OK Mal Thanks for the info. It is reasuring. However I have just built the driver stage as per Rik's Transv. and get a nice square wave from very rippling input, into the gate an IRF540 (50P ea incl p&p China) single ended arangement looks ok on the scope at the drain. I have 14v from the PSU not measured current yet. I get 110VP-P across a 50 Ohm load I think thats about 30watt's. The master osc is a Square wave from a DDS..... So far so good. I have a gate type keyer as per Stefans arangement .... real mix and match eh?. Removed connector from dummy load and connected to coupling link on my tuner not checked the volts across this yet. Counter goes nuts....... think its overload then look at DDS display it is dim. The ant
current indicator is good, not calibrated but can tell there is RF current (0.7A) there (but on what QRG?). Connect back on to the dummy load every thing is OK again counter OK DDS OK. I normally use Xtal control in my other rigs either div /8 or Xtal mixer with no problem. So I am getting RF back into the DDS or its PSU when the ant is connected ...... sorry about tail of woe. Think I will use a rig as the master osc (better screened) and divide by four like Stefan's hopfully it will not get RF into it, pitty won't be direct read out. Story so far will let you know the ending soon!!! Oh hoping for just over 100Watts with 24v PSU (25% FET max rating). 73 es GL Pete M0FMT IO91UX
Pete
I find for a class E PA or class C with series resonant
cct at frequency of interest you do not need a LPF.
To resonate the PA use an Air core inductor and fixed
capicator ie for 500 Khz use 5nF in series with 20uH inductor. Adjust inductor
slightly up/down in value to resonate at exact freq of interest. Use an old LOPT
core 8 turns Prim 24 turns Sec. The wire does not have to be Litz. Prim 2.5mm
elec wire and Sec 1mm. I have used both litz and elec wire and find no
difference.
A couple of hours should build this type of
TX
g3kev
----- Original Message -----
Sent: Friday, June 29, 2012 11:31
AM
Subject: Re: LF: Re: RE: 472 kHz PA
running at 12 VDC
Hi Rik
Many thanks for your very useful information. I hope others find
this as well.
Yes I said in an earlier mail what the PA and driver would cost
from E-bay (my supplier of choice... :o)) but at least they are
obtainable.
The current hold are the Ferrites or the Iron Dust toriods.
This is the usual problem because one black iron ring looks very much like
another. Once you have it in your hands you can do a test. What I do is to put
a few turns on an unknown core and measure the inductance then compare the
results with the theoretical value obtained from a range of cores with the
same dimensions using "Mini Core" SW. At the moment I have a source for the T1 core but not the choke. And surprisingly
I can't find the T80-2s they were every where once upon a time. Anyway I do
have a variety of cores that will probably do for the O/P filtre. One day
I will find a design I can kit perfectly.
I like your transverter, I am getting torn between projects right
now. Q6 / Q7 driver as you say can be reproduced for each FET gate cheaply in
fact I have the parts!
I will now crawl into my cave and start to build. Thanks to you and
Stefan.
73 es GL Pete M0FMT IO91UX
I assume that the more exotic parts are the
power FET (IRFZ48N) and the driver IC
(ICL7667).
The IRFZ48N is a 55V / 64A / 0.014 Ohm
device. For a first try you could replace it by far
more common IRF460 that is 100V / 28A / 0.06
Ohm.
A bit of reverse
engineering: Stefan states 50W output at 80% efficiency = 62.5W input = 5.2A at 12V.
So with a 28A device you are on the safe side.
The larger Ron (0.06
Ohm vs 0.014 Ohm) will (in theory) reduce the efficiency. But 5.2A in 0.014 Ohm
is only about 0.4W loss (for a IRFZ48N). For a IRF460 it is 5.2A in 0.06 Ohm = 1.6W
loss.
So the losses (12.5W) are probably
more due to the non perfect switching than due to the Ron of the FETs.
Hi Stefan
I like this and would like to make to make it. I so
far have tracked down most bits but as usual the Unobtainium of so many construction projects
that fall at the last hurdle unless they are Kits........ no chance of that I
suppose?
There will be those who tell me its easy to get the
bits but unless you work for a company that uses this stuff it can not be
slipped through the "Purchasing" it becomes a positive stop!!
Like the Litz I just put that in one of our big
outfits with min orders and it said it had over one hundred thousand Litz and promptly
collapsed. But they have
one of the toroids (T1) wow and its not a min order
either double wow! Litz can be fabricated from 0.125
Cu/En? Hmm
it's plaited ?
Wima caps are available from Maplin haven't checked
in depth but they usually are.
I like the idea of using a DDS or Transceiver as a master osc source div. by
four.
Generally I buy my power FETs by the hand full
off E-bay they may not be ideal but are cheap!! 50P (eer 60Ec) (Never mind the
quality but feel the width... principle) and the drivers TC range for
slightly more.
I'll keep searching for a bit but if anyone has a kitting
list (UK source) may be they could publish it?
And thanks Stefan nothing under the Sun is "New" but in
context it is new and worth while thanks!
73 es GL Pete M0FMT IO91UX
Hi MF,
Late at night, i made draw the schematic. Here it is: https://dl.dropbox.com/u/19882028/MF/100W%20475kHz%20PA.png
(picture of prototype https://dl.dropbox.com/u/19882028/MF/DK7FC%20472%20kHz%20PA%2015W%2B100W.JPG) Most
of the parts are not critical. The ferrite cores are the blue EPCOS types, N30
material. Details are in the image. But you may use a FT140-77 alternatively
and double the primary and secondary number of turns. There are many
alternatives, also to the FETs. Instead of the IRZF48N you may try a
IRF1405. This
one has a lower on-resistance but a higher input capacity. Maybe it works even
better, i didn't try to many parts. Switching times are not critical. So you
need to find a compromise between low on resistance and input capacity while
the Drain-Source voltage should be 55V for a 12V supply. Instead of RF
litz wire (i
use this one: http://www.reichelt.de/index.html?ACTION="">
) you can also use 1mm diam. enamel. Cu wire or better 5x0.4mm or what you
have. It just affects the efficiency, more or less.
The overall
dimensions of this prototype are 15cm x 10cm x 3cm, so it is quite compact and
since it runs from a 12V battery you can use it for /p tests. Also you can of
course use more than 2 power steps by using a multi step switch and more taps.
A
DDS signal generator with rectangular wave form or a sine signal coming from a
HF TX may be used. The waveform doesn't matter as long as the peak input
voltage exceeds 5 Vp.
73, Stefan/DK7FC
PS: Yes, it
is nothing new and someone has done it before!
Am
28.06.2012 11:46, schrieb M0FMT:
Hi again Stefan
The Litz is too, only by the Km or 10Kg
reel.......... :@((
Do you create your own I have 0.125 solid C/E guess that could be
stranded up to make a synth?
73 es GL
Pete M0FMT IO91 UX
Hi Stefan
Just over £5 a pop!
Just over £6 a pop!
But what is the “Unobtainium” ferrite?
…… ;0)
Chees Pete FMT
Hello Stephan , Me too 73 , Jean-Pierre F1AFJ
Le 28/06/2012 10:36, M0FMT a
écrit :
Hi Stephan
Me Too
73 es GL Pete M0FMT IO91UX
like
-----
Original Message -----
Sent:
Thursday, June 28, 2012 6:35 AM
Subject:
LF: RE: 472 kHz PA running at 12 VDC
If someone is interested
i may convert the
circuit from my mind into a schematic as a png
file.
Always interested ;-)
73, Rik ON7YD -
OR7T
|
