DIYbanter - View Single Post - FX-6A flashtube availability and data

In article ,
(DoN. Nichols) wrote:

According to Joseph Gwinn :
In article ,
(DoN. Nichols) wrote:

[ ... ]

One can still buy meters, and there were only a few standard sizes, so a
retrofit is certainly possible. What will be lost is the
device-specific artwork (including scales) on the meter face.

The scale is a DB scale, and not the common range of a VU meter.
It is -6 to +10, plus white zones (on a black background) for "CAL" and
"BAT", so swapping in another meter will take some artwork.

The face shape of the meter is also uncommon, except in GR
equipment.

And nothing tells me what the full-scale current (or voltage)
is, other than the fact that there are three forward-biased diodes in
series across the meter, so it should not be any more than 2.1V FS.

Are these diodes only for protection, or do they also implement the log
function? There were meter movements with overly strong springs, so it
took a significant voltage to lift the pointer off the "zero" stop.
Such a meter could indicate the voltage across the diode string, which
is logarithmic in current.

I've done a bit of exploratory work with a stereo zoom
microscope, and the movement out of the case. It turns out that the top
hairspring is not firmly attached at the axis of the coil, though it is
attached at the outer end, and the bottom hairspring flexes as would be
expected throughout the range. I'm going to try my hand at re-soldering
the inner end of that -- after unsoldering the upper end and lifting away
the bridge with the jeweled bearing. It is certainly the kind of work
which is best done under the stereo-zoom microscope (it turns out that
mine is the AO version, not the B&L, FWIW.)

And a very steady hand...

At the moment, if I turn on the voltage check function, and then
tap the meter face, it makes contact long enough to send the meter to
beyond full scale occasionally, and to have it vibrating at a lower
level most of the rest of the time -- about 5/8th scale. The reason
that it goes beyond full scale is because only one hairspring is
resisting the force generated by the moving coil, so it moves beyond
where it normally would.

Sounds right.

O.K. That could have been the purpose of the air-core inductor
in the supply which I duplicated. The wire was about #10 ga bare
copper, wound in a groove turned in a plexiglass cylinder the full
width of the case.

How many turns? What is the diameter and length of the winding? There
are standard formulas to compute the inductance given this data.

This is from about 1966 or so, IIRC. You expect such details?

The coil form was long enough so it just barely fit inside a 19"
cabinet rack with about 15" of vertical panel space. Since there was
some space left at the ends to avoid breakdown to the mounting hardware,
that limits the coil to perhaps 15" to 17". Coil form diameter was
about 3", minus 1/2 the diameter of the wire to allow it to bottom in
the turned groove. Something like about 5 to 8 turns per inch, so that
would make it 75 to 136 turns.

I rest my case - you do remember, well enough to compute approximate
inductance.

What was this supply for?

Laser pump for Nd-Yag optically pumped laser. Max rep rate of
10 PPS. It was intended to be an illuminator for a night vision scope
which was gated on when the pulse was at the right distance, and off
afterwards to mimize the brightness of light sources within the field of
view. It was viewed by a three-stage image intensifier system before
the changeover to the micro-channel ones which could get the same gain
in a single tube.

I probably did calculate the inductance back when I made it.
The original design was to have it tapped for the ten 20uF 1KV
capacitors along its length to stretch the pulse, but that turned out to
not work well, so the pulse was going through the entire length of the
inductor from all of the capacitors in parallel.

This tapped inductor with capacitors is a lumped-component transmission
line, one kind of pulse-forming network. These were widely used in WW2
radars to generate the radar pulse, where one wants a constant power for
a fixed length of time (rather than the exponential decline of a simple
RC network).

I recall a passing note in Edgerton's book saying that lots of people
had tried to use pulse-forming networks to drive flashlamps, but the
wild variations in flashlamp impedance during the flash undermined all
their efforts.

It was charged from two big gray transformers made by UTC to be
able to get the caps back to full voltage within 100 mS.

In the lab, it was good for erasing print on paper. :-) The
power supply also lit me up rather spectacularly -- *once*. Between
left finger and right elbow (which was leaning on the case of the scope
on a cart.

Ten 20-uF caps at 1000 Volts is 100 Joules. Ouch! Good it was not from
arm to arm. A defibrillator is 400 Joules.

Joe Gwinn