Hi I have designed a BPF filter with center frequency of 140MHz and 30MHz BW. My filter is 5-th butterworth with lumped Ls, Cs and trimmers. Unfortunately the insertion loss is about 9dB. Is there any way to reduce insertion loss or i should change my configuration?
Maybe try a different type of filter, chebyschev, would give better insertion loss.
As Don said, the inductors need to be fairly high-Q. Coilcraft models are available on their website, and these typically give Qs of about 50 for that frequency range.
Is the filter flat across the passband? Is it a constant (or near enough) 9dB insertion loss?
Almost certainly you need better inductors - this isn't a particularly high Q filter, and the loss should be better. How are your inductors constructed?
Can you post either an ASCII schematic here, or a real schematic on news:alt.binaries.schematics.electronics? It might help with really looking at what you've got so far.
Not that I know doodley-squat about filters, but for some reason,
9 dB insertion loss for a 5th order filter sounds almost normal. But upping the Qs, even I know would help. :-) The closest I've ever been to designing a filter was a couple of second- or maybe fourth-order BPFs in the audio range, lifted directly from The Active Filter Cookbook, and, of course, they had a gain element in the middle.
And presumably it makes a difference if they're simulated or on an actual breaboard or even prototype. I've never run a simulator, but Real Designers don't use simulators. ;-D
Thanks for your answers. The inductors i have used are from murata and their Q is about 30. Simulations show that insertion loss should be about 5dB, but the implemented one is 9dB. My questions now are:
1- Is 4dB diference with simulation acceptable?
2- Is there any other structure which i can used to increase performance?
if your simulator allows s-parameter data to be entered into your circuit, then you can go to coilcraft website, download their data and enter it in. this will give you a much better idea of the inductors performance. Q of about 30 is quite low, you can get Qs of about 60-80 from some manufacturers.
what is your filter specification? rejection cut-offs?, attenuation level, passband ripple, return loss match?
I've been designing, building and testing all kinds of filters for the past year, and you can certainly improve your performance.
Hi Thanks for your responses. The Q specified for filters is in 100MHz frequency and murata coils cyurve that Q for this kind of coil is more than 30 at 140MHz. I think if there is a way that i can upload filter schematics, we will can discuss about this topic precisly.
No, the 4dB difference is not acceptable. Is this for a one off, or a production unit? If it is a one off, I would suggest you wind your own coils with thick wire on large, low loss formers - you should be getting much better than 30 as a Q figure. I have built similar filters with insertion losses less than 1dB this way.
At what frequency do Murata quote this Q figure? If it is far from the frequency you are using, that could be where the discrepancy lies.
- you have to put up with their ads, of course. Or there might be some kind soul who would let you email the schematics to them, assuming that you have a graphic.
Or you could draw it in ASCII - browse the group and you'll be able to see some examples of how it's done. Like '-' is a horizontal wire, '|' is a vertical wire, [coil] is a coil, [cap] is a cap, and so on. There's also a program you can download, I think maybe this one:
If you go to Coilcraft, you will find coils with a Q of about 130 - which will make a vastly better filter for you. 30 is frankly a bit soggy. Are these chip coils, by any chance?
Can you export the schematic as an IFF file? I can upload them and see what's going on, if you email to snipped-for-privacy@yahoo.com (remove the NOSPAM).
Have look at the coilcraft website
formatting link
you can also order some free samples of any value and series you wish, also contains data such as floor plans.
chip inductors usually have low Q but are quite small, the high-Q chip inductors they do reach aboutr 60-70, but the air-core inductors go around 100.
From 100MHz or up is the range where discrete components are no longer what they appeared to be. Start doing some measurements what values the parts really have.
And Motorola made the 68488 chip, which allowed a simple microprocessor to become IEEE488 capable.
I first used it for controlling a Fluke high voltage power supply from a computer, putting it between the Fluke's own wide interface and the computer which needed to control the power supply (a HP 9825 IIRC).
Later, I designed a controller for Diablo daisywheel printers to allow some friends to connect them to their Commodore PETs, using the
68488 and a 6802 as the "smarts".
Enjoy, DoN.
--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.