Voltage follower recommendation for 10K balanced line out

[vitopower]

I have a Suhr Reactive load (guitar amp speaker load simulator with line output) that has a 10k balanced out for feeding a line level guitar signal to ADC input for speaker cabinet impulse response software.  With the unit’s potentiometer turned down to 5, the impedance is about 2.3k.  Many users have reported the 10k unbalanced out sounds better into an instrument input and the newer version has been changed to a buffered 600 ohm out.

I would like use the balanced out to present a clean, balanced low impedance line signal to my ADC via a voltage follower.

Is the attached fully differential line amplifier appropriate for this task?  It is described as “positive feedback / active termination”. 
It is described here on p.10

http://www.ti.com/lit/an/slyt143/slyt143.pdf

Is there a type that would be better suited?

Can I get away with a single supply voltage to use a wallwart instead of building another supply?

THANK YOU!

 

[Newmarket]

That application is for 'High Speed' stuff - look at the filtering that is up in the MHz.
You don't need to get into that stuff (rf termination etc etc)for analogue audio.
Do you a schematic for the output of the Suhr ?

 

[Bo Deadly]

That circuit is not applicable. That positive feedback business is to provide "active termination" which has to do with impedance matching at high frequencies as in like the MHz range.

A THAT 1646 is the sort of IC one might use to make a balanced out (although for a guitar signal it's probably overkill unless you're trying to drive really long cables).

Note that if the 10K pot is connected directly across the dummy load (very possible) you might be able to just change the potentiometer to something like 1K in which case the output impedance will be more like 250 ohms in the middle position.

 

[vitopower]

That application is for 'High Speed' stuff - look at the filtering that is up in the MHz.
You don't need to get into that stuff (rf termination etc etc)for analogue audio.
Do you a schematic for the output of the Suhr ?

Thanks for reply.
No schematic, but I know it's a 10k:10k output transformer with a pot and the impedance is 6k with the pot at 10.

 

[Bo Deadly]

Open it up and post some pics. It's very possible that it's just a giant resistor and a few wires in which case we might be able to tell you what the best way to lower the impedance is. If the output is unbalanced from the wiper of a 10K pot, that would be not good.

 

[vitopower]

Open it up and post some pics. It's very possible that it's just a giant resistor and a few wires in which case we might be able to tell you what the best way to lower the impedance is. If the output is unbalanced from the wiper of a 10K pot, that would be not good.

Thanks for reply.
Its reactive, so there is a combination of dummy coil/inductance/resistance, etc.. Its still in production, but in an improved version, so I'm not sure if posting pics would be appropriate, but I will take a look to better describe the output section. 

I know there is a 10k:10k transformer balanced out and an unbalanced out.  If I want to use the balanced out, I would need a fully differential opamp, I think, right?  I see that THAT is single ended to balanced.

 

[pucho812]

I have a Suhr Reactive load (guitar amp speaker load simulator with line output) that has a 10k balanced out for feeding a line level guitar signal to ADC input for speaker cabinet impulse response software.  With the unit’s potentiometer turned down to 5, the impedance is about 2.3k.  Many users have reported the 10k unbalanced out sounds better into an instrument input and the newer version has been changed to a buffered 600 ohm out.

I would like use the balanced out to present a clean, balanced low impedance line signal to my ADC via a voltage follower.

the real question you first need to ask yourself is how does it sound to you? if you are happy with it, don't mess with it.

 

[vitopower]

Pucho, it sounds like getting 10k balanced to 600 ohms balanced via opamp really is a bigger deal than I thought?..

If it was something common I'd like to be able to compare buffered and unbuffered as it would appear that the impedance mismatch is likely what is causing the issue and why the company added a buffer in the next version (they confirmed this).

This looks like it might work -

https://www.amb.org/audio/alpha24/

 

[pucho812]

O.k. so if I am understanding you correctly, it's not really an issue.  you didn't hear it as an issue. But you were told it was an issue and they did an updated design and that the update sounds better?

 

[vitopower]

The issue was that many users were finding the balanced out (on my version) didn’t sound as good as the unbalanced out and the company has recently added a buffer on the output, likely to address this.

I haven’t heard it yet for myself, period.  I’ve heard enough through demos to know its sounds close enough to warrant using the technology when I want to crank an amp louder than I am able to with a speaker. I’ve heard gear incorrectly loaded enough times to know that more often than not  it results in a less than desirable result.  I’m trying establish the likelihood of being able to hear both options for myself with a few inexpensive parts I have lying around and the good natured help of people who know more about these things than I do on this forum, who might be able to steer me in the right direction.  Then I would know which sounded better and not have to wonder, as it is so often the accumulation of these little differences that end up making big differences in the final product.

 

[pucho812]

I understand. But if you want my 2 cents and you probably don't, if you have not heard it personally, then I would not bother with it.  sounds like it works fine as is.

 

[vitopower]

Right on, thanks Pucho

 

[Bo Deadly]

Debalancing the transformer with a differential buffer to get a low impedance differential output is quite pointless. You could just put a differential output on the unbalanced out using something like THAT 1646.

But there is probably a better way to get low impedance seeing as how the dummy load itself is very low impedance. It sounds like it's just a 10K pot feeding the unbalanced out and 10K:10K transformer. It is strange that they chose to use a 10K:10K transformer since 10K is a little high for a line output. Maybe that is what was corrected. I very much doubt that it sounds different though. The 10K is just more prone to noise over long cables.

To figure out the correct way to get good low impedance out, you would have to post a pic of the inside and / or make a drawing. It could be that you could swap the transformer with a 10K:600 (although this will yield an additional 12dB of attenuation) or also swap the pot to 1K and use a 600:600.

Or if you're just itching to make a circuit, you could put a 1646 on the unbalanced out. But a simple "impedance balanced" output with a single op amp would work equally well seeing as how a guitar amp output is going to have a relatively high noise floor. So I would not get too carried away.

 

[vitopower]

Thanks, I got it wrong.  They replaced the transformer with an output driver IC.  The new one is electrically balanced and buffered.

 

[vitopower]

Debalancing the transformer with a differential buffer to get a low impedance differential output is quite pointless. You could just put a differential output on the unbalanced out using something like THAT 1646.

But there is probably a better way to get low impedance seeing as how the dummy load itself is very low impedance. It sounds like it's just a 10K pot feeding the unbalanced out and 10K:10K transformer. It is strange that they chose to use a 10K:10K transformer since 10K is a little high for a line output. Maybe that is what was corrected. I very much doubt that it sounds different though. The 10K is just more prone to noise over long cables.

To figure out the correct way to get good low impedance out, you would have to post a pic of the inside and / or make a drawing. It could be that you could swap the transformer with a 10K:600 (although this will yield an additional 12dB of attenuation) or also swap the pot to 1K and use a 600:600.

Or if you're just itching to make a circuit, you could put a 1646 on the unbalanced out. But a simple "impedance balanced" output with a single op amp would work equally well seeing as how a guitar amp output is going to have a relatively high noise floor. So I would not get too carried away.

Thanks squarewave, this was good info, and the 1606/1646 is perfect for this application (it also has differential input, which I now understand doesn’t matter).  I will try a buffer followed by this line driver and compare to the original setup.

 

[abbey road d enfer]

Thanks, I got it wrong.  They replaced the transformer with an output driver IC.  The new one is electrically balanced and buffered.

I guess the xfmr was not a very good one, and they judged an electronically-balanced circuit would be much cheaper (about $3 in components) than a really good xfmr ($30+).