[Filburt]

Quote:

Originally Posted by amb It's because we want to keep the combined resistances at the + input and - input to be close to each other for the lowest DC offset. When C1 is installed, the resistance on the + input is the value of R2 (100KΩ). To keep the - input side at 100KΩ you will need very high values of R3 and R4 (the parallel value of these, e.g., 120KΩ || 620KΩ ~= 100KΩ).

Ah, okay. I thought it might be this. Thanks for the helpful reply!

One last question: is that what is raising the noise floor so much? That is, the use of very large value resistors when C1 is installed? The discussion here leads me to believe that this is the case.

 

[amb]

Quote:

Originally Posted by Filburt One last question: is that what is raising the noise floor so much? That is, the use of very large value resistors when C1 is installed? The discussion here leads me to believe that this is the case.

Yep! This is the dilemma with these bipolar opamps. If you want to have an input cap then you'd have to live with high resistances and high noise, or eliminate the input cap and have the DC offset vary a bit with the volume pot.

 

[BavariaBarbarian]

Hmm...

Just for kicks I've run Tangent's calculator for Johnson noise:

R3=120k R4=620k (Gain ~6)

99.77 µV, -81 dB relative to 1 Vrms

R3=1.2k R4=6.2k (Gain ~6)


38.47 µV, -89 dB relative to 1 Vrms

R3=1.5k R4=4.5k (Gain ~4)

25.66 µV, -92 dB relative to 1 Vrms

R3=120k R4=470k (Gain ~5)

72.51 µV, -83 dB relative to 1 Vrms

And the combination I'd probably use:

R3=150k R4=300k (Gain=3)

42.9 µV, -88 dB relative to 1 Vrms

All with R[size=x-small]pot[/size]=10k and R2=100k respectively 100k+1k.

I've no way to verify these by measurement, though.
I don't plan do build and RMAA that many Pints.

Btw, cheers for posting your impressions darkisz, and you did a beautiful job with your Pint!

 

[Filburt]

Doh! Looks like maybe I should have gone with Gain 4! I ordered resistors for 6, though

Would Gain 4 have been powerful enough for things like AKG K701s or DT880s? If so, I am now a sad panda.

 

[Teerawit]

Dunno about the K701 and DT880 but a gain of 4 drives the HD600 just fine (my suitable listening levels are around the middle of the pot's sweep, which is good). This is from a fairly strong line level output (Micro DAC).

 

[Filburt]

Quote:

Originally Posted by Teerawit Dunno about the K701 and DT880 but a gain of 4 drives the HD600 just fine (my suitable listening levels are around the middle of the pot's sweep, which is good). This is from a fairly strong line level output (Micro DAC).

How far do you have to turn the volume up?

 

[Filburt]

Alright, I grabbed 4,53k and 1,5k resistors, and also a 1,13k resistor since the calculation I got with tangent's formula is ~1,127k instead of ~1,0k as in the case of 1,2k/6,0k.

 

[Teerawit]

Quote:

Originally Posted by Filburt How far do you have to turn the volume up?

Around the middle of the pot's sweep (starts at 7 o'clock, listen at ~12 o'clock)

 

[Filburt]

Quote:

Originally Posted by Teerawit Around the middle of the pot's sweep (starts at 7 o'clock, listen at ~12 o'clock)

Oh wow, that's not bad at all. I'm glad I grabbed the extra resistors, then, as I'm probably going to make it Gain 4.

Do you think 4,53k will work alright? Also, what did you use for the C1 resistor? I have 1k resistors and 1,13k resistors both available to me.

 

[Teerawit]

Quote:

Originally Posted by Teerawit Anyone tried (or willing to try ) the PINT with AD8397, no input caps, 1.2K and 6K gain resistors (or something close), and no ferrite beads?

No one? I don't have a scope so...

 

[Filburt]

Quote:

Originally Posted by Teerawit No one? I don't have a scope so...

I won't be able to test it for you until this summer, unfortunately. If for some bizarre reason, no one has tried it by then, I'll give it a shot if I have a spare board around.

 

[Teerawit]

Quote:

Originally Posted by Filburt I won't be able to test it for you until this summer, unfortunately. If for some bizarre reason, no one has tried it by then, I'll give it a shot if I have a spare board around.

Ah. I just have a hunch that the ferrite beads aren't necessary with the AD8397 in that specific circuit

 

[nysulli]

Quote:

Originally Posted by amb No. You want to select R3 and R4 such that their paralleled value is about the same as the paralleled value of R1 and R2. If you have R1 = 1KΩ and R2 = 100KΩ, and you want a gain of 4, then R3 = 1.5KΩ amd R4 = 4.7KΩ gets close.

one more question, wouldn't aiming for a parallel resistance above 1k for R3 and R4 be somewhat of a good thing, in that you could more closely match the parallel resistance of R2 and R1+pot in the resistance range of the pot where you'd be listening to the amp, or am I not considering something that could cause me some issues

 

[Filburt]

Quote:

Originally Posted by Teerawit Ah. I just have a hunch that the ferrite beads aren't necessary with the AD8397 in that specific circuit

How come? Even at Gain 6, you're well above the audio band in your bandwidth...

Also, I'm wondering what nysulli is wondering. I got 1,13k resistors just in case, but I also have 1k resistors available.

 

[morsel]

Teerawit, the ferrite beads are required, not optional. Leaving them out will result in oscillation that can easily toast the opamps. This oscillation can be caused by an unterminated cable on the output, among other things. We blew over a dozen opamps testing so far, and we have a scope, current limiting power supply, realtime distortion analyzer, etc. The AD8397 is not a forgiving opamp.

For those of you wishing to Mini³ify your PINT, here are some gain resistor values:

Reqp = R3*R4/(R3+R4) = equivalent parallel resistance = 1k
R4 = Gain*Reqp
R3 = R4/(Gain-1)

Code:

[left]GainR3R4 22k2k 31.5k3k 41.33k4k 51.25k5k 61.2k6k 111.1k11k[/left]