F5 with 2SK2013/2SJ313

Status
Not open for further replies.
Actually the peak output for these will be in excess of twice the bias,
depending on the amount of Source resistance degenerating the gain.

For purely square law devices operated push-pull (no resistance in series
with Source pin) the theoretical number is 4 times bias current. With large
values of Source resistance, it declines toward 2 times.

This being an imperfect world, you usually see un-degenerated Fets doing
about 3 times bias.
 
Mr Pass,
Thanks for explaining that - I got it the other way around. So if I am run I g 1.5amp bias in my amp, it can push about 3amps to as much as 4.5amps in class A through the speaker? Well that's good to know. If I am running 25v rails isthe max power about same as regular F5 at 25w? Do you see any issue with me changing the gain to 20dB?
 
I think in class A the peak current is about half of bias current. If I did this right it means you need 0.45amps for 5w into 8ohms. These are rated at 1A and should operate probably no more than 500mA ea to be safe. So having two should give you room for peak 500mA into speakers. Rail voltage needs to be maintained so that there is enough room for 6.3v rms. Maybe +/-12v could work?
P = IV = I²R = V²/R, for unchanging DC voltage and current.
For AC waveforms this becomes:
P = IrmsVrms = Irms²R = Vrms²/R
and for a sinewave it can also be expressed using peak value of the waveform:
P = IpkVpk/2 = Ipk²R/2 = Vpk²/R/2

using this last P=Ipk²R
we see that 0.45Apk will allow 0.81W to be dissipated in an 8r0 dummy test load.

If you want 5W then rearrange the equation to become Ipk = sqrt(2P/R) =1.18Apk into 8r0 needing a drive voltage of 8.94Vpk
 
Last edited:
P = IV = I²R = V²/R, for unchanging DC voltage and current.
For AC waveforms this becomes:
P = IrmsVrms = Irms²R = Vrms²/R
and for a sinewave it can also be expressed using peak value of the waveform:
P = IpkVpk/2 = Ipk²R/2 = Vpk²/R/2

using this last P=Ipk²R
we see that 0.45Apk will allow 0.81W to be dissipated in an 8r0 dummy test load.

If you want 5W then rearrange the equation to become Ipk = sqrt(2P/R) =1.18Apk into 8r0 needing a drive voltage of 8.94Vpk

I don't think this is consistent with what Mr. Pass said:

Actually the peak output for these will be in excess of twice the bias,
depending on the amount of Source resistance degenerating the gain.

For purely square law devices operated push-pull (no resistance in series
with Source pin) the theoretical number is 4 times bias current. With large
values of Source resistance, it declines toward 2 times.

This being an imperfect world, you usually see un-degenerated Fets doing
about 3 times bias.

using this last P=Ipk²R
we see that 0.45Apk will allow 0.81W to be dissipated in an 8r0 dummy test load.

P=0.45A squared x 8ohms = 1.62w
 
I want to add the regulated PSU that Juma designed:

post-20-074595800%201275676845.gif


It's a very simple circult and I am hoping that it helps reduce the hum I am getting from rail ripples under load. Can someone explain to me how this circuit works and where the 20V reference is coming from? If I want to run it at say, 22V what do I need to adjust?
 
So if I am run I g 1.5amp bias in my amp, it can push about 3amps to as much as 4.5amps in class A through the speaker?

If I am running 25v rails is the max power about same as regular F5 at 25w?

Do you see any issue with me changing the gain to 20dB?

1) With a .47 ohm Source resistor, probably about 3.5A pk

2) Yes, 25W

3) 6 dB more gain = twice as much distortion, half the damping factor.
 
Capacitance Multiplier

I implemented Juma's capacitance multiplier on a separate heatsink. I have to say that it was simple and I works very well. I am getting 22v rails no load and 19.5v rails under full load with 1.5amps bias current. Ripple measured with AC volts on Fluke is less than 1mV as shown below. Quite an amazing performance from the previous 40mV ripple coming from the CRC filter.

post-20-074595800%201275676845.gif


The circuit is so simple, I easily did it with P2P right on top of the IRFP240/9240's.

574333d1476267068-f5-2sk2013-2sj313-f5-j313-k2013-capacitance-multiplier-works.jpg
 

Attachments

  • F5-J313-K2013-Capacitance-Multiplier-Works.jpg
    F5-J313-K2013-Capacitance-Multiplier-Works.jpg
    141.8 KB · Views: 27,177
I cannot find the post where Juma explained the "mkII" version of his amp any more, but I still have the schematics. It was May 2011...
 

Attachments

  • Toshiba F5-sch_by Juma_new2011.gif
    Toshiba F5-sch_by Juma_new2011.gif
    9.1 KB · Views: 878
  • Toshiba F5-psu_by Juma_new2011.gif
    Toshiba F5-psu_by Juma_new2011.gif
    10.1 KB · Views: 589
Status
Not open for further replies.