[Amps] Making a broadband transformer

[Mike Waters]

Bill,

I know very little about what it takes to build a solid-state amp, but I
have an experience to share that may be of help to you.

I have a broadband matching transformer on the input of my swamped-grid
dual-833C triode amp made from 3 stacked small binocular cores made of 73
material. It handles 100 watts continuous without hardly getting warm on
160m, and I've run 800 watts through similar transformers on 160 through 20
using heavily processed SSB (and possibly also CW) without exceeding the
Curie temperature (310 F). (I don't remember if tried it at higher
frequencies, but I think I did).

The actual grid input transformer in there now has three cores. This one
shows a two-core transformer that worked fine in the same place.
http://www.w0btu.com/files/misc/833C_linear_amplifier/
http://www.w0btu.com/files/misc/833C_linear_amplifier/100_4944.JPG

Here's the test photos --power vs. temperature-- of a similar transformer:
http://www.w0btu.com/files/misc/Binocular%20core%20power%20vs%20temp%20test/

Now, that material in that application should be good from 160 through 15
meters, after which the reactance goes up somewhat. For receiving (our
usual application) that's fine, but you'll have to experiment with it on 12
and 10 meters in your amp.

Fair-Rite P/N is 2873000202 (Amidon BN-73-202)

Trouble is, I don't know if Fair-Rite makes in the size and shape you might
need. But if they do, I'll bet #73 might work for you.

73, Mike
www.w0btu.com


---------- Forwarded message ----------
> From: Bill Turner <dezrat1242 at yahoo.com>
> To: Amps <amps at contesting.com>
> Cc:
> Date: Sun, 03 Nov 2013 18:15:09 -0800
> Subject: [Amps] Making a broadband transformer
> I have been doing some bench testing of broadband toroid transformers with
> the intention of using them in a legal-limit solid state amp. I don't have
> a
> "real" impedance analyzer, but I find the MFJ 259B SWR analyzer gives me
> readings for both resistance and reactance which seem to be reasonably
> accurate and make sense.
>
> I am finding that creating a transformer that will cover 160-10 meters is
> not an easy thing. No surprise I suppose.  :-)
>
> My best result so far has been with a pair of FT-280-43 toroids stacked on
> top of each other (not binocular). For test purposes I am keeping the turns
> ratio 1:1 to make it easy with the MFJ analyzer. I have a 51 ohm non
> inductive resistor across one winding and the MFJ connected to the other.
>
> I have tried many different numbers of turns using a twisted pair and so
> far
> just four turns is working pretty well on the lower bands but the leakage
> reactance starts to become excessive above about 11 MHz. The resistive
> component only varies between 50 and 60 ohms on 160 through 10, which is
> acceptable, but by the time I get up to 28 MHz the leakage reactance is up
> to about 80 ohms.
>
> So I would like some advice on what to do. I have read all the books and
> papers I can find so I have a pretty good grasp of the basics, but there
> must be some tips and tricks I have yet to learn.
>
> One question I have not seed addressed anywhere is the use of multiple
> mixes
> in the toroids. For example, a 43 mis and a 61 mix stacked. Has anyone
> tried
> this? It would seem that would help extending the bandwidth. I have some 61
> mix toroids on order so I will be finding out myself pretty soon, but I
> would like to hear from anyone who has tried it.
>
> One thing I am really curious about is how the commercial amp makers do it.
> Have they found the secret to reducing leakage reactance, or do they
> compensate for it somehow? Or just ignore it and power on through? I see
> the
> Ameritron ALS-1300 has a 470 pF capacitor from drain to ground on each of
> the transistor pairs. I don't see any other compensation for the
> transformers. Interesting. That's the only schematic I have for a
> commercial
> amp. I would love to know how Icom, Tokyo Hi-Power and SPE do it.
>
> Just out of curiosity, I found it is quite easy to tune out the reactance
> with a capacitor, but then of course, the transformer is no longer
> broadbanded, but it does give a neared perfect resistive match.
>
> All comments welcome.
>
> 73, Bill W6WRT
>