Building a PINT, couple quick questions.

Apr 4, 2006 at 9:40 PM Thread Starter Post #1 of 23

Jam_Master_J

Headphoneus Supremus
Joined
Aug 25, 2003
Posts
1,851
Likes
13
Going to build a PINT for portable duty only. Since I'm using IEMs mainly and prefer low-volumes I imagine a gain of 2 will be more than I'll need.

I've been reading the PINT section and the few discussions on the web about the amp and just wanted to make sure I wasn't forgetting anything critical. I'm going to be using the smallest Hammond 1455 case with one 9-volt battery. I plan on using one 9-V nimh and a wal-wart(for charging only).

Anyway, onto my question: I've been reading up on Tangents site and it seems that switching the gain in this amp is a bit more involved than other designs. I'm not really sure what resistors I need to change to get a gain of around 2 and a very low noise floor(since these will be used with IEMs). I'm using the AD8397 opamp. I've read somewhere that in some cases you may want to place a resistor in place of C1 but I don't really know what the best choice for my needs are.

Thanks for any advice,
Justin
 
Apr 4, 2006 at 10:00 PM Post #2 of 23
You may end up having to install C3 if you're going to run at Gain 2 to keep the 8397 stable, although I don't know for sure. You will probably want to not use input caps, so you'll need a resistor in C1.

For gain 2, use 2k (SMD) resistors for both R3 and R4, and put a 1k (axial) resistor in each C1. Gain 2 seems kind of overkill to me, especially for portable use (and greatly limits the amp's versatility), but I wish you luck with it
smily_headphones1.gif
 
Apr 4, 2006 at 10:04 PM Post #3 of 23
Unless I am way off here, to get a gain of 2, you should use a 200k resistor in both r3 and r4, unless you decide to DC couple the amp. To balance the input bias currents, you need to have the same resistance at both the inverting and noninverting inputs. The resistance of R3 and R4 in parallel must be as close to R2, which is 100k, as possible. Since gain is calculated as (r4/r3)+1, to get a gain of 2, use 200k resistors for both, as 1/((1/200k)+(1/200k))=100k is the parallel resistance of r3 and r4. 200k is a standard value, so should be easy to get and ideal for your amp. As for DC coupling or AC coupling, I will leave that up to you; noise floor will be higher because of the increased values of the feedback resistors needed in the DC coupled scenario, but dc offset will be best when AC coupled. When AC coupled, the parallel resistances of the pot and r2, so 10k and 100k, is ~9.1k maximum, which leads to much lower values of r3 and r4. Someone check my thinking though, I havnt actually done this yet.
very_evil_smiley.gif


EDIT
Quote:

Originally Posted by filburt
For gain 2, use 1,5k (R3) and 3k (R4) resistors


I do believe those values gives you a gain of 3. Sorry about the name goof.
 
Apr 4, 2006 at 10:07 PM Post #4 of 23
I gave him instructions for DC coupling. If he AC couples, then the values you suggested are what he'd use, I think. And you're right, I goofed on the gain calculation, forgot to add the +1. Why does everyone always misspell my forum handle?
 
Apr 4, 2006 at 10:07 PM Post #5 of 23
Thanks for the help so far. I'm not too worried about Gain 2 not being enough. I currently have a 5-gain Dynalo and I can barely use 1/5 of the volume knob with my IEMs. I used to have 2 gain on my pimeta and it worked great with Grados and other low-impedance headphones.
 
Apr 4, 2006 at 10:12 PM Post #6 of 23
Quote:

Originally Posted by Jam_Master_J
Thanks for the help so far. I'm not too worried about Gain 2 not being enough. I currently have a 5-gain Dynalo and I can barely use 1/5 of the volume knob with my IEMs. I used to have 2 gain on my pimeta and it worked great with Grados and other low-impedance headphones.


Well, you may want to install C3 just to be on the safe side. I don't know how unstable the 8397 will be at gain 2 on the PINT, but you may need it. If I recall from the data sheet, it does get significantly more stable at gain 2 versus unity, but I don't know if that's enough or not since the PINT already needs ferrites to get it stable at gain 6.
 
Apr 4, 2006 at 10:32 PM Post #7 of 23
Quote:

Originally Posted by Filburt
Well, you may want to install C3 just to be on the safe side. I don't know how unstable the 8397 will be at gain 2 on the PINT, but you may need it. If I recall from the data sheet, it does get significantly more stable at gain 2 versus unity, but I don't know if that's enough or not since the PINT already needs ferrites to get it stable at gain 6.


I've built a couple with a gain of 2: one with input caps, one without. With the input caps, it was stable but very noisy, which I thought was a little surprising, since tangent's noise calculator spit out -88 dB for the noise.

The DC coupled amp has a 2.5 MHz oscillation that I'm trying to fix. Could be my fault (bad soldering), or could be my bad luck -- I'm not really sure at this point, although amb said that the lower resistor values make it less stable. If anyone has any ideas ... (perhaps this is thread-jacking, but I thought it might be useful to the parent too)

With low impedance cans, I think that 6 dB is the perfect gain, but evidently my ears are more sensitive to pain than most people's.
 
Apr 4, 2006 at 11:30 PM Post #8 of 23
I'm using a similar cmoy-ish type AD8397 amp at a gain of 2, and it drives my HD600s to comfortable listening levels at 12 o'clock.
 
Apr 4, 2006 at 11:38 PM Post #9 of 23
Is that out of your micro-dac though? The micro DAC has a rather hot output, especially compared to portable players.
 
Apr 4, 2006 at 11:51 PM Post #10 of 23
Both Micro DAC and ipod. Do you get the same results in your portable setup too?
wink.gif


Using an ipod line-out, I use my low-gain amp at say, 1 o'clock with the HD600s. Still a good region of the pot to be using it at.

A gain of 6 would be more appropriate for like a K340 or something.
 
Apr 5, 2006 at 8:48 PM Post #11 of 23
Think I'm gonna go with Filburt's recommendation and go for R3=R4=2k and the 1-k Axial resistors in C1. I'll also order C3 in case it's needed.


Edit: Can't seem to get 2K resistors in that same resistor line. Closest is 2.21k
 

Users who are viewing this thread

Back
Top