4 Band Lowpass Filter Board

Board Design 

Ok, in the spirit of sharing and making things easy for people, I am going to start releasing some gerbers for the various boards I have designed for my various projects, that work HAHAHA, because, yes, not everything i do works.

I am not into this for making money, I do everything that i do simply for the love of it and for the learning experience. So selling boards and things is not high on my agenda. And besides, you can get 10 boards made and send to you from JLCpcb in China for not much more than it would cost me to send you one from Australia.

First cab off the ranks 7th order low pass filter boards. So you will notice there are no part values on the boards. Why? Well because i do not know what bands you might like to build for. The relays are double pole double throw in a standard 18x9mm package size. The transistors are anything you have in CBE footprint, I used 2n3904 and the resistor is there for current limiting, just pick a typical value, i think i used 4.8K.

SMA edge connectors bring RF in on the left side of the board. By default all relays are in the OFF position and no RF can pass through. The 6 pin header brings control signals and power to the board. To turn a relay on, a 5v control control signal is placed on one of the first 4 pins to turn 1 pair of relays on. The last 2 pins are ground and VCC of the relays you used.

For designing the filters I use an online calculator like this one at Calculators Edge: http://www.calculatoredge.com/electronics/bw%20pi%20low%20pass.htm Or a software tool like Elsie.

Lowpass Filter Board

Gerber Files can be downloaded from HERE:  http://robs-blog.net/Files/LowPass_Filter.zip


18 thoughts on “4 Band Lowpass Filter Board

  1. Hi rob and thanks for sharing the pcb!
    Wich type of toroid have you used?

    Thanks and 73’s de Marco IW5EJM

  2. Hi Rob, I designed a comparable pcb for 7 pole bandpass filter. When measuring the filters on a spectrum analyser I got good results but when I combined all on a pcb I lost 20db (the from the -50db). Some how the input signal is radiating to the output. If I by pass the relays and measure direct on the filter output it was ok. Researching this issue one flaw I found was that the ground plain is not continues I also see this on your boards. Did you make any measurements? DId you have any experience with this effect? Thanks Paul

    1. Hi Paul. I did not experience the issues you did. In my final design all my relays defaulted to the off position and grounded. By grounding them in the off position any radiated RF is shunted to ground and not passed through the filters. I also used a lot of via stitching and had via fences between each filter for isolation.

      I have got some new boards being made with are single band LPF. I should have them monday and I can build one and do some tests with the VNA and compare the filters in the switchable bank and report back to you the results.

      Do you have a picture of your board you can link to, so i can take a look at what you did and what might help to fix your issue.

  3. Hi Rob, I found the issue in my board. The groundplane was not continues. There was a loop on the output connection. If I closed the loop. The problem was gone. Your right you need good grounding and use a lot of via stitching.

    1. I try and leave the bottom copper layer as unbroken as possible to give the greatest area of unbroken ground as possible, with both the top and bottom flood fills connected to the ground net and then stitched together with vias. Anywhere there are long traces I try and fence those off also with vias. The amp board i respun that is on order atm has 380 vias on it. I might go a little over board on the vias but hey more is better right LOL

    1. No its probably good for about 50W at the most. You would need to go up a size in toriods and use 1Kw rated ceramic caps and test that the relays can handle 100w. Not sure if T68 toriods would fit on the board. I do not have any here to try.

      1. Hey Rob,
        Thanks for answering my question, actually is 50 watt more than enough. I was actually planning the T68 to use instead of the T50, hopefully it will fit! Anyway, thank for sharing all this beautiful stuff you made!
        73 Renaat

        1. It should fit. I have a couple of T68 cores here, I will print out the PCB later on paper and place the cores on to see, but i think they should be fine.

  4. Hello Rob,
    I ordered 10 LPF pcb’s, made by pcbway.com, total costs € 13 including shipping, and received today!
    Thanks for sharing the Gerber files, T68 ring cores won’t work, it’s too tight, but I have already ordered T50 ring cores. The pcb’s are very good quality 🙂 thanks.

  5. Hi Rob…
    I assume you are using C0G ceramics for this. Have you tried using SMDs? I thing 1201 size might just fit.
    A lot cheaper and more available too.
    Also, so you have any known good cap/ind values for common HF bands? I am trying to design filters but am finding it hard tp get the degree of attenuation I want at the 2nd harmonic while keeping a flat response and a the attenuation not bleeding into the band(s) of interest.
    / Gerry

    1. Hey Gerry,

      Everything I am doing now days is SMD, in low pass filters I am using 2KV rated Ceramic caps in 1206 packages mostly X7R not C0G/NP0. Way cheaper than using exotics like mica or even higher voltage NP0. As ceramic caps reach 50% and higher of their voltage rating, they start to change capacitance, and so you need to derate them accordingly. So a 10w signal is 30V peak, this would make non NP0 50V ceramic caps marginal.

      As for design tables I do not use any, I use online calculators for designing low pass filters at places like calculators edge, or software like Elsie or RFsim99. Not getting enough attenuation of the 2nd harmonic might not actually be a filter problem, it could be a distortion problem in the amp, or some preceding stage. Better to eliminate as much harmonic content by elimination distortion than it is trying to filter it out at the end. A 7th order filter really should be overkill in all but the most extreme of situations.

  6. Hi,

    Thanks for sharing your designs. I got 5 boards made by JLCPCB before I realized that the gerbers posted don’t match the picture in your blog or the description 🙂 My bad. It seems the posted schematic does match (?)

    I have a couple of questions about the gerbers/schematic:
    – are the capacitors 1206?
    – what transistor are you using to control the relays? The schematic seems to have some text but the jpeg is not good enough for me to decipher. It looks like DTCT14 but I’m not sure and can’t find any info on that. I could bodge something with a 3904 but I’d rather not.

    1. Hi Dave,

      Caps are 1206 and the transistor is a DTC114 SOT23 package. You could bodge any old SOT23 in but it will be lacking the bias resistor thats built into the DTC114. Should not make much of a difference really, its just a switch to turn on the relays with a 5V logic level to the base.

      1. Thanks Rob. That helps.

        I have some SOT23 3904s I’ll try.
        Looks like Omron G6S relays will work. I’m going with 5V relays but I don’t see why that would be a big issue.
        And I guess any silicon diode e.g.1N4148-ish should?

        1. Yeah any old diode should work, ultimately the diode should be able to handle a couple thousand volts generated from the field collapse from the coil in the relay, but its not really critical, the part might just have a shorter life, which in terms of homebrewing will be orders of magnitude longer than the project in use anyway. HAHAHAHA, pretty sure the pinout on the 3904 is the same as the btc114 BCE so that will work with a shorter part life because of no bias resistor. not a big deal.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.