For some reason I haven't been getting email notification on this thread for a while. I've gone back a few pages to read the discussions.
In one segment Roam is talking about the operating point in the front end.
Any amp involves trade-offs. In this case, there was a current budget for the PS, a desire to not use a CCS for the plate load on the gain stage, and a desire to use the same B+ for the input and output stages to keep the transformer requirements to one HV secondary. Many of these trade-offs were to keep the cost down so that the amp $$ budget could fall into a realm where many diyers who wanted a tube amp could build one while still making a very good sounding amp.
The amp was originally designed to only use 6922s for the input stages. Nevertheless, there were quite a few folks who wanted to try 6N1P and 6H30 and I felt that the best answer was to provide a value for the cathode resistor that would set the voltage op points correctly. This is a better solution than having builders directly sub these tubes with no changes because not changing the cathode resistor really messes up the amp and, well, some folks were going to do this whether it was a good idea or not.
Some of the reviewers have simply substituted 6N6P into the front end and then noticed that, when they put 6922s in, the amp sounds a whole lot better.
The 6922 is running a bit lean in the first two stages. Doing so lets us use the 100k plate resistor to get a respectable curve from the 6922 without a CCS. But it is still running in the lower part of the plate curves where it is less linear than if we were running 10mA through it.
A 6N1P also runs a bit lean and at less than optimum voltages. Adjusting the cathode resistor sets the voltage operating points to where they should be for the amp and lets the 6N1P function reasonably well.
The 6H30 is a bit more problematic because it really wants to run at higher idle currents. But the amp is not designed to support 10mA in the front end on both channels. So the 6H30 has to run lean too.
Most folks tend to judge an amp, ultimately, by the way it sounds and it was clear that most of the builders wanted roll the at least one or two tubes in and out to decide which ones they liked best. Some folks are liking the sound of the 6H30 even though it is not in optimal operating condition.
Well, who am I to say who is right when it comes to nuances in the sound? I want everyone to use 6922s because that's what I designed, but when you build it it's your amp.
On the NFB. The NFB has two purposes. First, as Roam said, it lowers the Zo for low Z headphones. It also reduces the THD for the lower Z headphones. Driving 32R directly is not easy for the output tubes. Some of their non-linearities become more prominent. The easiest way to fix this is to provide some amount of NFB. The NFB also has the added effect of reducing the gain so that the vol pot has some range for the low Z headphones.
Prototypers have also noticed that the adjustable NFB affects the sound stage in a way that the listeners can easily hear. Different listeners find different settings that suite their listening styles.
I hope this clarifies some of the thinking behind a few of the things that are happening with this amp. If not, please let me know.
