On the scope I see about 22V peak to peak before clipping the opamp
[overlapping a little with the post from Nick Salis because his posted while I was still typing this]
Audio levels are defined in RMS values relative to a sine wave of the same amplitude. 22V p-p is 11V peak, which for a sine wave is 7.78V RMS. 7.78V RMS is right at 20 dBu (which is calculated as 20*log(Vrms/0.775) ). Just to make sure everyone is on the same page about the levels being discussed in different terms (peak voltage, RMS voltage, dBu levels).
Running from +/-15V supplies an NE5532 should be able to hit 12V or almost 13V peaks, but even going to 26Vp-p levels is only a little more than 21 dBu, so the difference between 11V peaks and 13V peaks is not much to worry about.
In either case, an NE5532 is more than capable of driving to the maximum level that the interface can handle.
There are some details of the input circuitry you would need to know to verify that the interface can actually handle 18dBu in all configurations, or if it actually requires symmetrical drive to handle that.
I was under impression that I could double the level by using transformer...I have tried to connect the primary single ended side of a standard 1:1 10k transformer
To double the output level you would need to use a 1:2 transformer. A 1:1 transformer simply transfers the input voltage to the secondary winding, providing galvanic isolation but no change in level (other than possibly a slight level drop due to the resistance which the secondary winding is adding in the signal path).
When I do this I observe half the levels (11V peak to peak) on each secondary
The terms seems to be mixed a little bit. A "1:1" transformer usually implies one primary winding and one secondary winding, but when you write "each secondary" that implies there is more than one. Is the transformer a 1:1CT (center-tapped secondary)?
I observe 22V peak to peak on one of the secondaries when plugging in the TRS jack but 0V on the other secondary.
A specific transformer part number and schematic will help clarify what you are seeing. The schematic you drew in
post #4 shows a transformer with a single secondary, so what does "other secondary" refer to?
I have only a one channel oscilloscope but if I probe them separately they seem to have the same phase,
Phase (or even polarity, i.e. a complete positive/negative swap independent of frequency) can only be measured relative to a reference signal, so you cannot tell with a single channel 'scope. There are certain asymmetric test signals you could use to check with only a single channel scope, but with a sine wave you have no way to tell what the relationship is to the other signal pin.
When I connect the TRS to the interface, hot gets twice the swing and cold goes to 0V which I find weird
What connections do you have between the transformer, pin 2, pin 3, the audio interface, and the oscilloscope when measuring with the TRS connection?
Measuring from hot or cold to somewhere else is meaningless.
I know that Ian (ruffrecords) knows this, but just so it is really clear to any less experienced people:
Almost every oscilloscope (except a battery powered portable 'scope) will have the probe reference pin labeled as "gnd" and tied to the chassis 0V level, which is going to be referenced to the power line protective earth connection.
An AC powered audio interface (such as your PreSonus) may or may not have a protective earth connection depending on how the power supply is wired and whether the computer interface is fully isolated or not. The safest expectation is that the interface will also have the circuit reference/0V/"gnd" connection tied to power line protective earth in some way as well.
So if you connect an oscilloscope probe between pins 2 and 3, that will connect the pin to which the probe gnd is connected to the 0V potential defined by power line ground. That may or may not change the measurement depending on whether the output is transformer coupled, directly driven with asymmetric drive (e.g. impedance balanced ouptut), directly driven with symmetric drive but using a cross-coupled circuit to emulate a transformer, or directly driven with symmetric drive using independent op-amp channels.
In other words you need to have some fairly sophisticated understanding of the circuitry you are measuring in order to interpret the results correctly.
(re: new measuring setup)
I connect my signal generator directly to an unbalanced TS jack that I connect to the audio interface.
The audio interface is balanced, so how exactly did you connect the signal generator? Are you using the TRS portion of the combo jack on one of the mic/line switchable inputs (input jacks 3-8)?
I configure the signal generator to generate a sine at roughly 20 Vpp
That is 19dBu, so should be slightly overdriving the interface since maximum input is specified as 18dBu.
The DAW reports this as being -18.7 dBFS
Something is wrong there, that should be 0 dB FS, or at least -6 dB FS even if the interface has some kind of brain-dead non-differential design (which I don't expect from PreSonus).
Are you sure that the signal level is still 20V p-p when the signal generator is connected to the interface? Most signal generators should have no problem at all driving into 10k.
It would also be good to double check that there are no software level controls anywhere in the data path which might be attenuating the signal in the digital domain.
Second question: Am I wrong in assuming that this [transformer coupled 20V p-p signal] should +6dB relative to -18.7dBFS?
That is wrong, 20V peak-to-peak (or 7V RMS as it would more typically be described in audio context) is the same voltage no matter whether you transformer couple it or not. A decent audio input design will have true differential inputs, so all the signal amplitude on pin 2 only and 0V on pin 3 (as you get with a TR plug inserted into the combo jack) will be the same as half amplitude on pin 2 and half the amplitude on pin 3. The exception to that may be if the input accepts a maximum level which is difficult to handle with typical 15V power supplies. For example 24 dBu is over 17V peak (about 34.7V peak-to-peak) so that obviously cannot be input directly to an op-amp running from 15V bipolar supplies without attenuation. Some input designs will attenuate the input signal so that even asymmetric 24dBu input can be handled, some will require that signals over some value are symmetrically driven, so you just have to know the limits of the device you are using. Does not likely apply to your Quantum which only accepts up to 18 dBu signals (which is just under 6.2V RMS, or about 17.4V peak-to-peak) maximum.
So a 20V peak-to-peak signal should be showing up as over 0 dB FS (clipping indicator should be on) whether directly coupled or transformer coupled.
Both connections should be the same, the question is why do they get displayed as around -19 dB FS instead of 0 dB FS? The signal is actually around 20 dBu, so if accurate the meter would be indicating that to hit full scale the signal would have to reach 39 dBu. That is obviously nonsense (a signal that high would be almost 70V RMS, or over 195V peak-to-peak) so the question is where is the level being attenuated between the physical inputs and the metering software.
I cannot find an online user manual for that interface. Do the LED's beside the gain knobs turn red when clipping? The 2nd harmonic is pretty high in that meter picture, I would think that a clipping indicator would be on somewhere.
