I got this idea to build a really good valve tester. I was thinking
in terms of a variable DC motor running a rotary port, but didn't like
that too much. So I instead decided on a little 8-ohm plastic cone
3-1/2" loudspeaker with a 3/4" voice coil to be my "valve operator."
That really works nicely, and gives me an audible "pop" every time
it opens (no pop to close). Even that is an advantage. It presently
opens three ports, which I can "load" with different bleeds, simulating
a trackerbar and longer tubing.
For example, the resistance encountered between the trackerbar and the
valve in an Ampico A and B can be easily measured and simulated with a
fixed bleed in one of the ports.
The valve tester basic circuit is as stable as a rock and works like
a champ. It's really simple, having an astable multivibrator and a
monostable multivibrator connected, giving me a separate frequency
adjustment and a pulse width adjustment. That translates into a
repetition speed adjustment and a valve actuation time adjustment to
test unloaded player valves (valves that aren't driving a pneumatic).
Now you ask, "How is he going to decide how that relates to the valve
he's testing?" Well, for one application, I decided to built a test
stand that incorporates several valve block patterns in one, so I can
either set a Gully, Simplex, Coinola, etc. valve unit on it, or a block
The pneumatic would be adjustably spring-loaded for weight. The
moveable leaf of the pneumatic would operate two switch positions,
one up, one down, and would attach temporarily to the fixed leaf with
a tiny pinch clamp. When the pneumatic had exerted a fixed amount of
force (travel) in each case, the switch would close, actuating both
a debounce circuit and an LED indicator light. The debouncer output
will open my power driver circuit (a power Darlington), dropping out
the control over the valve momentarily until the pneumatic returns
sufficiently to strike the note again (all adjustable).
What this does is to get the pneumatic into the act for a timing
measurement. The open/closed switch will be designed to retrofit
onto about any pneumatic anywhere, and plug right in to the circuit,
so I can use it for stacks, too. (The power supply is a Dell laptop
supply, 12 volts 2.2 A .)
The first measurement could be with the pneumatic closed. The speed
and duration adjusted by the two pots. (Maybe 10- or 15-turn pots,
if I can find the right taper and reasonably priced dials to indicate
the setting). The second measurement will be with the pneumatic
position switch plugged in. In this setting, the pulse width pot
(duration) will be disconnected with a switch in the miniature phone
jack as soon as it is plugged into the circuit, and takes the place of
the monostable multivibrator circuit. But it still has to keep up with
the frequency set by the other -- unless I want it to use just the
switch to self-oscillate with my speaker port.
I can first check valve speed with no pneumatic operating. That speed
will display either on a pulse counter or indirectly on my knobs. Then
I will substitute the pneumatic and test again, and see the change it
makes, by readjusting the frequency until the pneumatic is able to make
the full excursion reliably. From this, I will be able to tell two
1. The differences between valve types and their ability to respond,
loaded and unloaded, and;
2. The optimum valve travel settings for any valve, not to mention all
the other variables which I can also adjust before I restore them all,
and to whatever vacuum pressure range they optimize for.
By building the feedback system, I should be able to optimize any valve
type I may want to rebuild (or build), and set the valve gaps precisely
for best combination of speed and power. This would be very valuable
in reproducers, I think. Not to mention the fact that I could test any
kind of valve and know without question which ones are the most respon-
sive and powerful. It would end this question, once and for all,
because the results are sitting there on my pulse counter.
Also, by forcing the repetition to follow the frequency set, I am able
to tell if the valve is skipping a few beats with what is known as a
"missing pulse" detector. That's just another simple comparator cir-
cuit that can automatically adjust itself to the astable multivibrator
frequency and then display any missing actuations or returns in a
number of different ways. (All these refinements of course will have
to be added as I have time)
One last possibility also is -- by disconnecting the pulse generator
and relying strictly on the speaker valve, power driver, and switch
circuit, the valve/pneumatic combination could be allowed to oscillate,
turning itself off and on, and that rate could be measured. Each
valve/pneumatic system would then have a different rate, and by varying
vacuum, spring tension, and travel on the pneumatic, I should be able
to assure that the power and travel are optimized in this way, too.
By building different little clamp configurations, the switch clamp
can attach to Pitman rods and actuation fingers, abstracts, etc. to
check out any valve on any player piano. Yes, even the actual key!
The next step would be to use one of Larry Broadmoore's trackerbar
valves instead of the "speaker " valve operator, and with that, go
directly to a trackerbar hole with the travel switch operated by the
key. Now, nothing can hide!
I'd be interested in anyone's comment regarding this tester, ideas,
or anything they may think to add to it or change it. Now's a good
time -- ha.
[ The Ampico B valve (the old leaky valves, anyhow) seem to have
[ troubles at low vacuum and slow paper speed, like Tempo 50.
[ Instead of opening suddenly, the port is gradually opened.
[ Under this condition (old B valves) a magnet valve gives a
[ noticeably better performance than the music roll, apparently
[ because it operates quickly. Could your tester be adapted to
[ simulate the lethargic signal of a hole passing slowly over the
[ tracker bar, like the rotating-plate valve produces? -- Robbie