Filter Design Software

THS3201 ...Jim Thompson

I find,

Tim

-- Seven Transistor Labs Electrical Engineering Consultation Website:

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Or 200kHz, to ease the requirements on the coils even more at the expense of worrying about images. (There were a number of amateur and military radios in the 1950's that would convert down to 455, then again down to

Collins Radio used to make an impressive array of mechanical filters for

Den onsdag den 28. januar 2015 kl. 03.19.19 UTC+1 skrev Tim Williams:

then all you need is eight hands and the patience of a saint to trim it

-Lasse

That's one of our favorite opamps. TI makes nice fast stuff, relatively high voltage.

How are you going to nail those frequencies to ballpark 0.1%?

A couple of IQ mixer chips would allow a zero-IF structure. Mix down to baseband, lowpass filter the I and Q, and mix back up. That would be a no-trims design. Stopband attenuation might be a problem, but there are hacks for that.

I'm not the biggest fan of direct conversion, because all the spurs land right on top of you. With some reasonable IF, you can filter out some of them.

Of course if the phase response is more important than amplitude linearity, the performance of RC lowpasses is pretty attractive. (I use them for noise measurements a lot.)

Cheers

Phil Hobbs

It seems hard to me too. How do you tune such a beast? But I would guess Jim is doing it with opamps, and caps. If the cap ratio tracked with temperature...?

George H.

Bingo! (I guess you'll have to take my word that I didn't read ahead.)

George H.

Further head scratching suggests that an injection locked loop may be the easier-to-implement solution. ...Jim Thompson

Are you doing clock recovery? Why not a PLL?

Disappearing carrier, although I have some ways around that problem, but may or may not work in this situation. ...Jim Thompson

Den torsdag den 29. januar 2015 kl. 01.42.34 UTC+1 skrev John Larkin:

if it is just for data recovery and at that low frequency just oversample the data and a simple state machine to extract

-Lasse

Off-the-wall proprietary, so I can't say much. ...Jim Thompson

OK, I don't know injection locking (IL)... maybe you (and James A. ?) will share? I guess I see it (IL) as pulling a broader resonance to one side or the other.

George H. (opamps are going to be hard because of GBW variation, I think...)

THS3201 ?? ...Jim Thompson

Yup. If JT gets that working in a production circuit I'll send him a bottle of Scotch.

Yeah I saw that before.. 2 GHz. You certainly have more active filter experience than me. My only hard data points are from a 100 kHz, Q = .707, filter I made. It is used to do ~1% noise measurements and I measure it to the ~0.1% level. It's made with a 8MHz GBW opamp and there is some Q-enhancement. I see variations in the Q at the ~0.2% level (A quick survey of the last 10 instruments.) I'm not sure where the variation comes from, but I always wondered about GBW variation in the opmaps. You never see that spec'ed.

George H.

Could you use XOR phase detector and an integrating loop filter? That makes the lock very tolerant of dropouts, but maybe you can't afford a big enough capacitor.

Cheers

Phil Hobbs