zen without feedback

[Blues]

Current at 8ohms

 

[Blues]

5Wrms into 8ohms @10kHz...

 

[apassgear]

I think you need to add one more 7.5 v zener to the string, you only have 30 volts with 4 of these, less 4 volts that the pass mosfet is going to drop. As the schematic stands you will have about 26 Volts, not the expected 33 as shown in the schematic. Other than that looks good

-D

You are right, unregulated is 33V but regulated comes to 26V. Have a closer look to the schematic.

I'm still somewhat perplexed as to Nelson's choice of 25V rail. The only reason I can think of is that he wanted a higher bias.

As Nelson said he is using 2.7 amp bias, that will come to some 34 W per device which is what he prefers for commercial products.

As the circuit stands now:
I estimate some 40W per device (maybe less). But that should be measured and adjusted as to ones taste by changing R1 and R13 values.

I estimated some 20 to 25W dissipation for the regulator so be prepared for a total dissipation of 100W for each channel.

 

[apassgear]

Here's more or less apassgear's F2 circuit with some ckt sims....+/-9Vmax, 2.8A bias, 5Wrms into 8ohms...just replace transistors with your favorite fets.

Hey Blues, you overdid yourself, thanks for your input.

Any particular reason for changing the the bias resistor value? and other values like source resistors?

 

[Blues]

Tony,

Thanks! If you look closely at the sim schematics, the ref voltage across the Vb-e of the bipolar trans is about 0.79V so to lower the bias to about 2.8A I used 0.28ohms power resistors and to match NP's F2 specs. Of course you'll replace that to whatever resistor you feel will match your available heatsinks.

Allan

 

[apassgear]

Tony,

Thanks! If you look closely at the sim schematics, the ref voltage across the Vb-e of the bipolar trans is about 0.79V so to lower the bias to about 2.8A I used 0.28ohms power resistors and to match NP's F2 specs. Of course you'll replace that to whatever resistor you feel will match your available heatsinks.

Allan

Does this means if we lower the source resistor to 0.22 Ohms we could expect problems with the CCS?

 

[Blues]

As long as your ref voltage in the CCS is about 0.6V, 0.22ohm would be good for 2.7A of bias. As NP would say, adjust to taste.

 

[apassgear]

As long as your ref voltage in the CCS is about 0.6V, 0.22ohm would be good for 2.7A of bias. As NP would say, adjust to taste.

Thanks Allan,

Though it seems odd that 0.280 Ohms will render 2.8 Amps of bias. I was shooting for some 3A.

 

[Blues]

With sim you can put in any value you like and you won't be limited to the standard resistor values. I was too lazy to create the specs for the usual semiconductor components used in an Aleph or Zen. Anyway...

Ibias = Vb-e/R = 0.79v/0.28 = 2.8A

 

[liberty]

any idea/schematic for Zen lite without feedback?

 

[Dozuki]

Replace the current source and associated parts with a light bulb. Look at the earlier posts, what you are asking for has been discussed with a couple different schematics.

 

[MikeW]

Has anyone built this amp? How does it sound?

 

[Nelson Pass]

Well of course, I have. I liked it quite a bit, and the lamp
glow is lovely on cold winter nights. A pair of 300W 120V
bulbs gives a damping factor slightly less than 1 at about 22
ohms each. You can play quite a few games with the number
of bulbs and the DC voltage setting.

You will recall that the JFET amplifier piece used this, and the
cascoding works great for Mosfets as well.

 

[MikeW]

Anybody else?

 

[tlf9999]

The way to go on this is to make P1 and P8 very high values,
greater than 200 Kohm or so, and use a smaller (film) input
capacitor. This will allow some feedback for DC biasing, but
will not allow significant feedback for AC purposes, as the
impedance of the signal source will be very small compared to
those values.

what if you put a large enough inductor in serial with R8? That way, it still has DC feedback but AC is blocked by the inductor.

Or maybe split R8 into a RCR network.

 

[mastertech]

tlf what he has said has solved the problem without further
complexities that are prone to create more problems, however
your suggestion can work, how good is for you to find out


cheers