Instrumentation Amplifier Buffer fixed my frequency response. Help me understand.

I was experimenting with running different preamps into a Hypex NCore amplifier. I modified the Hypex buffer to create a unity gain buffer so I have the right gain structure, by removing Rg shown here. Works great!

One preamp I have is singled ended tube, AC coupled through a cap, Output Z of 1500 Ohms. If I measure the frequency response into 100K ohms, I get a LF rolloff of almost 5 db (I should increase the cap size - I think its 1or 2 uF).

When I run this preamp into the Hypex, and measure at the output of the preamp + amp, the rolloff is completely gone. It's now flat to 20 Hz. Great!!

The Hypex has has an input impedance of 47K, but yet it behaves like a very light load. THD is even down as well. I tried reading it about it, and I see that the feedback of the instrumentation amplifier buffer somehow simulates a high impedance. But I don't quite get it. Could someone explain how this works? And also, if it doesn't behave like a 47K load, why is the input R listed that way? To everyone trying to bypass buffers of amps, more transparency blah blah, just take a look at the benefits! See pic below.

aloneVsBuffer.png
 
Could be.

To test the preamp I go RCA to an RCA to bnc adapter into a quantasylum qa400 here:
https://forums.melaudia.net/attachment.php?aid=11905
To test the amp out, I created a voltage divider with a 5.6 and 1R power resistor and then put a 1x scope probe at that junction (analyzer can only do 1V in). Other side of scope probe is bnc right to qa400.

From the manual:

The inputs have a series 68 ohm resistor, followed by a low-capacitance ESD protection device, followed by a shunt resistor of 100K ohms, followed by a series C of 10 uF. That signal is then presented directly to a JFET op-amp non-inverting input.”
 
What tube is your preamp pushing the signal with? 12AX7? Tungsol datasheet says plate resistance is 80000 ohms 100 v plate and 62500 ohms 250 v plate. That is way below 1500 ohms you state and higher than 47k quoted for your amp. Just the capacitance of a 2 m RCA coax cable stresses a dynakit PAS2 preamp with 12AX7, which I own. Coax capacitance is about 330 pf/m . 3 m of that alone can roll off highs. dynakit made sure their ST120 transistor power amp had 100kohm input impedance to be compatible with PAS2/3 preamps.
Since the PAS2 doesn't sound any better than an op amp mixer and draws 100 watts, I parked it in the attic. My tube ST70 with original 7199 board (1% HD) sounds worse than my solid state amps with .1% HD. 12AX7 7199 & 6CA7 tubes were great for 1970 when RCA 40636 transistors were $10 each and sounded bad due to 400 hz Ft. Time marches on.
 
Last edited: