PPAS Build Thread.
Sep 25, 2006 at 3:28 AM Post #106 of 254
Quote:

Originally Posted by ozshadow
I am using the 4.7k resistor to set a gain of just under 6

I do not understand the idea of optimizing the resistor values. What do you suggest and what differences would it make ?

I will be running the tread at about 26.6 which by the time it comes out of the charging circuit, it should be about 26.0 flat.



Alright, I'm going to try to summarize Walt Jung's article bluntly and quickly as I have to go to work in a while. If I butcher something, it is in haste and someone correct me.

Bandwidth is the operating range of your amplifier, with the values chosen previously, you would be running a functional equivalent of a low pass filter. This happens because the amplifier effectively doesn't amplify frequencies as much above the bandwidth. The higher above the cutoff, the lower the amplification. Your cutoff would have been somewhere in the audible range, probably 12-16kHz somewhere, depending on the tolerances of the opamp.

Basically, in a typical voltage feedback opamp circuit, gain and bandwidth share a generally linear relationship. I.E. if the gain is changed by a factor of 5, the bandwidth will change by a factor of 5. What we are doing in a multiloop is intentionally lower the overal bandwidth by using 2 loops. The inner loop sets the local gain and limits the bandwidth of the amp. The outer loop sets fairly close to the effective gain of the amplifier. The benefit of doing this is we can limit the frequencies outside of a preset range to gain stability, resistance to electrical noise, and improve phase margin. Basically you are going to set inner loop gain as (R5/R6)+1. This should be high such that GBP/ILG = somewhere from 60-100 kHz; this will be the operation bandwidth of the whole amplifier, the "cutoff" if you will. There are a couple guidelines in determining the whole set of values - general consensus is that R3 should be set somewhere above 220R, R5 should be 5-10x higher than R3, and lastly, you should avoid picking any of the values greater than 1M. Outer loop gain, which has a greater influence on the total gain of the amp is then adjusted using R4. Also, when picking values, remember the lower resistance will equal a lower noise floor and higher current use; the higher the resistance, the lower the current use but the higher the noise floor. From here, you want R1 = R3+R5. You are going to, of course, want to pick values that are available for purchase rather than arbitrary values.

Alright, I'm out of time. I can help you out or suggest some values for you in the morning if you want. PM me or respond here, I'll be around.

Edit - Wow, that was in a rush, fixed a bunch of spelling/grammar errors
 
Sep 29, 2006 at 10:58 PM Post #107 of 254
Hmmm, exactly how many SST502's are needed for Class A, because the BOM says ya need 2, but there are 3 pads for them.


And guess what, I ordered two.
 
Sep 30, 2006 at 3:38 AM Post #108 of 254
Class-a biasing ground has arguable benefit. 2 is the point of diminishing returns. Populate both L and R.
 
Sep 30, 2006 at 3:24 PM Post #109 of 254
Just an FYI, but if you want to use the C1201 case with 2 9v batteries, you will need to grind the inside rails down a bit as it is an extremely snug fit otherwise, the middle rail primarily.

http://www6.head-fi.org/forums/showp...6&postcount=37

I used a micro die grinder to give me an extra mill or two on each side, leaving just enough for the screw to thread into.
 
Sep 30, 2006 at 8:36 PM Post #110 of 254
Post some pictures, also where did you get the C1201? and do you have that charger running? I only see the 1202 available on the hammond site.

ATAT.
 
Sep 30, 2006 at 11:44 PM Post #111 of 254
It's this case:

http://www.mouser.com/search/Product...46-1455C1201BK

I could take a picture, but nothings in the case yet, and I already recolored most of the insides black again. I still have to order the pot from Tangent to complete the amp, but everything listed in the diagram fits, I've already tried it.

This is not my pic, but it fits very similar to this pic, which uses the same case:

http://www6.head-fi.org/forums/showp...&postcount=164

As for the micro die grinder, I used one at work. It's just like this one:
http://www.harborfreight.com/cpi/cta...emnumber=47869

Like I said, I only needed an extra 2 mills or so.

During final assembly, I'll take pics along the way. Prolly order my pot next week.
 
Oct 8, 2006 at 12:27 PM Post #112 of 254
Oh, I am pulling a flat 17.0 mA current draw with opa2107/627/buf634 on low/SST502 class A


How long should Two 9v 270 mah batteries last in that ?
 
Oct 8, 2006 at 7:07 PM Post #114 of 254
Oct 11, 2006 at 4:33 AM Post #115 of 254
Anyone still interested in Diamond buffer boards?

I'll be making a proto run sometime soon.
 
Oct 11, 2006 at 5:22 PM Post #118 of 254
Quote:

Originally Posted by Tedro
All SMD? SOT-23 NPN/PNP with 805 resisters?


TO-92 input and output transistors, 1206 emitter resistors on the output pair, and 0805 emitter resistors on the input pair, for class AB biasing. CCS is provided by a pair of SST50x precision jfet crds, in SOT-23.
 
Oct 12, 2006 at 12:46 AM Post #119 of 254
Finished LM6171 / OPA2107 amp (LM6172 had .5 v offset, probably still impedance balancing issues.. will sort that out later)

Sounds absolutely fantastic, the OPA2107 just fits really well with this amp.

Oh and the buffers are LMH6321, really has that nice slam to it =)

will o
 
Oct 12, 2006 at 9:09 PM Post #120 of 254
Might want to edit the original post before people start asking where you and I got those magical 2701 chips (the 2107's)
smily_headphones1.gif




I think I could now solder this board with my eyes closed. Freakin buffer ! I think I soldered/removed/resoldered everything but the caps on mine multiple times just to find out one buffer had issues.


But I completely agree, the 2107 is a beautiful chip in this amp.
 

Users who are viewing this thread

Back
Top