Crystal Filter Perfection

Sometimes, just sometimes, all your ducks align in a row and you build something that is dead set on the money. I think today I have done just that with this crystal filter.

Firstly I have redesigned the PCB board and test jig setup I am using, making construction just a little bit easier and allows me to easily use the slightly larger binocular core.

But secondly, I think i hit the nail squarely on the head, with this filter. I mean its almost perfect. 5dB loss, great, good shape factor, great, -55db stop band, great, exactly 2200Hz passband PRICELESS. I honestly could not have done better even if I tried. Now to build the receiver board and pray to Jebus that it works as good as this filter looks.

As a bonus, here is his friend the CW filter, 600Hz wide. A little more lossy but still good stop band and shape. I was shooting for 500Hz, bit I need maybe another 15pF per node. Good enough for me though, that is 2 filters for a receiver, 2 filters that should perform reasonably well.

Final EDIT:

So I was going to upload the gerbers and create a great big post about how to use them. But, i went looking for the files and it looks like I have deleted them by mistake when cleaning up my project files. I got no idea how I did that, but they are gone and its a lot of work to recreate them, so there will be no release of these PCB files. Lucky for me I can reorder them easily and I also have enough board here to make 20 filters, that should be a life times supply for me. Rather sad really, because they work so well.



Before and After.


While not entirely a necessity, I upgraded the crystal oscillator that I use when building crystal filters. It does the same thing just looks a little more fancy. HAHAHA, when it only costs $2 to throw a simple board like this onto an existing order, why now hey? 😉


5 Band Filter Board

Just like a bought one hey? This has been a while in the making and brings together many of the smaller project boards into one. We have 5 low pass filters, band pass filters, TX/RX switching, and + and – 20dB RF preamp. Low and behold, it actually all works right out of the box, no major issues, other than winding all those toriods is a nightmare LOL

In testing the losses were acceptable, 1/2 a dB on the low pass and 1.5dB on the band pass. With essentially 10 filters on the board, it was going to be interesting to see how much work they would require to bring into the correct band.

Most of the filters were actually just fine as you can see here with the 80m low pass filter. Only 1 low pass filter needs to have a turn taken out of each toriod to bring the corner frequency up a little so the corner is not in band, and 3 of the band pass filters need the same to move them up in frequency also.

I did my best to account for strays and variation in winding, but there is always going to be some error and fixing, especially in the band pass filters where I am using a specific value of inductance in the calculations for the filter and then using a turns calculator to get that inductance. These things are always rather hit and miss. With the low pass filters you can always set the corner frequency higher than needed, like with this 80m filter, our band here is from 3.5Mhz to 3.7Mhz, and I set the corner frequency to be 4Mhz, because if you end up with strays you always end up with more inductance and capacitance, not less and those lower the frequency response.

Anyway, thats it for me, I have another 6 boards to build now and I seriously hope that the receiver board is actually good, because if that works well enough to use, I will have a very useable and workable 5 band 50w transceiver. Wish me luck.


My Trace Is The Fuse

So i was slowly creeping up on giving this all the amps. Moved it from my 10A current limited lab power supplies to the 40A monster I use to run the radios, which has no over current protection at all. This is not normally an issue, when things go from production to daily drive, they get fancy things like input fuse protection to protect the wiring from melt down if something goes dead short.

Now the amp, it works, works really good. With very little gate bias and very little drive, at 13.8V supply, I was getting 50W out and it was not even breaking a sweat or getting hot. So i figured it was time to up the ante and turn the gate bias up, slipped with the multimeter probe and shorted it out and gave it all the amps.

Like Tim the Toolman I gave it 40A of raw power LOL and the main power trace became a fuse and it gave up the ghost. Anyway, 50W is the most stable and usable power I have ever made and this thing will go much higher than that. With the right drive and bias combination i can see 75W+ plus out of it. But even at 50W it will run all day and not break a sweat. And mostly that is what i care about, Having something that will run all day and not be stressed out rather than something making all the watts but on the verge of wanting to take an eye out when a FET explodes.

Next version of the board will be fuse protected. Actually, all the boards I have sitting ready to be send off for manufacture are about to get fuse protection 🙂


40 Watts


I am pretty sure that this is going to hit the magic 100w, but i need to use a power supply capable of delivering more than 10A. I am current limiting and this is as far as she will go. I have not been doing a lot of late, busy with real life though I have been working on some new boards. They are sitting on my desktop just waiting to be sent off. Oh and I have built and tested a few older things I had here, they were a bust so I just binned them and did not post about it. Slowly working my way thought the backlog. This one was the most promising of them all.



Been having an interesting discussion about filtering for VHF and UHF when using the HackRF as a transmitter. Which got me thinking, how hard can it be. Well I now know why VHF and up filters have the copper pours pulled back and everything stitched to the bottom side ground plane. The strays on this are killing me. Its certainly usable, but its rubbish. Boards that work quite fine at HF are almost totally unusable at VHF. There is something new i can incorporate into future designs to improve things more generally. Better design is better design yeah?


More Protoboard Madness

I have spend the last couple of days writing code for this thing. I am finally making inroads and next job is to actually build the receiver IF board and test it out. This current iteration of the PCB is so much better to work with, much more user friendly. We currently have working audio amps, band switching, variable rf gain, band pass and low pass filters, TX/RX switching and the like all happening.