NFB circuit not typical?

[roger2]

The NFB circuit in this amp does not look quite like what I have seen in other amps.

Re47 at 10KΩ and bypassed by Ce13, then Re21 at 470Ω, all looks typical. But in other amps I have usually seen a large value cap shunting Re21 to ground. Why does this amp not have that cap?

And what is the purpose of the two circuits that are marked on the schematic, both of which start with a cap on the NFB return line? Marked in RED: Ce21, Re49, Re53, and De9. And the similar, but opposite polarity, circuit marked in GREEN.

Do these two circuits have anything to do with NFB?

Thanks in advance for any enlightenment...

 

[ostripper]

Same as my "spooky" amp using voltage based feedback. Its SERVO'ed !

Servo replaces the blocking cap.
Edit - they might be doing it through the positive input - but it has to be
for this type 3 stage amp.
OS

 

[rayma]

But in other amps I have usually seen a large value cap shunting Re21 to ground. Why does this amp not have that cap?
And what is the purpose of the two circuits that are marked on the schematic.

This amp has only an output offset trim in the input diff amp. Presumably the circuit is stable and can do without the usual nfb cap
to ground that reduces the DC gain to unity. The other circuits are current limiters for the output stage, one for each polarity.

 

[ostripper]

This amp has only an output offset trim in the input diff amp. Presumably the circuit is stable and can do without the usual nfb cap
to ground that reduces the DC gain to unity. The other circuits are current limiters for the output stage, one for each polarity.

This is my "Kypton V" , 2nd stage amplifies any first stage unbalances.
Actually , it is the 2'nd stage beta as the multiplier of the the first stage's
error.

Compared to your standard LIN (blameless) , nearly impossible to get
a native low offset through just matching.
The dual FET is a matched siingle package , this might offset the above
a little.

Typically , most Japanese (a common 70's/80's topology) amps like this used
an inverting servo.
BTW , this is said to be one of the most musical (best sounding) of
input stages (Sansui/Pioneer).

OS

 

[vzaichenko]

Also, having jfets as a first diff pair, results in virtually zero input currents, so no need to match the resistor networks' impedances between the inputs, having good balance anyway. Matched single packaged jfets - even better.

 

[sregor]

And what is the purpose of the two circuits that are marked on the schematic, both of which start with a cap on the NFB return line? Marked in RED: Ce21, Re49, Re53, and De9. And the similar, but opposite polarity, circuit marked in GREEN.

Do these two circuits have anything to do with NFB?

Thanks in advance for any enlightenment...

They are there for current limiting purposes - they increase the max current level on the positive outputs as the signal goes positive, and on the negative outputs as the signal goes negative.

 

[roger2]

Ok, thank you all. That helps my understanding significantly.

By the way, this amp is a Kenwood KA-7100. Interestingly, I have another amp that uses the same dual jfet input (uPA63H) and the exact same NFB configuration and resistor values. The only difference being the value of the ceramic bypassing the larger resistor. (Sherwood S-602CP)


As a follow-up question: Would there be any risk of introducing instability or oscillation if I upgrade the NFB resistors to quality metal film and the ceramic bypass to a current NP0 ceramic?

The original NFB resistors are generic 1/4w carbon film. Not sure about the original ceramic, or whether NP0 even existed back then.

 

[sgrossklass]

I'd be pretty sure it's NP0. Low-K ceramic capacitors constituted a major advance in receiver construction around WW2, as they could be manufactured in various temperature coefficients for precise temperature compensation of oscillators, hence allowing for much improved frequency stability. They became available in Germany in about 1937-ish, if memory serves. (Read more if interested.) Things ceramic became big in Japan in the 1960s and 1970s, and if you were to buy some ceramic IF filters for an FM tuner or whatnot, they'd be from either Murata or (less commonly) Toko.

 

[roger2]

Thanks sgrossklass. I have asked the several times about when NP0 ceramics came into common use. No one seems to want to answer that question. Interesting link too. It is nice to think about how, when, and where things developed that today we take for granted.

I am still not crystal clear. If a ceramic cap came from Murata or Toko, that does not necessarily mean it was NP0 as they would likely have had a range of different types at that time.

Do you think that ceramic caps in mid-1970's Japanese amplifiers would have been NP0? Or, even if they were NP0 back then, could today's NP0 ceramics possibly be better in some way?

 

[rayma]

I have asked the several times about when NP0 ceramics came into common use. It is nice to think about how, when,
and where things developed that today we take for granted.

Surprisingly, Wikipedia has an excellent and detailed article on this subject...Ceramic capacitor - Wikipedia, the free encyclopedia

 

[roger2]

It is a nice article. Wikipedia is usually my first stop when researching anything.

I have read this article previously but was not able to find the specific answer I was looking for. But...I shall read it again later this evening, more info may have been added. Thanks.

 

[roger2]

Sincere thanks to all who contributed here Very interesting links offered and I did read them.

I think I should start a new thread as my OP inquiry has been well addressed. I have more questions about ceramic caps in NFB loop, but it might be best to start that as a new topic.

 

[vzaichenko]

In general, ceramics are good to use for HF compensation, HF filtering, etc. - outside of the audio range. Ceramics - as well as tantalums - have got dependency of their capacitance from the voltage across them. In other words, they introduce noticeable non-linearity. That's why they are not within the audio range in signal path, including NFB. Dielectric absorbtion effect adds to this. Some films are good fot signal path. See this one here: >Picking Capacitors - Walter G. Jung and Richard Marsh<

Cheers,
Valery

 

[sgrossklass]

Actually ceramics vary widely. The caps used for VAS feedback would be NP0/C0G type (as most anything below a few hundred pf would be), which have some of the lowest nonlinearity and dielectric losses of anything.