Lorick, I've got `egg all over my face' and am hearby `chastised'... My only `lame' excuse is that I have not had time to really look at it. I'm working from an old classic; "Rapid Practical Designs of Active Filters" by David Johnson and John Hilburn. I had the Op-Amp polarity reversed and confused a VCVS filter with a MFB filter... Correcting the above and adjusting the Bode plot shows a `band-pass' response around 300 Hz. I did a sweep from 10Hz to 1KHz and reduced the input level to your spec. My `public disgrace' aside, that still does not address your problem... Now that you have me `hooked' I want to `nail' this down... I had originally thought you were only getting around 600mv from the filter chain? From your latest message, it appears to be working? With these frequencies and signal levels, I would not expect much difference between Op-Amps. What is the current status? Have you looked at the switched capacitor alternative? Again, it's been a long time since I've worked with Sallen and Key filters.. As far as loading the output, this will vary depending on the Op-Amp specs and the load impedance which includes capacitance. Finally, I would back off those Motorola Op-Amps. There is a `wealth' of options from National, TI, Maxim, Analog Devices, and Linear Technology. All of which seem to be committed to their products for the `long term'. I would request their CDs if you don't have them, and check their web sites. Most provide parametric searches both on the web and the CD. - Tom At 02:02 PM 3/9/00 -0500, Lorick wrote: >From: Tom Handley > >> site to check out the circuit and the first obvious problem is that filter >> is NOT a `band-pass' filter, but a `band-reject' filter!... >Are you sure? The book I followed says it's a band pass configuration and >the bode plot shows clearly a small pass region just where I calculated it, >with continuous attenuation outside of that, and a small notch hump at some >frequency higher than my pass region which looks like a typical chebyshev to >me, and the lab test validated that I obtained my highest output in my pass >band frequencies and attenuated from there. >I re-ran my spice simulation with motorola op amps (they seem to not load >down the output like the TI ones did, thus giving me the calculated output >levels). I used the MC33172 model from Electronics Workbench and then >exported the spice file to run in Orcad, which I prefer for spice over >workbench scopes and bode plotters....and with a 1v peak input sine, I got >about 25mv out at 300Hz, which if I calculate properly, is around 32dB >attenuation, which looks relatively like a flat line but it's within my >design specs for having at least 20dB attenuation below about 310Hz. Then >I ran the simulation at that 310Hz corner and calculated almost 19dB >attenuation which is close to design spec. > >> Also, there is >> no need to have three pots for offset. Just use one and a voltage >follower. >Yeah I've already reduced those pots down to the one reference. I mostly >put the 3 in there for the purpose of clarity to indicate I'm biasing all 3. >I've revised that schematic many times. > >> Out of curiosity, I ran a SPICE simulation with a Bode plot on the >first >> stage of the filter. >My tests were the entire 3 stages and I got the same type of frequency >response no matter what op amp I chose. My only issue was they attenuated >differently. Someone suggested the problem with certain op amps was loading >on the output, including the feedback loops, and so I'm placing another >order today to get those motorola ones that worked well. If not I'll just >try whatever I have laying around until something works. > >I've added the spice file for the whole filter with the motorola op amps on >that page. The direct link is >http://rivendell.fortunecity.com/perilith/831/filter.cir >if you wanted to look at it again. I'm sure it's a band pass. > >> I would say, in general, if you go beyond a 2'nd order filter, you should >> look at an alternative to Op-Amps unless you have spares in a given >package >> and/or find a price advantage. >If I could do this over again I would have a completely different approach >than I ended up with but right now time is what is against me and I have a >lot of calculations and demonstrations I can "show" for what I've done so >it's adequate for my purposes for now. >If only I could find time to work on the pic code again. I'm always >seemingly too busy to do PIC things... I guess I'm doing my usual habit of >underestimating what's involved with gluing code together to have a >successful program based on individually working math routines and port >accessing. ------------------------------------------------------------------------ Tom Handley New Age Communications Since '75 before "New Age" and no one around here is waiting for UFOs ;-)