QRP-Labs broadcast band filter

wip_c2

 

Inspired by some fine projects in the SpratMagazine (the journal of the GQRP Club), just for fun, I build a WBR 40m regen receiver as described in the QST from August 2001One thing the receiver had problems with is that it was very sensitive to overdrive of a local AM radio station. Initially I added this high pass filter and it immediately fixed the overdrive problem.  

A while ago I found this page on the VK3IL Blog on which he describes his approach on broadcast interference reduction with  a highpass filter that could handle 100W without a problem. 

Now I'm a QRP guy so most of the time the max power I use will be less then 10W. So using T130 torroids will be a little overkill. Also I very content with the QRP-Labs LPF and BPF's and use these in a lot of my projects. So why not build a broadcast band filter which is pin-compatible with the QRP-Labs filters and can handle QRP power levels ?  

 

The schematic 

 I haven't used simulation software in years (except of a free account of CircuitLabs) and knew that in KiCad there is a simulation tool based on LTSpice. So for this project I wanted to get myself familiar with this option. So I took the VK3IL schematic and drawn it in KiCad and started to dig-in how to use the simulator.

bcfilter_08

TODO : howto configure the LTSPice components (with screen shots etc) 

The simulation result

bcfilter_02

 

For simulation to prototype

Theory is a nice thing, but practice may not be the same.  So I build the filter Manhattan style on a piece of PCB with the same header placements like the QRP-Labs filters. 

bcfilter_03

 I glued 3 Manhattan style soldering pads on the PCB to solder the components to. 

bcfilter_04

 Capacitors in place. 

bcfilter_05

 Torroids added. 

bcfilter_06

Time for a cup of thee. 

Simulation vs real-life.

Because I made it the same size as the QRP-Labs filters, I could use my QRP-Labs filter adapter for the NanoVNA to test the filter. 

bcfilter_07

I used to measure the filter with my NanoVNA and below the result is shown. I'm happy with the result. 

bcfilter_09

 

New PCB

From the schematic above I designed a PCB which could fit into the QRP-Labs filter socket.

bcfilter_pcb

I ordered a batch and when the PCB's arrived I collected the parts and documented the build process.

When all works out well, the left over PCB's will be sold bare, as a kit or on request build and tested (see note below). 

bcfilter_10

Lets start building

I used my QRP-Labs filter adapter for NanoVNA as a soldering jig. 

Step 1 : place the 4 pin headers in the 4 pin female headers.

bcfilter_11

 Step 2 : place the PCB on the headers (OUT = left, IN = right) as shown below.

bcfilter_12

Step 3 : solder both 4 pin headers

bcfilter_13

 Step 4 : place the capacitors (C1, C4 = 1.5nF, C2, C3 = 560pF) and solder them. 

bcfilter_14

Step 5 : While holding the components take the PCB out of the adapter and bend the lead a little so they don't drop out.

Place the PCB up-side down on the adapter and solder the lead. When done, cut the leads off. 

bcfilter_14b

Step 6 : Wind the toroid. Cut 3 pieces of the supplied laker wire to a length aprx 36 cm of 14 inch.

Wind the toroid L1 with 23 turns (remember the wire through the hole is one turn).

Wind the toroid L2 with 21 turns 

Wind the toroid L3 with 23 turns .

Tip : what I always do (when possible) is group the wire in bundles of 5. This makes the counting easier. 

bcfilter_15

Step 7 : Remove the enamel from a enamel wire (this is how I do that) and pre-tin them. That way you know that you have removed the enamel from the wire correctly. 

Step 8 : Place the toroids on the PCB and twist the wires (tight but not to tight). 

bcfilter_16

Step 9 : again place the PCB up-side down on the adapter and solder the leads. When done, cut the leads off. 

(sorry no photo)

Step 10 : place filter on the adapter, sit down, relax and enjoy the fruits of you labor. 

bcfilter_18

Lets taste the pudding aka get some measurements done.

Step 1 : Calibrate your NanoVNA for a span of 100KHz to 10 MHz. 

Step 2 : Place the filter in the adapter and watch the results. When everything is well, you should see a S21 LOGMAG as shown below (the light blue one).

bcfilter_19

bcfilter_20

bcfilter_21

bcfilter_22

Conclusion 

I placed it in my 40m WBR receiver and it works like a charm.

One thing I found out was that it would be better to use a capacitor foot print with a 0.2" of 5 mm distance between the leads. But a little lead bending fixes that. 

The story continues . . . 

Update 14 november 2022 

Based on my conclusion I re-designed the PCB and now the capacitors can handle 0.1" / 2.54 mm and 0.2" / 5 mm pin distance for the capacitors. 

bcfilter_v2_pcb

A small prototype batch of PCB's I ordered showed that it worked out. So now 2 types of PCB's are available. 

Update 27 november 2022 

When I went to my local electronics store I bought a hand full of traditional ceramic capacitors and found the out the reason why you should use CC4 type of capacitors (low loss).

bcfilter_31

Below are the same 4 plots as made from the original PCB but using the traditional ceramic disk capacitors.

Watch the S21 LOGMag values. Although the pass through isn't that dramatic, there is a big difference. 

bcfilter_23

bcfilter_24

bcfilter_25 

bcfilter_26

Same PCB but now with C2 and C3 instead of the clasic ceramic disk capacitors, modern CC4 capacitors.  

bcfilter_32

bcfilter_27

bcfilter_28

bcfilter_29

bcfilter_30

For easy reading I have put the 12 measurements (incl the prototype) into a table.

Frequency (MHz)

First prototype
C1&C2 CC4
C3&C4 Unkown

S21 Logmag dB

Second prototype
C1..C4 Trad

S21 Logmag (dB)

Second prototype
C1&C2 CC4
C3&C4 Trad 

S21 Logmag dB

1.21375 -62.39 -68.15 -67.24
1.8325 -36.66 -42.99 -37.42
3.49075 -0.37 -1.88 -0.36
9.967  -0.09 -0.42 -0.13

 

Conclusion 

Depending on your design, it might be interesting to use the more expensive low-loss CC4 capacitors above the traditional ceramic disk capacitors.

When you are building : 

  • a RX and use a small RF amp after the broadcast band filter, you can keep it cheap. And be fair, the difference of -1.52 dB isn't that much.
  • a QRP multiband RX/TX rig (excl the 160m band) its a good idea to use the more expensive CC4 and NP0 capacitors for C1 to C4.

Of course a dedicated BPF for the band you want to listen to is always the best option, but when you want more bands, you need more BPF's making your design more complex.

But playing around, doing experiments, check and compare results, re-design, and re-work. That's what's this great hobby is all about. 

  

Where to buy 

The left-over PCB's of this small batch will be sold through the For-Sale page on my website. When they are sold out and there is a demand, I might be ordering a new batch of PCB's to sell them, make new kits or even sell full assembled filters. So when they are sold out and your interested, let me know.   

The design of the broadcase band filter is publicly available (use the For-Sale page  to request GERBER files) so you order your own PCB's at OSHPark or a other PCB factory of your choice, but licensed it Attribution-NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0).
Meaning : 

  • Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
  • NonCommercial — You may not use the material for commercial purposes.
  • NoDerivatives — If you remix, transform, or build upon the material, you may not distribute the modified material.

It's not to be childish but it now happened a few times that a  designs I made was sold by other people just because is was "Open source, so I can do anything I want with it" without any credits or even a token of appreciation

When you like the design and PCB's. a small fee as a token of appreciation / commission to support my work is very appreciated (see : Support & Tip Jar).