some attenuation, but only 1.6dB.
The values will be different if you planned to use a different
attenuator value ahead of the TUF-3, but easy to calculate.

Good luck,
John, KU4AF
Pittsboro, NC

555 2007-03-22 07:48:42 Ed Almos Thanks KU4AF
Thank you John, your reply made me think of an alternative solution
based on your post.

Draw out a 2.3dB T attenuator, this is fitted between the MAR
amplifier and the mixer at the front of the SSB stage.

The two horizontal resistors are 6.58 ohms, a standard value of 6.8
ohms will be close enough. The vertical resistor to chassis is 186.63
ohms, this is not a real resistor but the 200-13j ohms impedance of
the FM detector chip at 45MHz.

This soluti