Low Pass Filter Designing

[Nikos A.]

Hi everyone,

I have a 10Mhz OCXO and I want to design a low pass filter to apply at its output. The reason for that is that besides the foundamental frequency (10Mhz) I can measure harmonics as well. I've never designed a filter before... Looking some tutorials on the web I calculate the RC values for a low pass filter with cutoff frequency of 15Mhz.

I tested the circuit on the breadboard and the results was frustratingly... The filter actually doesn't work at all..

Is ther any reliable free software/tool for filter desinging? Should I use LC instead of RC filter?

Thanks in advance

[mikerj]

The performance of a passive filter is very dependent on the source and load impedance, did you take these into account?  Additionally a single pole filter only rolls off at 6dB/octave, so a third harmonic at 30MHz would only be 6dB down even if your filter was working correctly.

[srce]

What values for the R & C did you use?

Useful filter design tool here: http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

Along with some other types: http://sim.okawa-denshi.jp/en/Fkeisan.htm

[MarkF]

And here is a LC filter design program with typical output:

https://www-users.cs.york.ac.uk/~fisher/lcfilter/

[JagV12]

This LC filter adds the rejection of any chosen frequency. You can cascade several to reject various harmonics. It is often used as output filter on RF transmitters. R1 & R2 are not actual components, they symbolize source and termination impedance. Of course, L & C components value has to be calculated to match your needs...

[Kleinstein]

The output filters of DDS generators are also an area to look for, though it likely would not need such a steep filter here.
They usually use elliptic filters, so not just a low pass, but also additional extra notch functions in the stop band.
The filter JaV12 showed is actually a simple elliptic filter - a pure low pass would be without C3.

[blackdog]

Hi Nikos A.

Maybe you can get some inspiration from a topic of mine on a Dutch electronics website about filters that I made around 10MHz.
For example, use Googel translate to read the text.
https://www.circuitsonline.net/forum/view/119653/1#highlight=pass+low%2010mhz

Topic number two.
https://www.circuitsonline.net/forum/view/119525/3#highlight=pass+low%2010mhz

I hope its helps.

Kind regarts,
Bram

[awallin]

And here is a LC filter design program with typical output:
https://www-users.cs.york.ac.uk/~fisher/lcfilter/

if that means pi/2 phase-shift at 10 MHz from the filter then you'd better make sure that phase-shift is stable over temperature to whatever spec you are trying to achieve.
the tempcos of typical LC components might be orders of magnitude worse than the tempco of the OCXO...

[JagV12]

Even a pi phase-shift over a (quick) 1 minute temperature rise is only half a wave more (or less) than normal. At 10MHz, that's .5/(10E6*60s) or 1/1.2E9% drift or 10.0000000083MHz or 0.0083Hz more (or less) than normal.

And that drift only occurs during warm-up. Or did I miss something ?