Can you improve these amplifiers?

here you have Pioneer circuit cascoded, RIAA correction should be ommited, with proper transistors you should get BW>2MHz
1687521374297.png
 
Thanks for all the leads: I'll have some work testing all of them, in isolation or combination.
Regarding the distortion of the input stage, in its environment, it is ~4ppm. A bit high to my taste, but a very minor contributor with my current options.
I don't intend to include everything in a GNFB, it would be too demanding. Having individual blocks with a low enough THD looks more sensible
 
here you have Pioneer circuit cascoded, RIAA correction should be ommited, with proper transistors you should get BW>2MHz
View attachment 1185820
The Pioneer circuit seems to use regenerative bias on the input diamond, which is something I was going to suggest. I'm not sure how well it works at 1MHz? They also seem to be obsessed with reverse bias clamping diodes?
 
The Pioneer circuit seems to use regenerative bias on the input diamond, which is something I was going to suggest. I'm not sure how well it works at 1MHz? They also seem to be obsessed with reverse bias clamping diodes?
maybe too much clamping...
attached asc file for playing
cheers
Pawel
 

Attachments

Thanx but I don't have models for BF469 and BF470. Perhaps I could sub BD139/140? It would be great if you could share your models.
Hi, you can use any HF TO126 device with high enough VCE and low capacitance,

.MODEL BF469 NPN(IS=7.974E-15 ISE=2.266E-16 ISC=4.33E-12 XTI=3
  • BF=122 BR=6.235 IKF=0.01029 IKR=0.02746 XTB=1.5
  • VAF=25.51 VAR=19.43 VJE=0.4581 VJC=0.197
  • RE=0.3814 RC=0.439 RB=1 RBM=0.5 IRB=1E-06
  • CJE=1.742E-11 CJC=5.045E-12 XCJC=0.1041 FC=0.8555
  • NF=0.993 NR=0.999 NE=1.18 NC=1.397 MJE=0.3092 MJC=0.1947
  • TF=7.073E-10 TR=1E-08 PTF=0 ITF=0.1495 VTF=6.144 XTF=289.5
  • EG=1.11 KF=1E-9 AF=1 MFG=PHILIPS)

.MODEL BF470 PNP(IS=9.124E-15 ISE=1.672E-15 ISC=2.139E-13 XTI=3
  • BF=198.2 BR=1.256 IKF=0.13 IKR=0.1 XTB=1.5
  • VAF=465.9 VAR=13 VJE=0.8484 VJC=0.6298
  • RE=0.635 RC=1.42 RB=5 RBM=0.5 IRB=1E-06
  • CJE=1.447E-11 CJC=8.483E-12 XCJC=0.619 FC=0.99
  • NF=0.9904 NR=0.99 NE=1.527 NC=1.08 MJE=0.3884 MJC=0.4561
  • TF=1.38E-09 TR=1E-32 PTF=0 ITF=0.065 VTF=2 XTF=21.78
  • EG=1.11 KF=1E-9 AF=1 MFG=PHILIPS)

anyway I found this circuitry has too much distortion @1MHz for Elvee's demands, and is little peaky at this frequency.
cheers
Pawel
PS for 1KHz it is excellent
 
I am still trying to improve my amplifiers, and the best candidate so far is the Groner-based one, but I am faced with a strange behaviour: when the compensation cap C1 is too small (12p), the THD is relatively good (0.003%), and logically the FR has a nasty peak due to insufficient margins.
If the cap is increased to 22p, the FR is ~flat, and the THD increases, which is to be expected. What is surprising is the magnitude: it shoots to 0.015%, a five-fold increase, when the cap value is not even doubled.
In addition, the circuit tolerates an output-inclusive scheme, which means that increasing the compensation should have a minimal effect on the THD.
Can someone explain that paradox?

1688289442814.png
 

Attachments