Reducing resistor values in opamp headphone?

[goody75]

I'm a relative newbie and recently modded the headphone section of my preamp.

First I swapped the stock opamp for an LME49720, replaced the input caps w/ Russian K71 polstyrene's and removed the output resistor. A very nice improvement.

Last week, I read on the datasheets for AD797 & LT1028 that it is better (for noise) to use a lower range for the feedback/ground resistors that set the gain (e.g. 300ohm & 100ohm instead of 30Kohm & 10Kohm). I have also found this recommendation on some threads/blogs/websites (I can't remember exact sources at the moment).

So anyway, I tried reducing the resistor values in my headphone section. On each channel there was a 27Kohm feedback resistor and a 10Kohm resistor to ground. I reduced the 27K to 1K and the 10K to 400ohm. This should keep roughly the same gain (3.5 instead of 3.7)
I also reduced -
1) input resistor (just prior to the input coupling cap) from 3.3K to 1K
2) resistor (just after the input cap) going from the signal path to ground from 7.5K to 400ohm.

Not sure if I should have touched the last two but it seemed like everything should be reduced (input resistor was reduced to a lesser extent).

Anyway, I thought the gain should have been pretty much the same but after changing the right channel, the right channel volume was weak. I let the resistors burn in some and it was still quite a bit weaker than the stock channel (all connections checked okay). So I increased the gain by lowering the ground resistor again from 400ohm to 100ohm (I didn't have anything in between). The new gain is now 11 and it sounded not too much louder than the stock gain of 3.7???

I duplicated these changes to the other channel and then noticed that the bass response was now pretty lousy. I recently read somewhere that the value for the ground (gain) resistor impacts the opamp's input impedance. If that is true then I think by lowering the resistor values and reducing the input impedance, the roll off frequency is now much higher and the 1uF input coupling cap is very inadequate. Am I on the right track?

Along the same line, how do you determine the opamp's input impedance (resistance) that would be used to calculate the roll off frequency for an input coupling cap? Because I just assumed that I could use the Zin from the datasheet - I should have known it's never that easy!

Any assistance is much appreciated,
goody75

[sofaspud]

Can you post a schematic of your circuit?
Along with a table showing original and new values, please.

[discrete]

Originally Posted by goody75

2) resistor (just after the input cap) going from the signal path to ground from 7.5K to 400ohm.

1st step, put back that 7.5K.
You lifted the high-pass filter cut-off frequency by a factor of ~20...

[AndrewT]

Let's see the schematic.

[goody75]

Thanks for the quick responses. I will upload a schematic tonight although the scan is difficult to read.

I will also put back the 7.5K resistor.

It also occurred to me in the car this morning that by lowering the resistor before the input cap (from 3.3K to 1K) I may have made it more difficult for the upstream opamp to drive. I'm not sure why but I think it's more difficult to drive a less resistive load. This is just speculation about why the signal sounded weak initially.

Anyway I will post the schematic tonight. Thanks.

[cbdb]

Last week, I read on the datasheets for AD797 & LT1028 that it is better (for noise) to use a lower range for the feedback/ground resistors that set the gain (e.g. 300ohm & 100ohm instead of 30Kohm & 10Kohm). I have also found this recommendation on some threads/blogs/websites (I can't remember exact sources at the moment).

Looks like youve created an input filter. Change the input resistors back. Lowering resistances will increase the load, making the opamp work harder, check the data sheet for distortion versus output to see if you have a problem. Was noise a problem? Leave one channel stock so you can compare your mods. Then try just the feedback resistor mod.I bet the noise difference is very small if its even audible. You may be wasting your time.

The input impedance of an opamp is usually so high you can ignore it, its the rest of the circuit that matters.

[goody75]

Originally Posted by discrete

1st step, put back that 7.5K.
You lifted the high-pass filter cut-off frequency by a factor of ~20...

Just curious... is this 7.5K input resistor-to-ground a 2nd high-pass filter in addition to the input coupling cap? and by reducing it's value, the cut-off frequency increases (like using a lower value coupling cap) - If so, why the additional filter? Or does this change the high pass cut-off frequency by changing the resistance that the input coupling cap sees?

Thanks.

[goody75]

Originally Posted by cbdb

The input impedance of an opamp is usually so high you can ignore it, its the rest of the circuit that matters.

My interest in the opamp's input impedance is for choosing input coupling cap values so when you say I can ignore it, does that mean that the Zin of the opamp is very high (say over 50Kohm) and any reasonable value cap I use will probably be okay?

Even more specifically, can I use the Zin of the opamp to calculate the cut-off frequency with f=1/(2*pi*R*C)? (I'm still a novice & I like specifics.)

Thanks.

[AndrewT]

I would normally expect to see two input filters.
A High Pass to remove DC and to attenuate extreme LF
A Low Pass to attenuate RF and some of the ultrasonics.

One is a CR filter, the other is an RC filter. That's 4 10c components.

[sofaspud]

I'd like to see the schemo, but I'd guess it worked with the capacitor to form a filter. And probably changed the circuit's input impedance enough to cause the weak signal.
The lower resistance recommendation is likely because smaller values of resistance have lower thermal noise than larger values. With audio, it's only important for those resistors in the signal path (ie an input resistor to ground shouldn't qualify).