XPost: rec.radio.amateur.moderated

Ham radio ON7IR

///////////////////////////////////////////
Selectable HF low pass filter

Posted: 13 Apr 2022 06:39 AM PDT https://on7ir.blogspot.com/2022/04/selectable-hf-low-pass-filter.html




A while ago I ordered a selectable HF low pass filter on AliExpress. This filter board contains 4 relais to select the desired bandfilter: 80m ,
40m , 20-17m and 15-10m. The relais need 12V to operate but basicly you can replace them by manual switches if you don't want to use a power supply.
The board can be found at many chinese webstores, some of them provide the circuit on their page but it is hard to read the values of the components.
You can also buy the lowpass filter as a kit, maybe then the value of the components is mentioned, I don't know. Anyway I did some investigation and found more information about the values so read further. But first
something about my build.
The hard to read values on the drawing found on seller's webpage.

First I checked the low pass filter board. There was a bad solder joint at
the 2-pins connector (J3 the power connection). Not a surpise with cheap chinese stuff! No other bad joints were visually found.
The low pass filter board as found on various webstores
Bad solder joint on my pcb


Because I didn't have a small 4-way switch available a simple button and a CD4017 are used to switch between the bands by just a push on the button.
First I made and tested below circuit on a breadboard. It works on 12V but
also works reliable at 5V. Afterwards all components were soldered on an experimental pcb.


The circuit.
Each time the clock input of the 4017 goes high the active Q port goes low
and the next Q port goes high. When the Reset pin is high then all Q ports
go low. A led is connected to each of the outputports Q0, Q1, Q2 and Q3 via
a current limiting resistor. One could connect port Q4 via the diode to the resetpin but then there is no option to not use the filter (passthru). So
after the next button push Q4 is high and there are no relais activated.
Next, when Q5 goes high the 4017 is being reset meaning that Q0 is high
again. This is the same situation at power on, Q0 is high.
I choose to feed the 4017 with 5V but the relais need 12V. So at each port Q0-Q3 a BC547 is connected which drives the corresponding relais on the filterboard.
The build.
In order not to change anything on the low pass filter board I added four
extra LED's and a pushbutton on a small separate piece of experimental pcb.
I have several external harddrives and through the years some of the
harddisks went to digital heaven. The metal enclosure is perfect to house
the low pass filter. One of them (a Toshiba hdd) has an external
powersupply delivering 5V and 12V. Inside the enclosure there is a small
pcb containing connectors for a harddisk, led, powersocket and USB
connector. Five holes were drilled in the frontpanel for the four LED's and
the pushbutton.
The USB connector has been removed to free up space for the BNC sockets
(also on a small pcb). Two pieces of RG316 connect the BNC's with the
lowpass filterboard.
Here some pictures from various stages of the build.Front and back of the harddisk enclosure

Electronic switch with 4017 and pcb with LED's and pushbutton,
ready to be mounted in the enclosure.
All wired up.
Final result.




The measurements.
I admit that I was too much focused on getting everything mounted and
working instead of finding out how the bandpass filter behaves. Using a
miniVNA the S21 measurements revealed not so good things...
80m.
At 3.5 MHz the filter already has a 1.5dB suppression and almost 3dB at the high end of the band. Also the curve is a bit strange. At 10MHz the filter performs less and less up to 12.5MHz.

40m.
Also bad. Suppression starts at 6MHz. At 7Mhz there is already more than
4dB loss.
20m and 17m.
Looks good. -3dB at 24.5MHz.


15m, 12m, 10m.
Loss is almost -1db at 28MHz and -2.17dB at the high end of the band, -3dB around 31MHz.
Except for 20m-17m all filters are working at incorrect frequencies! So
time get deeper into the design.
First I had to know the components values. The values indicated on the schematic that I found were difficult to read. The values of the capacitors
was easy to find, they are printed on them. The actual values of the
inductors can only be measured when one side of each toroid is detached
from the pcb.
So I searched and via the website of KC0WOX I found the original design of
the filter, described in QST March 1999 edition. The used capacitors have exactly the same value as the ones on the low pass filter board that I
bought. Also found this video about testing the low pass filter. The
results are much better than mine!
Next I compared the number of windings on each toroid with the values
mentioned in QST.
QST article Chinese pcb
80m T50-2 22 turns T50-2 22 turns
40m T50-2 16 turns T50-2 22 turns
20m T50-6 11 turns T50-6 10 turns
15m T50-6 8 turns T50-6 10 turns
It seems that our chinese friends just don't care and use the same number
of windings without knowing why. However that does not explain the strange
80m curve and I need to look further into this. The toroids might not be genuine or a capacitor is not working well. I have two T50-2 toroids
available, once bought from a seller on a ham flea market so first step
will be to replace the toroid for the 80m low pass filter. When done I'll describe my findings in a subsequent blog article.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)