less negative feedback in a an amp designed for less feedback might be good, but chip amps are not one of them
Most of the circuit is inside the chip, you can not get to it or change it.
if you want to get into that there are some solid state zero feedback designs. but not in the chipamp forums
Most of the circuit is inside the chip, you can not get to it or change it.
if you want to get into that there are some solid state zero feedback designs. but not in the chipamp forums
Yeah...Janneman...a pot can be trouble. I initially thought that I would have a small fixed resistor in the usual R1 position along with a 25K pot so that there would be always some feedback regardless of the pot wiper position. I'm still seriously considering a high/low gain setup with a DPDT/On-On switch and matched R1/R3 resistors. Let's say one gain set at 20 and another at 60. Or I may give up on the flexibility for this kids boombox and just go with 60 and be done with it...not sure.
So far I'm gravitating to a 100K R2 value and very small 1.6K/1.57K R2/R3 values...seems like an interesting compromise...yields a gain of 60.
So far I'm gravitating to a 100K R2 value and very small 1.6K/1.57K R2/R3 values...seems like an interesting compromise...yields a gain of 60.
Actually a pot in an inverted design wouldn't be all that bad except that it seems impossible to get a good R3 match. The default advice is that R3 should be set to the parallel value of R1 and R2 in order to reduce any bias. Well since the pot becomes part of the R1 resistance value and this varies all over the place, it's impossible to keep this matched. One thought that crossed my mind was a double-pot controlling R1 and R3 in effect coupling the two in near equal values throughout the range. However, this seemed kinda butch...I dunno. Any thoughts on this?
So a pre-amp is out? Didn't see if you said that, sorry if I missed it. For that high of a gain I would use a pre-amp to keep the gain of each stage within reason and overall gain where I want it.
For the switched out gain, be careful. If it is desgined wrong when the switch is changing positions you may get the chip amp to oscillate or head for one of the rails at full speed then back down when the switch closes causing some nice pops. I'd suggest having the high gain as a fixed resistor and switching in a resistor in parallel with the gain resistor to reduce gain. Then all the switch problems should not be an issue.
Since you are driving with a source that can drive low impedance, then use low values for best performance (resistor noise). Some of that gain in low noise is going to be lost with high gain but why not use say 47 - 100 ohms input and then set the feedback. I only say to use more than 32 ohms because most amps will have better performance with higher impedance loads and since you are going to gain up the signal as much as you say then any noise will make a big difference.
-SL
For the switched out gain, be careful. If it is desgined wrong when the switch is changing positions you may get the chip amp to oscillate or head for one of the rails at full speed then back down when the switch closes causing some nice pops. I'd suggest having the high gain as a fixed resistor and switching in a resistor in parallel with the gain resistor to reduce gain. Then all the switch problems should not be an issue.
Since you are driving with a source that can drive low impedance, then use low values for best performance (resistor noise). Some of that gain in low noise is going to be lost with high gain but why not use say 47 - 100 ohms input and then set the feedback. I only say to use more than 32 ohms because most amps will have better performance with higher impedance loads and since you are going to gain up the signal as much as you say then any noise will make a big difference.
-SL
SpittinLlama:
I'm trying to digest what you just said but yeah...
I'm not using a preamp. I am combining L-R channels though so I'm using little summing resistors of 1K each for some channel isolation. This effectively sets my resistance floor value at 500 ohm if I'm right (the 2 little 1K resistors in parallel effectively) and I'm adding R1 proper at 1.2K for a total effective R1 value of 1.7K. My gain would then be roughly 60. Am I thinking of this right?
C1 is high enough to keep the cutoff at about 10 Hz.
How about the doube pot controlling R1/R3 idea instead of the DPDT switch? Any better?
I'm trying to digest what you just said but yeah...
I'm not using a preamp. I am combining L-R channels though so I'm using little summing resistors of 1K each for some channel isolation. This effectively sets my resistance floor value at 500 ohm if I'm right (the 2 little 1K resistors in parallel effectively) and I'm adding R1 proper at 1.2K for a total effective R1 value of 1.7K. My gain would then be roughly 60. Am I thinking of this right?
C1 is high enough to keep the cutoff at about 10 Hz.
How about the doube pot controlling R1/R3 idea instead of the DPDT switch? Any better?
Attachments
What I meant about the switch if that if the contacts are completely open when switching then the LM3886 may do some funky stuff causing pops or other problems. So the idea is to have the higher gain set and then a switch that is normally open. When you close it another resistor is added in parallel with the R2 resistor in your circuit so the gain drops.
I have not thought enough about your input resistance to know if you are right. The problem you have with using the inverted is keeping R3 equal to the parallel combination of R1+input and R2. I think you can probably use a dual pot for R1 and R3 since their correct values are so close because R2 is 100K. Or go the non-inverting route and then no problem with the R3 value. Might want some resistance in series with the pot so you don't accidently go to 0 ohms.
-SL
I have not thought enough about your input resistance to know if you are right. The problem you have with using the inverted is keeping R3 equal to the parallel combination of R1+input and R2. I think you can probably use a dual pot for R1 and R3 since their correct values are so close because R2 is 100K. Or go the non-inverting route and then no problem with the R3 value. Might want some resistance in series with the pot so you don't accidently go to 0 ohms.
-SL
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.