Complementary diff input with JFET and BJT cascode

[AndrewT]

Hi Bensen,
is that phase shown on the gain plot?
What is the vertical scale? I'm guessing 30degrees per div (90degrees at -3db).
This gives a total delay of about 20degrees from 15Hz to 10kHz = 5.6uS. Tell me I'm wrong because this seems very high.

 

[Bensen]

Hi Andrew,

It is 20degrees/div:
10Hz --> 10°
15Hz --> 7°
10Khz --> -5.4°
15Khz -->-12°

So I guess this is to high for an amp. What are decent values? And what can I do to make the phaseshift smaller?

This is very good you mention this, I would have overlooked it. I understand that this is important to get good and correct feedback in higher frequencys.
I don't understand all these terms yet, but am I correct whan I say that this is bad for slew rate issues and TIM distortion?

Thanks
Ben

 

[Bensen]

When I play with the values of the input filter:

LP: -3dB = 270kHz, 390P + 1.5K
HP: -3dB = 0.73Hz, 4.7µ + 47K

Results:

10Hz --> 5°
15Hz --> 3.7°
10Khz --> -2.83°
15kHz --> -4.74°
20kHz --> -6.7°

Aren't the -3dB points a bit to low and to high?

 

[Bensen]

I think I was a bit to quick with the TIM and slew rate issue.
The phaseshift is there due to the input filter. After this filter is the phaseshift between input and output 2° in a frequency range of 1Hz till 20kHz. So this is of no concern, wright?
Or am I wrong here Andrew? Please tell me.

Can somebody tell me whether I'm wright in post #120. Is the bandwith 110kHz?

Ben

 

[AndrewT]

Hi,
it appears most of the phase change is due to your necessary input filters. Since you are in sym mode try measuring without the filters.

I think you are looking for 500nS to 1500nS delay through the amp for audio frequencies.

Then put your filters back in with the -3db points about 10 times outside the audio band say 2Hz to 200kHz. Then make adjustments from there. Remember the NFB and PSU high pass time constants will be even lower than 2Hz.

 

[ppl]

I have played around with this type of topology using both a BJT cascoded to another BJT in complementary diff amp input stage and then using a pair of P and a pair of N ch Jfets cascoded to BJT and the best sounding jfet combination was the J176 and J113 cascoded to 2N5087 and 2N5088 I liked the J113/J176 a lot better than the 2SK170/2SJ74 as the Toshiba transistors sounded thin in the bass

 

[Bensen]

Hi PPL,

A few questions,

Did you also used a second diff stage?
What type of output transistors/mosfet's have you used?
How was the sound in general compared to what else you have heared?

I started a next thread after this one, where I've put the total circuit on. I haven't got the time yet to make it, but I will one day, after my house is ready.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=70431

Cheers
Ben