[morsel]

Or go with a pair of nice MOSFETs like the IRFZ24N and IRF9Z34N.

 

[kevin gilmore]

You can certainly take off power transistors and reduce the standing
power and the amount of power supply. But it won't sound the same.
Will it sound similar for many headphones, probably yes. Will it be just
the bass that is effected, i don't think so. The damping factor is just
about flat across the entire audio band and beyond. Most other
amps out there have a damping factor curve that shows a significant
increase in resistance as output frequency rises.

Lets say i have one output transistor that outputs .5 amp.
Lets say i have 4 output transistors that output .125 amp each totalling
.5 amp. Do these two sound the same, certainly not. Typically the
silicon area of the 4 transistors because of thermal considerations will
be less than the silicon area of the one big transistor. This generally
points to less capacitance and a higher Ft. What about lead inductance?
4 transistors in parallel are going to have about 1/4 the lead inductance
of one device. This definitely matters too.

The absolute best transistors i can find for my uberamp have an Ft of
30mhz. While significantly better than all the old stuff out there, there
is no way these sound the same as transistors with a Ft of 120mhz.

I'm sure its going to be a few months before people build and then
listen to this thing. But once they do the oooh's and aaaah's will flow
like water.

Those fets might be great as pass elements for the power supply for
this unit, but certainly not as output devices. Rise and fall times
are not balanced, significant gate offset and capacitance and in general
and equivalent Ft of probably less than 10mhz...

 

[aos]

I don't question the validity of using multiple transistors, my question was more related to where would be the sweet spot. I would definitely use at least 2, but the question is whether going from 2 to 4 gives audible improvements, and past that, going to 8 and so on. I guess this is best left to experimentation.

 

[kevin gilmore]

The sweet spot for the output transistors i'm using is somewhere
in the 80 to 100ma range each.

 

[morsel]

Servo circuit added and minor alterations made to the schematic.

 

[pooooo]

Quote:

Originally Posted by kevin gilmore You can try the lm431 but my guess is that when you do you will find that the servo no longer functions. The servo depends on the very slight changes in voltage across the led when the current changes. Yes the second stage transistor gets warm, but not out of spec. You can certainly try other transistors there if you like. Like one of the output transistors. I've had my unit on now 24 hours a day since last saturday morning and can touch every component and nothing is getting too hot.

Hi Kevin,
Thanks for your explanation. The servo is very innovative but I personally don't want that servo in my amp. Just like you, I don't like capactors in signal paths. I think LM431 should be less noisy than LEDs, though such high-frequency noise might not be audible.
Are you planning to build a Dynahi with the power supply topology of the previous lower-power version? I have several pieces of OPA541AM at hand and I'm wondering if they are suitable for such usage.
BTW, may I ask about your room temperature? It's about 90 degree here at my place during summer, and my original amp (30W quiescent) gets really warm even with two heatsinks four times as big as yours. Almost untouchable unless fans are turned on.

 

[morsel]

Servos like this one are great. I don't like caps in the signal path, either, but this does not count, as its purpose is to integrate the signal to correct for DC offset. This is not the same thing as using a coupling cap on the input or output, in which case the entire signal passes through the cap.

 

[Prune]

This servo reminds me of one here: http://www.tubecad.com/articles_2002...ers/page5.html

 

[kevin gilmore]

opa541 are exactly what i'm planning on using. However the
pre-regulator needs to be some kind of pass fet because an
lm317 can't handle the voltage, and you have to keep the
pre-regulated rails less than +/-40

Without the ventilation on even in the winter, the media room
fully powered up can easily hit 85. Usually i leave the fan on
and this takes care of the problem. In the summer the ac is
always on (summer starts about april) and then it is no problem.

The projector alone puts out 650 watts in heat. The amplifiers
together put out depending on volume up to 2000 watts.

 

[rsaavedra]

Quote:

Originally Posted by kevin gilmore Caps. 33 pf (2) 470 uf/50v (4) .1uf (4) .47uf (2)

Kevin, other than the 470 uf/50v cap, how about the voltage spec for the others? Think I'll start getting the parts before the boards.

 

[pooooo]

Quote:

Originally Posted by kevin gilmore opa541 are exactly what i'm planning on using. However the pre-regulator needs to be some kind of pass fet because an lm317 can't handle the voltage, and you have to keep the pre-regulated rails less than +/-40

I think LM317/337 are OK for such voltage. I even tried to build a PS unit which outputs 250VDC with LM317 for a tube amp. LM317 works fine as long as the voltage difference between its input and output does not exceed 40V. For high voltage applications, just add some diodes to protect it from transient status such as power-on surge and power-off.

 

[kevin gilmore]

the .1uf need to be 15 volts or better
the 33pf would need to be 30 volts (minimum you can buy is probably 100v)
.47 needs to be 15 volts but to fit in the hole the
exact mouser part is 75-mkt1822447014 which is 100v

 

[gavinbirss]

I have seen from previous posts by master Gilmore that when he posts links to
hires photos of his circuits that some people hint to the fact that he has to
learn how to use photoshop in order to reduce the size of the files.
This will obviously reduce the quality and the use of a his 15 megapixel camera.

You can't give a child candy and not expect his to eat it.
(my point is not that Mr Gilmore is a child ....)

I am refering to picture file : dynahif2.jpg

Why does only three led's light up ?

Is the fourth led located on the right hand side, camera shy or dead ?

Why does the three led light up anyway and indicate DC being integrated by
the DC Servo , considering that there is no input ?

Is the DC Servo going MAD ?

humour my please since I am only a Kevin Gilmore follower and you, Kevin are the leader.

Great work ...


"The problem is the range of impedances this amp has to drive.
From 32 to 300 ohms. For low impedance phones you need the
bias turned up to keep the unit in class A for the entire audio
signal. For high impedance phones you could definitely turn down
the bias and still keep it in class A."

How would one go about changing the bias on the fly.

Does your preamp have a stepped attenuator and how many steps is needed for
a 32 ohm load (grado) headphone to produce a dynamic (medium to loud) level.

How much is the gain in db for your prototype to power 32 ohm headphones.

Gavin

 

[rsaavedra]

Bumping this, wondering about:
1) Group buy of the Dynahi boards
2) Whether the Dynahi fat edition PCB's will be available any time soon.
3) Any news on the power supplies designed specifically for the Dynahi?

 

[Sbear]

Regarding a pcb group buy- count me in.

A question regarding these condor power supplies- Are they any good?
I mean it doesn't seem like you could build something from scratch for much less than 40 bucks a pop, even with less than best transformers and res. caps. So are these power supplies worthy to be used with this amp or they just to get by until Mr. Gilmore comes up with his own design?