Hello, i've recently got back into an old project of mine that i never finished, and thus started to gather parts and knowledge to finish it. The project includes a TPA3116 amplifier from ebay, and after hearing it again and testing with a func. generator, i wouldn't wish this low pass filter(or whatever the hell this is) to my worst enemy, basically it remains the same volume at higher frequencies as the tweeters, no difference at all (some low pass filter...) thus i have started to design my own to couple it with the circuit and remove the existing one. After a lot of research and revisiting old threads of mine here, i have come up with this sketch; i would apreciate if someone could give a quick scan for errors and any lack of dc paths, missing components, mistaken isolation or feedback problems.
Thank you all in advance.
Thank you all in advance.
An externally hosted image should be here but it was not working when we last tested it.
EDIT: noticed a mistake, the input of the inverting should be attached to the output of the sallen key not the input, here it is corrected and in multisim; in multisim simulation i get good waveforms, the only problem is that i have around 50mv of offset between outputs, is this a problem? will this offset be amplified as noise?
An externally hosted image should be here but it was not working when we last tested it.
If you want to design active filters there is no need to reinvent the wheel. Download TI's excellent Filterpro software and it will design full circuits with BOM's using real valued components.
WEBENCH® Filter Designer Tool - TI.com
Similarly, their TINA spice program is great too.
WEBENCH® Filter Designer Tool - TI.com
Similarly, their TINA spice program is great too.
To get lowest DC offset the resistances 'seen' at the + and - inputs need to be fairly closely matched.
Take your U2A as an example - here the + input sees 100k plus 50k from the divider network, total 150k. The - input sees 25k. You'd also do better to use just one divider network (halve the resistor values) as they'll typically give different '0V' references due to resistor tolerances.
Take your U2A as an example - here the + input sees 100k plus 50k from the divider network, total 150k. The - input sees 25k. You'd also do better to use just one divider network (halve the resistor values) as they'll typically give different '0V' references due to resistor tolerances.
- Status
- Not open for further replies.