Bassman Micro: Negative Feedback Resistor Value???

[Lancer X]

There was another thread puzzling out this resistor value, but it was from a while back and the consensus was basically "@robrob did this for a reason". I am trying to figure out why @robrob chose a 15K NFB resistor for his Bassman Micro designs.

In his 5F6A Switched NFB Mod page, he discussed that the original 5F6A featured a 27K NFB resistor, and the fact that:

"If your 5F6A feedback is tapped off an 8 ohm speaker output like mine then your 5F6A feedback resistor should be 56k to give the same amount of feedback as the original 5F6A's 2 ohm output with a 27k feedback resistor. You should use a JTM45 equivalent resistor of 20k."

In that same page, he also provides a conversion factor for amps featuring speaker taps other than 2 ohm, and the JTM45 equivalent NFB resistor for 8 ohm would be 20K. As such, it makes sense that in his Deluxe Micro Mod design, the NFB resistor for the JTM45 setting is 22K (close enough to 20k.)

What has me puzzled is why the Bassman Micro NFB resistor is specced out to 15K. Why not 56K???

I plan to include a 3-Way NFB selector in my Bassman Micro build - that's why I'm trying to understand and sort this out.

 

[2L man]

Output voltage on 8 ohm output is double what 2 ohm output so double value voltage divider is used to get about the same NFB.

Lower power power stage output voltage is perhaps proportionally lower what the PI input signal is and voltage divider circuit is tuned for that. When you tune the amp for your liking just test different ratios.

 

[Lancer X]

Output voltage on 8 ohm output is double what 2 ohm output so double value voltage divider is used to get about the same NFB.

Exactly. The original 5F6A had a 27K NFB resistor and a 2 ohm output, so for 8 ohms output it seems that the Bassman Micro should have a 56K resistor. Rob has a 15k specced in - that's what I don't understand.

 

[Ten Over]

Exactly. The original 5F6A had a 27K NFB resistor and a 2 ohm output, so for 8 ohms output it seems that the Bassman Micro should have a 56K resistor. Rob has a 15k specced in - that's what I don't understand.

It's not that simple. You have to calculate the open loop gain first and then the closed loop gain and then compare the ratio between the two to the ratio of the 5F6-A. The nominal impedance of the OT secondary is only one part of the calculations.



After a quick half-page of calculations, I came up with something between 27k and 33k for NFB similar to a 5F6-A.

After writing this, it came to my attention that he has more than one micro Bassman. I calculated the NFB for the EF80 version.

 

[Lancer X]

Holy smokes. Thanks @Ten Over. Looks like I need to hit the books on this one!

 

[Lancer X]

So, at this point I have successfully been able to work through the math to show that a 27k FB resistor in the 5F6A created a 45% ratio of closed-loop to open-loop gain.

Ditto for the JTM45, which has a ratio of ~27.5%, and that a 10k resistor will convert a 5F6A to 27.5%, as @robrob shows in his NFB Switch Mod page.

When I apply the same math to the EF80 LTP Micro, however, I can't determine why he chose a 15k FB resistor (assuming that he was indeed aiming for ~45% closed-to-open gain). I keep coming up with a number more like 195k ohms to replicate 45%. (I am referencing Richard Kuehnel's Basic Theory and Bassman books, as well as Rob's pages of course. So thankful for these resources!)

I know there's some nuance I am missing, but it's making me nuts...