Reply to message

John Byrns wrote:

I have scanned "The Transistor Amplifier" article I mentioned in an
earlier post, and placed the resulting files on my web pages at the
following URLs:

http://users.rcn.com/jbyrns/p23.jpg
http://users.rcn.com/jbyrns/p24.jpg
http://users.rcn.com/jbyrns/p26.jpg

This 1959/62 design for a KISSASS amplifier could use a few updates, like
silicon transistors, a higher input impedance, and a lower output
impedance or higher "damping factor". Accomplishing these improvements
would probably require adding another transistor or two, but the article
serves as a good starting point.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/


Thanks. This looks more like it, especially figure 2. You would get
enough gain with a modern transistor, I think. A darlington pair looks
good but in any event, if you can restrain yourself from the control
freakery that overcomes all of us the moment we sniff silicon, once
you have the gain you won't need to add too many correction circuits,
so the parts count could be impressively low.

But I do think the input impedance will be a bigger problem than you
think. Patrick will shortly be dancing around here with a whole
'shirtload' of NFB in his pocket that he wants to give away to a
caring, sharing person! Whatever you will require will make the 14dB
you say I have inside my 300B as natural feedback look like
chicken-feedback.

To match a 300B amp like the KISS really well, the transistor Class A
amp must work with direct input from a CD player. My Quad 67 CD
player, for instance, demands a minimum 10K load at all times from the
next device, in this case is the pot on the amplifier.

I'm not overly bothered about a transformer on the output of the
KISSASS. Or I would be happy to relax the requirement so that the
KISSASS is a headphone amp standing beside the KISS 300B to drive,
say, the common 32 ohm headphones like my open-back Sennheisers.

Now we're cooking with gas!

Andre Jute

Andre Jute wrote:

Reply to message

John Byrns wrote:

I have scanned "The Transistor Amplifier" article I mentioned in an
earlier post, and placed the resulting files on my web pages at the
following URLs:

http://users.rcn.com/jbyrns/p23.jpg
http://users.rcn.com/jbyrns/p24.jpg
http://users.rcn.com/jbyrns/p26.jpg

This 1959/62 design for a KISSASS amplifier could use a few updates, like
silicon transistors, a higher input impedance, and a lower output
impedance or higher "damping factor". Accomplishing these improvements
would probably require adding another transistor or two, but the article
serves as a good starting point.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

Thanks. This looks more like it, especially figure 2. You would get
enough gain with a modern transistor, I think. A darlington pair looks
good but in any event, if you can restrain yourself from the control
freakery that overcomes all of us the moment we sniff silicon, once
you have the gain you won't need to add too many correction circuits,
so the parts count could be impressively low.

But I do think the input impedance will be a bigger problem than you
think.

The input transistor in fig3 powers the power transistor from its emitter, but
there is considerable
collector resistance.
It doesn't quite look right the way it is set up.
But one could use a pnp mosfet to drive an npn bjt
in a normal cascaded gain pair, and have normal global FB
applied to the mosfet source circuit like one would with a power tube and an
input tube.
There would then be no problem with input impedance, or drive level,
and the NFB would reduce the miller effect which isn't too good in some
smaller T0220 types.

Such a pair is I think called a sziclai pair, maybe I am wrong.

Patrick will shortly be dancing around here with a whole
'shirtload' of NFB in his pocket that he wants to give away to a
caring, sharing person! Whatever you will require will make the 14dB
you say I have inside my 300B as natural feedback look like
chicken-feedback.

A bjt used in class A to get about 15 watts isn't too hard to do
but the thd will be rather like a 6V6 with no NFB, pretty darn awful....
But 20 dB of FB would allow the first 3 watts to not be too bad
as is the case with a 6V6 if you only wanted 1.5 watts from such a tube.

I'd give away 20 dB of NFB, but any extra amount up to and including a
shirtload
will cost yer somethink.

I have a pair of class A mosfet amps I built a few years back and they have
a diffgain pair as the input driver stage, and PP class A outputs.
There is a total of 22 dB of NFB.
Rin is high, Ro is less than 0.5 ohms, and thd is about
0.2% at 45 watts, declining to SFA at 3 watts.


Class A SS amps should never need more than 20 dB of NFB.

If the amp is to be SE, then the thd will be greater than a PP amp,
since you'll get some 2H.
If the gain of the power bjt is too great, some local current FB
can be used in its emitter circuit, maybe 0.47 ohms would be about right...


To match a 300B amp like the KISS really well, the transistor Class A
amp must work with direct input from a CD player. My Quad 67 CD
player, for instance, demands a minimum 10K load at all times from the
next device, in this case is the pot on the amplifier.

I'm not overly bothered about a transformer on the output of the
KISSASS. Or I would be happy to relax the requirement so that the
KISSASS is a headphone amp standing beside the KISS 300B to drive,
say, the common 32 ohm headphones like my open-back Sennheisers.

I think the OPT is a very sensible way to do it.
Otherwise one has to have a CCS, or choke, and a cap coupling to the load...
The OPT keeps DC out of any speaker, and the OPT isn't as difficult as one
would think,
using fewer turns but much thicker wire.
If the load is 6 ohms, then the primary L need only be 100 mH for a -3 dB point

at 10 Hz.
The core size would still need to be the same large size as any other decent 15
watt SE transformer,
so that the Lp can be achieved with a few turns of thick wire, since Ic would
be about 3 amps.
A special tap off the sec for the heaphones and a dummy R load would be correct

for when phones are used.

Some sort of diode limiting of the collector voltage swing is needed with an
OPT due
to back emfs that could cause the power bjt to have its Vce exceeded.
The power bjt need only be rated for about 50v, since the lower voltage types
usually
are better suited to low voltage apps and class A.

Now we're cooking with gas!

Use a big frypan as a heatsink, and cook with transistors.

Patrick Turner.



Andre Jute