Hi Ken,
As far as I know, filters have been designed for 50 Ohms at both ports. At
the beginning, I terminated the filter with the help of two 100 Ohm
resistors (with the combiner in place). After finding a bad VSWR response,
I opted for testing the filter in isolation (which is easily achieved
soldering a 50 Ohm resistor in one end and a short piece of coax for the
antenna analyzer in the other end), getting almost the same (bad) results.
You get an slightly different result when testing the filter in isolation,
as the parasitic effects of the PCB tracks (the common RF tracks where the
relays are connected) as well as the broadband combiner are no longer
there. By "slightly different results" I mean reading VSWR of 2.5:1 instead
of 2.6:1, for example. Interestingly, at some frequencies the parasitic
effects of the PCB tracks "help a bit" while in other frequencies "make
things a bit worse".
Anyway: 160 m and 80 m filters had the right shape (the cut-off frequency
of the filter was about right) but VSWR was bad. I think that the lists at
contesting.com do not allow for attachments, so I'll send you an email
directly with some captures and photos.
So after the modifications, I have all seven filters behaving beautifully
when terminated with 50 Ohms (directly soldering a single 50 Ohms load or
by using the relays and terminating the two amplifier ports with a 100 Ohm
resistor each).
However, I still don't know if that response was there to compensate
something at the RF decks. Maybe during the development somebody found
that, at frequencies lower than 40 MHz, the filter's input impedance had to
be increasingly capacitive to maximize power transfer into a high-power
dummy load... I don't now.
So what I'm going to do next (time permitting) is to build a directional
coupler for 100 Ohms, a 100 Ohm load (there are inexpensive RF loads on
ebay, they just need a heatsink) and then, with a manual tuner in between
the directional coupler and the load, move the knobs around to see what
settings allow for maximum RF voltage developed in the 100 Ohms load. That
should clarify (I hope) if what I observed in the 160 m and 80 m was a
design issue or a design feature.
The last thing I want is to modify the filters thinking that I'm improving
the life of the transistors when in reality I'm making things worse.
I'll send you that email in a minute, it is a bit late here in VK.
73,
Carlos VK1EA
On Mon, May 10, 2021 at 12:21 AM Ken Brown <kenradiobrown@gmail.com> wrote:
> Hi Carlos,
>
> Thanks for the interesting information. I'm curious about the 160 and 80
> meter filters. If their bandpasses were not at the right frequencies, where
> were they before you changed the capacitor values? And what was the VSWR at
> 100 Ohms input impedance and 50 Ohms output at the frequencies they
> previously worked best at?
>
> I have a Hercules II and I like it a lot. Mine was not performing up to par
> and I replaced all of the PA transistors. That brought the power back up to
> specifications. Now after reading your post I wonder if it would work even
> better and longer if I made the filter modifications you have made.
>
> By the way, there are some very small PCB traces on the board that has the
> band switch on it, and one of those traces needed to be repaired on mine. I
> don't recall which band would not work because of that open PCB trace.
>
> Thanks, Ken N6KB
>
> On Sat, May 8, 2021 at 10:59 PM Carlos PECO BERROCAL <
> carlos.peco@gmail.com>
> wrote:
>
> > Hi all,
> >
> > Someone from the list kindly offered me a "tech special" Hercules II a
> > couple of years ago. I finally found the time to start working on it.
> >
> > This amplifier has two separate RF decks with an output impedance of 100
> > Ohms each (not 50) that get combined into a single 50 Ohm line and then
> > filtered. To test the LPF board (in isolation), I placed two small 100
> Ohm
> > resistors in lieu of the amplifiers, and connected an antenna analyzer at
> > the output port. Then, with a small 12 VDC power supply, I was energizing
> > the relays corresponding to the filter under test.
> >
> > I found that the 20 m filter presented a VSWR of 7:1 (that was probably
> the
> > cause of the amplifier's failure in the first place !) and traced the
> root
> > cause to the rivets that connect the top and the bottom PCB ground
> tracks.
> > So I renewed those connections (and added some more) and the filter
> showed
> > a VSWR better than 1.3:1
> >
> > HOWEVER, the 160 m and 80 m filters still showed a really bad response,
> > with a VSWR around 2.5:1
> > This is where the mystery begins:
> > - I grabbed the schematics and simulated the circuits on SimSmith. For
> the
> > inductors, I just measured the core size and the number of turns, that
> gave
> > me an initial ballpark figure to play with (that ended being a very
> > accurate estimate, by the way).
> > - Leaving the capacitor values fixed, no amount of inductor tweaking on
> > SimSmith fixed the off frequency response.
> > - I ended up unsoldering ALL the 160 m and 80 m components, measuring
> them,
> > and simulating the filter with its actual values on SimSmith. It matched
> > perfectly my measurements with the antenna analyzer.
> >
> > So,
> >
> > How is that five of the seven filters are perfectly fine, while the two
> > bottom ones (160 m and 80 m) present a VSWR of 2.5:1 at the ham
> frequencies
> > ?
> >
> > I could not restrain myself and ended up modifying the capacitor values
> > (while keeping the inductors constant) and now I have filters presenting
> a
> > VSWR better than 1:3:1 on both 160 m and 80 m. However... I'm a bit
> > curious about this. I wonder if there is something going on at
> frequencies
> > under 4 MHz with the RF pallets (like departing a lot from 100 Ohms) and
> > the designer decided to compensate those issues in the LPF block ???
> > The broadband power combiner is actually broadband... instead of two 100
> > Ohm resistors I also used a single 50 Ohm resistor connected to one
> filter
> > port and the antenna analyzer on the other.
> >
> > If someone wants to simulate the filters, the values (not present in the
> > schematics) are:
> > L1: 4.8 uH
> > L2: 2.46 uH
> > L3: 2.3 uH
> > L4: 1.5 uH
> >
> > Sorry for the long post. In any case, it is worth looking at the ground
> > rivets in the LPF filter block, as the built-in VSWR detector is located
> > AFTER the filter, unaware of what is going on at the input.
> >
> >
> > 73,
> > Carlos VK1EA
> > _______________________________________________
> > TenTec mailing list
> > TenTec@contesting.com
> > http://lists.contesting.com/mailman/listinfo/tentec
> >
> _______________________________________________
> TenTec mailing list
> TenTec@contesting.com
> http://lists.contesting.com/mailman/listinfo/tentec
>
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|