On Dec 3, 10:02 pm, Andre Jute wrote:
On Dec 3, 2:01 pm, Patrick Turner wrote:



Ian Thompson-Bell wrote:

Don Pearce wrote:
On Sun, 02 Dec 2007 13:51:44 +0000, Ian Thompson-Bell
wrote:

I have being playing around with plate characteristics graphs, looking
at how basic parameters (gm,rp and mu) vary with operating point. The
most frustrating thing about this is that the published curves all
extend to rather large plate currents and voltages. For my preamp I am
working at plate currents of less than 5mA and the curves here are
rather cramped. Is there any source of curves that are expanded for the
lower plate current region?

Cheers

Ian

It's a long, boring Sunday afternoon. Perfect time to do a set of
measurements like this and post the results somewhere. If this is an
area of valve performance that is poorly documented it would be
interesting to know what kind of variation there is between samples.

d

Been there, done that, wrote the spreadsheet.

Some months ago I ran 32 different 6AU6 tubes in triode mode and
measured gm, rp and mu at an Ia of 3.8mA and Vp of 100V. I have the
results in a spreadsheet if you are interested.

I don't really want to repeat that work every time I decide to use a
different tube. What I don't understand is why the manufacturers data
extends so far away from normal operating points such that the data at
the operating point is of poor resolution - but I guess most them aren't
around to ask. The best data for the 6AU6 in triode mode I have lists Ia
up to over 30mA and Vb to over 400V. OK so you load line is likely to
hit the Ia axis at twice your quiescent, but if that is say 5mA and Vb
is 250V then you are trying to read info off the graph in the bottom
left hand eighth of the graph.

Hving said that, I have just found a similar graph for a 12AT7 in which
Ia only goes up to 5mA - much more sensible - so maybe its just this
particular tube.

Ian

There are very few data sheets which are inadequate as you suggest.

Search and yee shall find!

Otherwise, simply make your own data sheets.

Ra, µ, and gm can all be easily measured using 2 different load values
and for say 20Vrms output on the higher value load of say 68k for 6AU6.

As long as THD 4% then your results will be fairly accurate enough.

µ = gm x Ra for all tubes and gain, A, = µ x RL / ( RL + Ra ) for all
tubes.
Use an accurate DVM at 400Hz sine wave, simple.

By simple high school algebra, you can do the rest.

Patrick Turner.

Nah, all those measurements and plotting besides is a six-Tylenol job.
Ian can do this very simply in a single shot, and get a curve, or many
curves if he wants to be fancy, complete with grid markings to plot
on. All he needs is a filament tranny and a socket and a distribution
strip screwed to a piece of scrap ply, a single resistor in the
cathode circuit, his scope and an amp with a working power supply to
tap off, plus a variac to get the right voltage if it is different
from what the amp PS supplies. The B+ goes in to the anode with the OV
attached to the grounded side of the fil. The negative grid bias goes
into the grid in series with the fil. This is most easily done by
attaching the negative end of whatever makes the bias (batteries are
good; I used a Lego trainset power supply which is a small AC/DC
variac) to the grid and the positive end to ungrounded side of the
filament. Scope horizontal or X also to the grid, scope vertical or Y
between the resistor (100 ohm is good) and the cathode, the other end
of the resistor to 0V. Connect grid screen and anode and bingo, that
is all the connections; it takes less time to dig out the parts and
screw and solder it up than to describe. The scope will show an
elongated J lying at an angle and that is the transfer curve for the
negative bias you feed, with the anode current up the left scale and
the anode voltage along the bottom scale.

Er, you have to rewire to get anode voltage on the scope horizontal or
X; as described you get grid voltage on the X.

Photograph, feed photo to
computer or print out, draw loadline over scope's own gridlines, and
Bob's your uncle. To make many curves on the same sheet, put camera on
tripod for stability and position, change electrical parameters, take
multiple exposures (I took singles and just overlaid them in the
computer; lost faster than fartarsing around with tricky multiple part-
exposures). Don't forget to label each negative bias curve.

The labelling requirement is a good reason to tape some of your wife's
pattern-cutting paper or baking tray paper (both are translucent
enough) over the oscilloscope screen and take a tracing as well as a
photograph. Common projection film and Koki pens also work nicely.

Whole procedure is a hell of a lot faster than fiddling with a DMM and
the uncertainties of how it loads your circuit, or pulling out a VTVM
and setting it up.

I don't know if Ian has said how he took his earlier measurements, but
I'd be surprised if he hasn't already worked out this junkbox curve
tracer for himself.

Andre Jute
Visit Jute on Amps athttp://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review