This is a rather nice implementation. I would add emitter resistors to Q1 and Q2, use a KSA1381 for Q7, remove Q3, C11 and C13, reduce the value of R30 and try to omit C9.
Hello.Thanks for your answer.The schematic is from a blown rotel RA-913 and the parts are the ones i used to replace the originals.
Emitter resistors to Q1,Q2 will be very dificult to accomodate due to PCB layout being very tight at this point. the current source around Q3 will stay as is,regarding Q7 being replaced with a KSA1381 i dont know if KSA1381 it is suitable as a driver due to its low Ic max current rating (100mA).At the moment i am using 2SA1837-2SC4793 as drivers.
D5,D6 are mounted on main heatsing fot thermal tracking.Bytheway what's the purpose of C11?
I dont know if C9 miller compensation cap should be ommited or maybe increased to something like 10p or so?
Emitter resistors to Q1,Q2 will be very dificult to accomodate due to PCB layout being very tight at this point. the current source around Q3 will stay as is,regarding Q7 being replaced with a KSA1381 i dont know if KSA1381 it is suitable as a driver due to its low Ic max current rating (100mA).At the moment i am using 2SA1837-2SC4793 as drivers.
D5,D6 are mounted on main heatsing fot thermal tracking.Bytheway what's the purpose of C11?
I dont know if C9 miller compensation cap should be ommited or maybe increased to something like 10p or so?
Thanks for the quick reply and modification bimo.what else can i use instead of drivers 2sc5171 and 2sa1930?I have 2SC4793,2SA1837,2SD669A,2SB649A
C11 averages the voltage difference between the bases of the drivers over a long time. Voltage difference between driver bases doesn't need to change instantaneously. Often 10-47 uf is used here.
This topology is inherently unstable. The number of 6 harmful phase compensation capacitors may be reduced in conjunction with other measures taken. I especially dislike the presence of Q3.
An emitter resistor would be a better choice than C11. Emitter resistors greatly reduce instability.
Neither the 2SB649 nor the 2SA1837 are good voltage amplifiers. Anyway, maybe you should keep the 2SA1837. I would use the KSA1381 at not much more than 10 mA.
I thought that either C9 or C10 could be adequate.
An emitter resistor would be a better choice than C11. Emitter resistors greatly reduce instability.
Neither the 2SB649 nor the 2SA1837 are good voltage amplifiers. Anyway, maybe you should keep the 2SA1837. I would use the KSA1381 at not much more than 10 mA.
I thought that either C9 or C10 could be adequate.
You can use 2SC4793,2SA1837. But in real amplifier 2sc5171 and 2sa1930 is better. I never use 2SD669A,2SB649A because it hard to get in my country.
This is the year 2020, and the only good reason you might want to build a new quasi-comp amp is that you are rebuilding one from old parts.
There are a couple unnecessary peak voltage losses like R12 and the placement of R23.
A newer amp should have a degenerated LTP, a current mirror and a Darlington VAS.
You mentioned this was a blown Rotel. That should tell you that you need current limiting, at least. Newer power transistors with better SOA will help but +/-46VDC into a short or even 4 Ohms will easily blow a single set of power transistors. I would use two pair of power transistors for a voltage that high, with the possible exception of a hearty set of MOS power transistors.
R18 and R19 should never be larger than 100 Ohms. Otherwise the outputs potentially float a massive shoot through current when driven hard at ~10KHz. Complimentary outputs are easily cross-coupled to prevent this problem but there is no good way to cross-couple a quasi output.
You don't need to include everything from Bob's book but you should at least start with a basic "Blameless" circuit which you can see on Douglas Self's web site.
There are a couple unnecessary peak voltage losses like R12 and the placement of R23.
A newer amp should have a degenerated LTP, a current mirror and a Darlington VAS.
You mentioned this was a blown Rotel. That should tell you that you need current limiting, at least. Newer power transistors with better SOA will help but +/-46VDC into a short or even 4 Ohms will easily blow a single set of power transistors. I would use two pair of power transistors for a voltage that high, with the possible exception of a hearty set of MOS power transistors.
R18 and R19 should never be larger than 100 Ohms. Otherwise the outputs potentially float a massive shoot through current when driven hard at ~10KHz. Complimentary outputs are easily cross-coupled to prevent this problem but there is no good way to cross-couple a quasi output.
You don't need to include everything from Bob's book but you should at least start with a basic "Blameless" circuit which you can see on Douglas Self's web site.