Semisouth SJEP120R100 Power Loss

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
K

kasey197

Member
Joined 2009
In an F6 style setup, the dc bias and ac drive are set independently. If we use one set of output R100s, then while it may sound a bit umm, left field, it seems to me that Vgs matching is less important than matching of gain characteristic (eg slope and shape of Vgs vs Id curve) ...
 
S

Semisouthfan

Member
Joined 2012
kasey197 said:
Makes sense ... the curve straightens up higher I go in bias so I guess that makes the matches a bit easier to make compared to lower bias ...
Click to expand...

My estimate is that the output curves that participate in load-line canceling are going to be the bigger factor in matching, but really, the output and transfer curves "work together" to set the residual harmonics (assuming this linearisation effect dominates other NF) and so I'm doubtful that most device matching methods are all that useful. The fact that dsdjoy reported random selection was as good or better at finding the results he wanted (post #69 of this thread) reinforces my hunch.
 
flg

flg

Member
Joined 2005
Yep, I like 'em to Buzz, Mike?
I'm still here, I've just been workin on the auto a little lately. Blew a rear shock out in a pot hole at 25 degrees a few weeks ago. While I was driven around trying to figure out the noise and stuff the other one blew out? Had to go back to my OEM shocks. The bouncy feeling was to much... Another "high performance" Chinese component :nownow:
Gotta attend to the other day to day and work stuff too :D
 
Last edited:
K

kasey197

Member
Joined 2009
Semisouthfan said:
My estimate is that the output curves that participate in load-line canceling are going to be the bigger factor in matching, but really, the output and transfer curves "work together" to set the residual harmonics (assuming this linearisation effect dominates other NF) and so I'm doubtful that most device matching methods are all that useful. The fact that dsdjoy reported random selection was as good or better at finding the results he wanted (post #69 of this thread) reinforces my hunch.
Click to expand...

Am skeptical to be honest, but also happy to be convinced ;) dsd has access to his ears, I only have a distortion analyzer (in the first instance)
 
S

Semisouthfan

Member
Joined 2012
kasey197 said:
Am skeptical to be honest, but also happy to be convinced ;) dsd has access to his ears, I only have a distortion analyzer (in the first instance)
Click to expand...

Well, that's interesting. dsd was honest about saying it worked for what he "wanted," and that is pretty clearly the rub. I'm getting the idea that managing residual harmonics is where objective and subjective mix most vigorously. Matching channels in the context of residual harmonics looks to me to be pretty challenging, so I'm not trying to talk anyone out of matching transistors. I'm interested in understanding the answers to "matching what?" and "why?" Sounds like a great future article. But for now my next article is focused on objective measures predicted by shape of output and transfer curves in the context of load-line canceling, where some matching methods appear to me to be tenuously helpful at best. Maybe there is a general trend? I'm looking...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom