View Single Post
  #4   Report Post  
Andre Jute
 
Posts: n/a
Default 5K OPT in place of 3K ?

wrote:
Hello,

I have an opportunity to have SE OPT 5K / 0,4,8 ohms. Big good ones
second hand, cheap price.
But i want to build 300B SE, usually it is more 3K OPT.

Is there a way to use them for a 300b amp. ?

Thank's
BR,
Luc D.


Consider your speakers, Luc, and their power requirement. How much
power do they really require for the sound pressure level you use?
Clearly, if you are interested in 300B, they will work with 7W or less,
which is what you get out of a standard sort of 3K OPT 300B amp.

Theoretically going from a standard 3K OPT TRX to one with a primary
impedance of 6K reduces peak SPL. It is not a big deal. If you are
designing so close that 3dB or even 5dB in your available transient
reserve becomes important, you should get more sensitive speakers or go
to PSE or PP. I use Lundahls on the 5K6 tap all the time and going from
3K3 to 5K6 dropped the output of High Voltage, High Current design from
7W6 to 4W8. My current KISS "Fiultra" is a Medium Voltage, Medium
Current (320V operating, 65mA, -65V) design to make it safer for
DIYers and to ensure longevity of the tubes; it is further limited by
the signal available from the single driver, so that it puts out 3W8. I
don't notice the difference. My speakers are Lowther horns, which can
be seen during construction and development on my netsite.

So what do I gain (except a safe construction and longevity) by giving
up half the power? The main answer is silence.

What happens is that a different impedance on the plate of the power
tube (any tube, actually) turns the transfer curve on the pivot of the
quiescent operating point (you are building a Class A amp, by
definition, if it is an SE amp, so the Q point is at the standing
current drawn constantly at zero signal--there is a chapter on my
netsite to show how to use the transfer curves or try the RDH or
Crowhurst). The higher impedance pivots the transfer curve
anti-clockwise, thereby making it flatter, raising the right hand side
further above the turning bottoms of the negative grid bias transfer
functions (the diagonal -Eg lines), and removing the end of the
positive going signal (the left hand side) further from the possibility
of grid current. The result is more linear power transfer, for which in
English we say: "Silence!"

The theoretical limit of load impedance increase, given infinitely
sensitive speakers of course, is set by a horizontal transfer curve, at
which point zero current change occurs and no power is transferred. A
more realistic theoretical aspiration is a loadline running perfectly
perpendicularly through precisely evenly spaced Eg-En-Ia curves (a
condition the 300B and very few other power tubes approach at all
closely); the limit here is set by how much current maximum dissipation
spec allows you to draw. In fact, the highest common load-impedance
pivots the loadline (which is the transfer function of that particular
OPT impedance with that particular tube and those particular operating
parameters) well past the perpendicular-to-transfer-curves to get
enough signal swing to get good power without running into the
nastiness where the curves fall over at the bottom and without
dissipating unconservative amounts of power. Thus, beyond a certain
point, which might be about 6-8K depending precise operating
parameters, a higher load will again lead to a less linear sound, i.e.
more distortion. Today it is widely considered that 5-6K is the ideal
compromise. (You may have heard about the notorious flame wars on RAT
and elsewhere. The winder and his pals assaulting me were in part
inflamed by my early advocacy of higher load impedances on DHT, which
undermined their commercial interest since the winder had only
old-fashioned low-impedance designs in his catalogue, at that time
nothing higher than 3K5.)
****
If you're going to the expense of 300B in the first instance, you
should already be in a position with your speakers where you do not
need to hog the tubes out for maximum power. Frankly, I now consider
2K5 (common 30 and 40 years ago, actually the max to be inferred from
the RDH recommendations!) a grotesque mismatch to 300B, a waste of good
tubes. Nor is 3K fabulous; for quite a bit less money if you shop
carefully you can build a better amp using trioded PP EL34 with film
caps everywhere except in the power section. You'll be amazed at the
quality of components for such an amp that you can buy straight out of
the RS catalogue. Someone wrote to me the other day, after I said good
amps cost roughly the same to build whatever the chosen topology, that
a careful costing of my T113 Triple Threat (PP trioded EL34), T44bis
KISS Populaire (standard three-stage 6SN7-300B, a very silent amp at
any power) and T39 KISS Ultrafi (what it says, WE 417A into 300B, a
low-power ultra-silent amp), all three of which he has built, reveals
that the T113 EL34, which I designed to be built out of the RS
catalogue, costs about three-quarters of the price of either of the
300B amps. According to him the price difference isn't in the tubes or
the transformers (Plitron/Amplimo for the PP amp and Lundahl for the
300B amps) but because the 300B amps depend on very expensive polyprop
caps in the power supply.

The best 300B until ten or fifteen years ago used 3K5, a very popular
load in Japan and among the Japanese-inspired Americans. It seems to me
that today even a really good commercial Japanese 300B amp starts at 4K
load (the PSE Miyabe from Triode Supply Japan, known in Australia as
the VPD3000, uses a 2K OPT for the parallel tubes). Your top-end DIY
audiophile generally starts with 5K, a great many of them like my own
fave of the Lundahls on 5K6 if my mailbag is anything to go by, and the
niche guys like Lucas Cant's 6K.

HTH.

Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"an unbelievably comprehensive web site" -- Hi-Fi News & Record Review