Variable feedback.
(if the voltage on the amp output node exceeds the output at the connection of Q11/13 by more than a diode drop, there's a feedback path from amp output to input stage. Part-time global negative feedback)
Last edited:
So, with tone controls circuit disengaged in normal operation, would you say that this circuit has global NFB and how much of it? When does this circuit become global NFB circuit? In practice, is there an fixed value of circuit parameters when it becomes ordinary global NFB circuit? When will such condition appear, at high volumes, when some tone control freak ODs with tone controls usage, etc?
Last edited:
No this is only a local feedback loop. This loop do not include the output transistors.
But the output stage has its on feedback loop from the output into inverting input of a M5218 opamp.
As for the D13-D15 try read this
Analog Devices : Analog Dialogue : Input Protection
But the output stage has its on feedback loop from the output into inverting input of a M5218 opamp.
As for the D13-D15 try read this
Analog Devices : Analog Dialogue : Input Protection
Ivane
-First GNFB loop is closed around entire input/VAS/predriver/variable tone control stage , automatically that Denon amp is not GNFB free amp.
-I have no internal schematic of M5218P OP-amp , but I believe that OP-amp not serve only as DC-servo but as OPS error correction circuit too , which I count as second GNFB loop .
-First GNFB loop is closed around entire input/VAS/predriver/variable tone control stage , automatically that Denon amp is not GNFB free amp.
-I have no internal schematic of M5218P OP-amp , but I believe that OP-amp not serve only as DC-servo but as OPS error correction circuit too , which I count as second GNFB loop .
Last edited:
Unity gain, non-inverting, power buffer. D13, D15 are protections against exceeding the max differential input voltage.
The preamp has NFB through the tone control, and is providing the overall gain. As such, there is no global network feedback, but two stages (preamp+unity gain power buffer) each with it's NFB loop.
IMO this amp was designed only to satisfy a marketing gimmick request. I can't see anything special in this configuration, compared to a regular amp.
Attachments
Last edited:
Some people like to define anything less than global feedback as not really being feedback. They also define global feedback as being from output point to input point - anything less is not global feedback and so not feedback. Hence an amp with lots of feedback can be described as free from feedback. This is strange, because the theoretically most excellent form of feedback is global feedback - anything in the output which is not in the input gets reduced. Anything else does less than this. In most cases the feedback is taken from somewhere just short of the output point (e.g. just before the Zobel network and output inductor) and fed back to something other than the input point (e.g. the other LTP input) - so some might say this is not global feedback and hence not really feedback. All this is just playing with words, because feedback is feedback however you do it.
Heck, I can do that- that's not what I said at all
The AR-1 was designed for an amplifier with a 7-ohm output impedance. That's not going to be a voltage source! IOW it did not change how amplifiers were designed at the time- it supported them. And if you look on the back you will see level controls just as I have described earlier.
Very true, but irrelevant for fashion audio and the associated marketoids. "They won't hear a d*mn thing anyway".
One remark: the long-debated generation of higher-order products from simpler ones in given stages occurs with cascaded stages too. For some reason this seems to bother people who peek less than global feedback. The old cliche about the dog chasing his tail comes in, despite virtually all examples lacking any significant delay to support it.
As suggested the overall topology may have been part of a marketing ploy. Probably the pitch was along the lines of no GNF and a nearly perfect output voltage follower.
As usual, given the same bill of materials, one looks for ways to do better.
Also, if the result is truly sensitive to loading enough to make everyday cables an issue, however doubtful that is given the L-R isolating output network, there may be so much loop gain in the output section already that it could be classified as a marginal design. One could go further to investigate if the characteristics of the op amp (or whatever is in the little triangle) were known. I haven't searched to see if that part is described anywhere.
The somewhat-similar approach of driving output and driver devices via the power supply terminals has been used a fair amount. It suffers from the effect of uncertainty in the op amp quiescent current drastically affecting the output devices' bias when the output stage is common-emitter (PNP emitters to V+, NPN to V-). Op amps usually have rather broad specs on Iq, even if they are there to satisfy the six sigma QA people. It can appeal as the output swing can go to the rails, almost---but saturate them in doing so and you are in a world of hurt from storage time. By the time one makes all of that work one looks back at the emitter followers and may realize that they are not so bad after all.
The synthesis of a given output resistance by current feedback is an interesting technique. When the equivalent resistance is close to a match to the load, it is remarkable how close to a true power-regulated output it is, given substantial load resistance variations. To do better one needs real multipliers in the loop.
Although hardly needed, I proposed once that such configurations be used for tube filaments. The surge currents for cold startup are much lower than when powering from voltage sources, and you don't have the potential thermal runaway issues powering from current sources. So you may have better control over temperature, although this won't help long-term degradation of emissivity.