NIGC Buffer

[Assesears]

G'Day

Been looking through the threads can't find anything on nigc buffers, does it not work out in the 'ni' setup or does it sound sh!7e. I thought the nigc setup is the latest craze, it seem's not many dudes are building buffers with them or am I totally wrong.


Cheers Bruce

 

[Sheldon]

I've done it. Details here: http://www.diyaudio.com/forums/showthread.php?s=&threadid=35442

It's not quite as common, because the NIGC generally has higher input impedence and therefore, in theory, benefits less. I built Pedja's discreet jfet buffer and it sounds very nice to me. I haven't compared with non-buffered. No reason not to try if you are interested. Carlos's buffer should work too. I might try that one next time.

Sheldon

 

[Nuuk]

You can use a buffer with a NIGC just as you can with the inverted type.

Does it make the amp sound better? Only you can say by trying one.

 

[mikelm]

It's not quite as common, because the NIGC generally has higher input impedence and therefore, in theory, benefits less.
Sheldon

Mmm... I think it would be the other way round.

the inverting i/p GC always has the i/p resistor inbetween the amp i/p and the o/p of the buffer so the amp i/p sees just under 10K minimum

on the NIGC the low impedance of the buffer can be directly injected into the amp i/p so the amp sees a very low impedance which should reduce distortion and give a 'bigger' sound.

also on the NIGC this effect can augmented as well by reducing the values of the feedback resistors - but it will be a matter of taste as to wether this is desirable and low impedance FB will probably erode stability margins.

mike

 

[planet10]

The buffer has a theoretically greater benefit on the inverted gainclone because it removes whatever you have plugged into the input from the feebback loop.

dave

 

[Nuuk]

The buffer has a theoretically greater benefit on the inverted gainclone because it removes whatever you have plugged into the input from the feebback loop.

Dave, I think you meant 'input loop".

 

[planet10]

I think you meant 'input loop".

Maybe feedback circuit would be more correct ... in an unbuffered IGC, the cable, and at least part of the source device (ie CD player back to the eoutput buffer) are part of the feedback circuit.

dave

 

[mikelm]

Maybe feedback circuit would be more correct ... in an unbuffered IGC, the cable, and at least part of the source device (ie CD player back to the eoutput buffer) are part of the feedback circuit.
dave

Whatever we choose to call it there seem to be good auguments for having a buffer for all kinds of GC.

I have only tried a transformer volume control for this function. I wonder if anyone out there has compaired TVC, Valves, opa627 etc, and Jfets.

I would be interested to hear opinions of how they compare.

mike

 

[skyraider]

buffered nigc

I used my NIGC with 2132/buf634 buffer.
I just like it better, even though its a bit noisier. I get tighter bass, and less harsh highs. I wouldnt say buffer is better, maybe its an overall combination.

 

[jan.didden]

Maybe feedback circuit would be more correct ... in an unbuffered IGC, the cable, and at least part of the source device (ie CD player back to the eoutput buffer) are part of the feedback circuit.

dave

Dave,

Good point for the IGC. On the other hand, any decent CD player or preamp has its own buffer (or at least low output impedance stage) so you end up buffering the buffer...
But, again, tube sources can be a different kettle of fish due to highish and/or non-linear output impedance.

Jan Didden

 

[planet10]

any decent CD player or preamp has its own buffer (or at least low output impedance stage) so you end up buffering the buffer...

The interconnect (and the connections it connects too) are being taken out of the loop when you add the buffer at the front of the IGC.

dave

 

[GregGC]

The interconnect (and the connections it connects too) are being taken out of the loop when you add the buffer at the front of the IGC.

dave

Hope that doesn't take you off topic:
Isn't the "NF loop" from Out of the amp to "-in"? And "-in" has GND caracteristic (0 volts DC or whatever "+in" sets it too be and extreamly low AC impedance to GND). So the influance of the cable and connectors for the IGC would be as strong as it would be to NIGC.
Just my understanding, though.

/Greg

 

[planet10]

Isn't the "NF loop" from Out of the amp to "-in"? And "-in" has GND caracteristic (0 volts DC or whatever "+in" sets it too be and extreamly low AC impedance to GND). So the influance of the cable and connectors for the IGC would be as strong as it would be to NIGC.

I don't pretend to really understand what is going on, my comment is just a summary from a whole slew of posts elsewhere on the subject. My understanding is that a NIGC avoids the problem.

dave

 

[peranders]

You are quite right Dave. An inverting amp is much more sensitive against the signal source (including cable) and I don't really understand why a buffer doesn't seems to be a rule. I don't take any changes so I use a buffer and also lower the resistor values for the LM3886 (in my case).

http://home.swipnet.se/~w-50719/hifi/qrp02

 

[digi01]

peranders,your qrp02 project layout is very neat.really good work

thanks for your share it.


digi

 

[peranders]

Thanks, I like it also

 

[Joe Rasmussen]

I don't pretend to really understand what is going on, my comment is just a summary from a whole slew of posts elsewhere on the subject. My understanding is that a NIGC avoids the problem.

dave

The confusion lies in the fact that the "-" input of an IGC is a "virtual earth" and as such there is no theoritical signal at that point. So some would argue that the feedback loop stops there. The other side of the argument is that the input Z defines one arm of the components that determines the gain, and also imparts its qualities of that gain (like source Z, cable characteristics etc), so how can you say it is not in the feedback loop. It really depends how you view it, I am inclined towards the latter.

Has anyone noticed how the IGC resembles an I/V converter? Really, when you think of it, the input voltage is converted to equavalent current. This current is maintained in the feedback resistor as the virtual earth really doesn't have any way of sinking the current, it's only virtual, not real. But the critical input current isn't just defined by the input resistor alone. It is source sensitive.

So there are so many ways to view what is happening in an IGC.

Joe R.

PS: Has anyone forgotten who started this silly notion re buffers?

 

[planet10]

Has anyone forgotten who started this silly notion re buffers?

dave

 

[carlosfm]

Good point for the IGC. On the other hand, any decent CD player or preamp has its own buffer (or at least low output impedance stage) so you end up buffering the buffer...

Jan, I suppose you are talking about a power amp.
But on an (unbuffered) integrated amp, you have a pot before the chip, and even that affects the feedback and the overall gain of the (inverting) amp.
Dave is right.

I have a inverting LM3886 power amp and a preamp (OPA627+BUF634) and long ago decided to use input buffers on the amp (OPA627).
I have a 5 meter shielded interconnect between the pre and the power amp.
My pre has no problems with that.
But with the unbuffered inverted amp the signal cable made a good antenna.
I knew some seconds before that my mobile phone was going to ring.
Adding the NI buffer cured that.
Dead silent, no interferences, no anything.
And better sound IMHO.

 

[carlosfm]

It is source sensitive.

Yes, it is.
For me, it has better sound than NI, but needs a good buffer.
NI with a good buffer is better than without it, but still I prefer buffer+inverting.