Question About Audio Research Preamp Specs

@dsper 

In the meantime, I will explore different 6922 tubes that might help to mask the LF roll-off. Not sure if that makes any sense...

It doesn't. Changing the tube won't affect the frequency response.

I do not understand technically why a particular preamp may have a greater capability to drive a low impedance amplifier.

While, I try to follow these types of discussions, something basic is lacking in my current understanding as well as how it impacts how it sounds..

Could you perhaps elaborate further or direct me to another thread where I am sure that this has been covered before?

No need- I can explain it.
There are two factors, output impedance and low frequency cutoff. They are related.
First is output impedance and generally to avoid distortion, the amplifier should have an input impedance 10x greater than the output impedance of the preamp. That's easy enough- any preamp mentioned so far does that.
The second bit- low frequency cutoff- is a measure of the value of the output coupling capacitor of the preamp vs the input impedance of the amp. The two together form a timing constant. Here's the formula:
F = 1,000,000 / C x R x2Pi

F is frequency, C is microfarads and R is resistance. Normally the formula looks a bit different (1 divided by the other factors) but since microfarads is a convenient capacitive value I adjusted it.
So if we have a 10uf capacitor driving a 10K load at the input of an amplifier, plugging in the values we see that:
1.59 = 1,000,000/ 10 x 10,000 x 6.28
IOW it will be 3dB down at 1.59Hz.

The problem is that a 10uf cap has colorations, even if its the best Teflon cap money can buy. But otherwise this cutoff frequency is good. But a preamp manufacturer has to weigh options and one of them might be that they want it to sound better with their own amps. So they might limit the value of the coupling capacitor- and thus increase the cutoff frequency into lower impedances. IOW they sacrifice bass response and impact for greater transparency. But if their ideal amp has a high input impedance they might not be sacrificing any bass at all.

Now if you can find a review of a preamp that graphs its output impedance vs frequency, you can see the effect of the coupling capacitor- the output impedance rises as you approach the cutoff (-3dB) point. This shows up in a lot of tube preamps so you can see that many manufacturers regard 10K as perhaps not worth sacrificing the transparency they get with a smaller coupling cap.

Because our preamp has a direct-coupled output, our output impedance curve is identical to the frequency response curve. With capacitor coupled preamps this isn't the case. One further thing of note- its a good design practice to set the -3dB point at to at least 1/10th the lowest frequency you want to play, so 2Hz if you want to be good to 20Hz. This insures that there will be no phase shift at 20Hz, which gives you better bass impact. So this is part of the issue- getting that solid bass all the way down. You need that margin in order to avoid artifacts caused by phase shift. IOW if you want proper bass at 20Hz, your preamp should be able to go to flat down to 2Hz while driving the input of your amplifier.

Take that with a grain of salt. I do prefer direct coupling (no cap), best cap is no cap mentality.

But dsper OP one of the best tube preamps hailed even by many tube and solid state preamp manufactures as being the best tube preamp ever made is the Aesthetix Calypso preamp, and it has massive output coupling caps, as shown in this pic circled in red.
https://ibb.co/4Jb0bSV

This is why we direct couple... this limitation is eliminated.
In the photo of the Aesthetix its not clear what the actual value of the coupling capacitors is. But obviously the Aesthetix faces the same issue of what is the lowest impedance it can actually drive actually; this information does not show up in their specs or operating manual. 

1: If you own the AR pre, tell me the size (uF) of the output coupling cap, (it will be the big ones near each output) and what the poweramp input impedance is you want to use it with, and i’ll calculate at what low frequency it’s at -3db for you.
It’s a simple to correct if too small, with a $20 larger coupling cap, don’t go buying another preamp because of it.

2: Or you can go the other way and get the input impedance resistor (20cents) raised to 50kohm on the poweramp, if it’s fet input no problem, even a 100kohm

Cheers George audioconnectionGeorge
 Makes great sense here.
  easy to have a tech make the resistor changes on the power amp.

The reason ARC didn't do this initially was simply because the larger you make a coupling cap the more coloration no matter what type of construction. Larger caps start to get considerably higher inductance so even if you have the best Teflon its still going to be an audible degradation.
A secondary reason might also have to do with space. Adding long leads to make a part fit in a place that wasn't designed for it is a good recipe for noise pickup.

What does the  "20K ohms minimum load" tell me?

Exactly what it says. Most ARC preamps are not recommended for amps with anything less than 30K but this one is good to 20Kohms. Your amps though have a 10K input impedance, which essentially means that the combination will be bass shy.
As others have pointed out, the output impedance of the LS28 **seems** sufficiently low, but that spec is done at 1KHz. At 20Hz, its output impedance is much higher owing to the size of the coupling caps at the output of the preamp. As a result, the 10K load of your amps will result in a bass rolloff.
So ARC is essentially telling you not to use this preamp with your amps.

We developed a direct-coupled output on our preamps so loads like the McCormick are no worries. At any rate, if you want to use a balanced tube preamp with your McCormick, ARC is not an option if you want it to sound right.