low power high voltage PSU for PMT
Smile, it can only get WORSE
Good day everyone,
I need a portability (low power battery operation) of a
photomultiplier tube and conclude that the cockcroft-walton power
supply is the way to go. The design of the voltage multiplier is
straight forward-it's done and ready to go.
However, generating the AC input (150 to 350 volts peak to peak at 80
to 300 khz) is easier said than done.
Hamamatsu has a nice cockcfort-walton based PMT assembly, but they do
not offer the entire power supply as a building block component. They
only sell their cockcroft-walton power supply as part of a complete
(potted) module-I can't use that.
The Hamamatsu HC120-3 draws 7 ma from a 15 volt dc source
because they do not use power sucking divider resistors to get the
stepped voltage needed to power the PMT. A 'conventional' supply with
divider resistors draws 125 to 150 ma! The Hamamatsu HC120 is a self
excited oscillator based on an a ferrite core autotransformer-but
that's all I know about it.
Does anyone know how to make the self excited oscillator that
generates the input voltage for the cockcroft-walton multiplier???
Anyone considering this needs to be aware that the cockcroft-walton
pmt supply only produces the power needed based on the light input of
the tube-so there is a very wide dynamic range (load current varies
widely) and there is very little power loss. Conventional high voltage
generators with voltage divider resistors present a nearly constant
load to the power supply, so they are different beasts.
Does anyone have a schematic for a cockcroft-walton based self excited
oscillator supply for a PM tube?
Also, if anyone knows where I can purchase a reasonably priced supply
of this type, please pass the word.
Thanks,
Art
Bill Sloman
The simplest self-excited oscillator to drive the Cockroft-Walton
multiplier would be a Royer inverter
http://www.mag-inc.com/pdf/TWC-500p20-22.pdf
You can use power MOSFETs rather than BJT's.
This isn't all that happy driving a very capacitative load, such as a
Cockroft-Walton multiplier.
Note that it generates a square wave, usually with a biggish switching
spike on the rising edges of the waveform. The circuit switches when
either the transformer core or the switching devices moves into
saturation.
Peter Baxandall devised his class-D oscillator in the first instance to
drive a nasty capacitative load. It looks quite like the Royer circuit,
but there is a capacitor across one of the transformer windings, which -
with the inductance of the winding - forms an LC oscillator whose
resonant frequency determines the frequency of oscillation. The output
is a sine wave - more or less.
Jim Williams of Linear Technology rediscovered this circuit some years
ago, but - with a fine disregard for historical reality, describes it as
Royer inverter. It is written up in Linear application notes AN49, AN55
and AN65
http://www.linear.com/pdf/an65f.pdf
The target application is back-lighting panels for the LCD screens of
portable computers, but the circuit has other applications.
-----
Bill Sloman, Nijmegen
David Reeve
Smile, it can only get WORSE <art...@uninets.net> wrote in message
news:3d16344e...@news.uninets.net...
I designed and manufactured a portable PMT powersupply for use in our own
instrumentation. It ran from a sinlgle 5V supply and took about 7mA for a
1000V output. However, for a one-off requirements I'd go for the Hamamatsu
CockcroftWalton socket module. Unless you have experience of this type of
thing, you'll waste a lot of time. Its easy to get a circuit pushing out
1000V and drawing 200mA at no load, it's not so easy getting a low power
cicuit working. What exactly is the problem with using the Hamamatsu
module?
Dave
Frihtiof Andreas Jensen
"Bill Sloman" <bill....@ieee.org> skrev i en meddelelse
news:3D179C4F...@ieee.org...
> The simplest self-excited oscillator to drive the Cockroft-Walton
> multiplier would be a Royer inverter
>
> http://www.mag-inc.com/pdf/TWC-500p20-22.pdf
>
> You can use power MOSFETs rather than BJT's.
>
> This isn't all that happy driving a very capacitative load, such as a
> Cockroft-Walton multiplier.
This can be fixed by adding a series inductor to the centre-point of the
transformer, turning the "classic Royer" into a current-fed Royer. It still
requires some means of voltage regulation though.
Smile, it can only get WORSE
wrote:
Thanks Bill!
I'm not sure about the Royer osc tho...
The actaul amount of power required is milliwatts at full PMT light
levels and microwatts at dark/very low light levels.
The royer saturates the core with each transition and I'm not sure it
is compatable with such a wide dynamic range load, such as a PMT.
At full load, they can be good, but at light loads, they have horrible
efficiency.
I'll check out the Jim Williams apo note that is called a Royer>:>
Thanks for your comments and input.
Regards,
Art
Smile, it can only get WORSE
Does this fix allow the enery stored in the magnetic field of the
transformer to flow into and out of the inductor? This is the only way
I can see to make a Rioyer work over the wide range of loads that a
PMT needs as it goes from full light to near-no light loads.
Any comments appreciated.
Regards,
Art
Smile, it can only get WORSE
>
>I designed and manufactured a portable PMT powersupply for use in our own
>instrumentation. It ran from a sinlgle 5V supply and took about 7mA for a
>1000V output. However, for a one-off requirements I'd go for the Hamamatsu
>CockcroftWalton socket module. Unless you have experience of this type of
>thing, you'll waste a lot of time. Its easy to get a circuit pushing out
>1000V and drawing 200mA at no load, it's not so easy getting a low power
>cicuit working. What exactly is the problem with using the Hamamatsu
>module?
>
>Dave
>
>
Thanks Dave,
Last things first...
Problem with the Hamamatsu module??? Hamamatsu is horribly expensive
and my least preferred vendor for that reason. Further, their 'module'
is not available as a building block component-it's only available
with a PM Tube that THEY choose to make it for. I've begged them to
make a universal cockcroft-walton supply available, but they have not
seen the light (yet).
Even if I had the $$$, they do not make it for the tube I need...
-----------------
5 volts and 5 ma sounds outstanding! I am tying to drive a Hamamatsu
R-1617, which is an S-20 coating with 10 stages and 10e6 gain in a 3/4
head on package. I'm curious Dave, is this tube similar to the one you
ended up desigining your supply for???
Will you share your circuit?
Thanks,
Art
Bill Sloman
"Smile, it can only get WORSE" wrote:
>
> On Tue, 25 Jun 2002 11:09:34 +0200, "Frihtiof Andreas Jensen"
> <frithio...@justremovethis.jensen.tdcadsl.dk> wrote:
>
> >
> >"Bill Sloman" <bill....@ieee.org> skrev i en meddelelse
> >news:3D179C4F...@ieee.org...
> >> The simplest self-excited oscillator to drive the Cockroft-Walton
> >> multiplier would be a Royer inverter
> >>
> >> http://www.mag-inc.com/pdf/TWC-500p20-22.pdf
> >>
> >> You can use power MOSFETs rather than BJT's.
> >>
> >> This isn't all that happy driving a very capacitative load, such as a
> >> Cockroft-Walton multiplier.
> >
> >This can be fixed by adding a series inductor to the centre-point of the
> >transformer, turning the "classic Royer" into a current-fed Royer. It still
> >requires some means of voltage regulation though.
> >
> >
>
> Does this fix allow the enery stored in the magnetic field of the
> transformer to flow into and out of the inductor? This is the only way
> I can see to make a Royer work over the wide range of loads that a
> PMT needs as it goes from full light to near-no light loads.
If Frihtiof Andreas Jensen had read the rest of my post, he might have
woken up to the fact that his "current-fed Royer" is actually the
classic Baxandall class-D oscillator. You can control the current into
the centre-tap of the transformer in other ways than just using an
inductor, and you have to, if you want to sustain a good looking sine
wave under load - as I did for the load cell driver in the Metals
Research gallium arsenide crystal puller - but Baxandall's paper
describes a circuit with an inductor
Both the Royer and the Baxandall oscillators will work reasonably
happily over the wide range of loads the PMT faces (not that the
photocathode current on a PMT ever represents a serious load ...). The
catch with the Royer oscillator - unless you snub it hard, when it
starts looking uncannily like a half-baked Baxandall circuit - is that
under light loads the Cockroft-Walton multiplier will multiply up the
tips of the switching spikes, which can be embarrassing to the point of
being destructive.
Transformers are voltage output devices, and - within the limits set by
the coil resistances - will deliver as much current as the load will
take at a given voltage. This makes capacitative loads a bit tricky.
----
Bill Sloman, Nijmegen
John Larkin
only get WORSE) wrote:
>
>I'm not sure about the Royer osc tho...
>
>The actaul amount of power required is milliwatts at full PMT light
>levels and microwatts at dark/very low light levels.
>
>The royer saturates the core with each transition and I'm not sure it
>is compatable with such a wide dynamic range load, such as a PMT.
>
>At full load, they can be good, but at light loads, they have horrible
>efficiency.
>
>I'll check out the Jim Williams apo note that is called a Royer>:>
>
>Thanks for your comments and input.
>
>Regards,
>
>Art
Art,
why not just pump the C-W tree with a simple flyback switcher? That
just needs one fet and an off-the-shelf inductor. I did a similar
thing a while back to drive avalanche transistors from a 9 volt
battery, with a simple regulator based on a 4000-series CMOS gate. If
I can find the schematic, I'll post it to a.b.s.e.
A Royer seems like a lot of work to me.
John
Smile, it can only get WORSE
>
>why not just pump the C-W tree with a simple flyback switcher? That
>just needs one fet and an off-the-shelf inductor. I did a similar
>thing a while back to drive avalanche transistors from a 9 volt
>battery, with a simple regulator based on a 4000-series CMOS gate. If
>I can find the schematic, I'll post it to a.b.s.e.
>
Several reasons come to mind immediately.....
-An inductor will not take low voltage and convert it to 350 volts!
-The overhead from the switching chip will suck up all the efficiency,
there are no switching chips designed to run with the power input that
a PM tube would need. If someone makes a low power switcher in the
future, I'd be on it in a flash.
>A Royer seems like a lot of work to me.
>
A Royer is probably NOT the answer either. They need to saturate the
core each time they switch, so they are lossy at low power becasue the
cores can't be made small enough! The Royer also puts out spikes-not a
clean AC sinewave.
The only implementation of a cockcroft-walton supply that works is the
Hamamatsu hc-120. It uses a single transistor LINEAR (class A) power
oscillator driving a ferrite core transformer that operates at 140
khz.
So far, all are usenet users are suggesting switching type supplies,
despite the losses associated with repetative and high peak switching
currents. I'm not sure a Royer or any other switching supply will work
at the extreme low power levels needed by the PMT.
Sure would like to see some dissenting opinions and schematics.
Regards,
Art
PS: What's a 'simple' flyback switcher?? All the ones I've seen
require tons of components and are far from simple. Yes, inductor
based switchers are simple, but no vendor makes a simple switcher to
drive a flyback, especially at low power.
Sir Charles W. Shults III
news:3d192227...@news.uninets.net...
>
<snip>
> -An inductor will not take low voltage and convert it to 350 volts!
Sure it can. Almost any small DC motor produces spikes on the order of 200
volts without trying. A few turns of wire around a nail can easily generate
spikes of 500 to 600 volts when you tap its connections on a 9 volt battery.
The trick is in the switching.
And think about transformers- they can convert any voltage to any other,
within reason. A small strobe trigger transformer often takes a piddly 6 volt
DC pulse and converts it to 10 kilovolts to trigger the "leader stroke" effect
that causes a strobe tube to fire.
A simple boost regulator is just the low end of the spectrum. With a pulsed
inductor, you can drive a simple Cockcroft-Walton multiplier and get many tens
of thousands of volts. You are limited only by losses and the stoutness of your
rectifiers and capacitors.
Cheers!
Chip Shults
My robotics, space and CGI web page - http://home.cfl.rr.com/aichip
John Larkin
only get WORSE) wrote:
>
>>
>>why not just pump the C-W tree with a simple flyback switcher? That
>>just needs one fet and an off-the-shelf inductor. I did a similar
>>thing a while back to drive avalanche transistors from a 9 volt
>>battery, with a simple regulator based on a 4000-series CMOS gate. If
>>I can find the schematic, I'll post it to a.b.s.e.
>>
>
>Several reasons come to mind immediately.....
>
>-An inductor will not take low voltage and convert it to 350 volts!
>
Sure will!
>-The overhead from the switching chip will suck up all the efficiency,
>there are no switching chips designed to run with the power input that
>a PM tube would need. If someone makes a low power switcher in the
>future, I'd be on it in a flash.
That's why I designed my own.
>
>>A Royer seems like a lot of work to me.
>>
>
>A Royer is probably NOT the answer either. They need to saturate the
>core each time they switch, so they are lossy at low power becasue the
>cores can't be made small enough! The Royer also puts out spikes-not a
>clean AC sinewave.
>
>The only implementation of a cockcroft-walton supply that works is the
>Hamamatsu hc-120.
Hmmm... a kinda extreme statement. Lots of people do this.
>It uses a single transistor LINEAR (class A) power
>oscillator driving a ferrite core transformer that operates at 140
>khz.
A resonant circuit would be OK here too, but switching is inherently
more efficient than class A, if that matters. The old Tek scopes used
a resonant oscillator HV supply, with a custom step-up transformer and
a tripler to get to 8 KV or so.
>
>So far, all are usenet users are suggesting switching type supplies,
>despite the losses associated with repetative and high peak switching
>currents. I'm not sure a Royer or any other switching supply will work
>at the extreme low power levels needed by the PMT.
There's no reason a switcher has to be inefficient, just because the
power is low. I recall a WWII IR night vision power supply that was a
switcher. It ran on one D-cell battery and delivered about 2KV to a
sniperscope tube. The switcher was a mechanical contact driven by a
watch spring at about 1 Hz. Ran for weeks on one battery. The
rectifier was interesting, too.
John
>
>PS: What's a 'simple' flyback switcher?? All the ones I've seen
>require tons of components and are far from simple. Yes, inductor
>based switchers are simple, but no vendor makes a simple switcher to
>drive a flyback, especially at low power.
I guess I'll have to post the critter. I'll put it on a.b.s.e.
tomorrow, if I can find it!
Bill Sloman
The Royer oscillator doesn't have to saturate the core - you can
saturate the driver transistors instead. And there is nothing to stop
you building a driven inverter, where the period is set by by a
monostable or a crystal controlled osscillator and some logic, where
you can turn off one driver before you turn on the other, avoid
efficiency-damaging fearues like croe adn driver saturation and driver
"shoot through".
I rather fancy the idea of building a circuit which uses a 4046
phase-locked loop and VCO with a divide-by-16 counter to produce
better switch driver waveforms for a Baxandall inverter while still
keeping them synchronised to the rather uncertain resonant frequency.
Messy circuit design and too many components for comfort, but it
needn't draw much current.
> At full load, they can be good, but at light loads, they have horrible
> efficiency.
Don't get too hung up about efficiency under light load - what you are
actually worried about is current drawn, and as soon as you can get
that down below the self-discharge current of your battery, you
couldn't care less about efficiency.
What is a worry with the Royer under light load are the switching
spikes (as I've mentioned elsewhere in this thread). Under light load
your Cockroft-Walton voltage multiplier will charge up to a voltage
equivalent to the peak of the switching spikes. Baxandall's class-D
oscillator or "current-fed Royer" finesses this problem rather nicely.
See
Baxandall, P.J, Proc I.E.E 106, B, page 748 (1959)
which Tony Williams posted - with other comment and another reference
- on s.e.d. 2000/06/16.
> I'll check out the Jim Williams ap. note that is called a Royer>:>
Jim Williams describes something that is recognisably the Baxandall
class-D oscillator, even if he call it a Royer inverter. From what I
can remember of Baxandall's paper, it gives you a much better insight
into what is going on than do any of the Jim Williams application
notes, but Jim Williams has put a lot of time into the circuit and the
result has a certain glossy finish which the rest of us can only envy.
> Thanks for your comments and input.
Always grateful for a chnce to propagandise for Peter Baxandall - he
is famous for his active tone-control but he did a lot of other good
stuff. I never met him, but I learned a lot from one of his
students/apprentices.
----
Bill Sloman, Nijmegen
Tony Williams
Bill Sloman <bill....@ieee.org> wrote:
[snip]
> I rather fancy the idea of building a circuit which uses a 4046
> phase-locked loop and VCO with a divide-by-16 counter to produce
> better switch driver waveforms for a Baxandall inverter while still
> keeping them synchronised to the rather uncertain resonant frequency.
> Messy circuit design and too many components for comfort, but it
> needn't draw much current.
In explaining the Baxandall oscillator, Thomas
Roddam remarked on the key features of the
transistor drive-timing.... If the tuned circuit
is parallel (Baxandall's) then the transistors
are switched at the zero-crossover of the tank
voltage. If it is the series-tuned variant, the
switch must happen at the zero-crossover of the
tank current.
It could be handy then (in the parallel-tuned osc)
to have a voltage sense winding, with comparators
looking for zero-crossing. This could get early-OFF,
late-ON switchings, with the Q of the tank carrying
the waveform through the short time when both devices
are OFF. Comparators are also useful with MOSFETs.
--
Tony Williams.
Smile, it can only get WORSE
>>The only implementation of a cockcroft-walton supply that works is the
>>Hamamatsu hc-120.
>
Hey John,
I should have stated that the only KNOWN comm'l implementation that I
know of IS the HC-120 by Hamamatsu. These use a single transistor free
running oscillator that produces a fairly clean sine wave and runs
class A at (about) 140 khz.
It uses a pot core transformer and varies the input voltage to the
oscillator circuit to compensate for load changes.
When I started this thread, my hope was that someone had reverse
engineered this Hamamatsu device and that they might have a schematic-
However, I'm not opposed to using inductor based non-linear voltage
generators or one that is designed around an inductor!
I've used Maxim switching regulators before however, and they all have
very poor efficiency at low power, mostly becasue they are designed to
produce multi-watt outputs, not the milliwatt levels a pmt needs. I
have no doubt that a switcher that was DESIGNED to drive a flyback
transformer at low power would be a good performer! To the best of my
knowledge, a low power switcher capable of driving a flyback suitable
for this ap doesn't exist. If this is incorrect, please point me to a
suitable chip.
The voltage references needed to maintain regulation are built into
these switchers, and that's a big PLUS....but they require alot of
components and there aren't many vendors into low power transformers.
Would love to know more about an inductor based high voltge generator,
even if it uses discrete components to achieve the necessary
efficiency at low power levels.
Hey John, what's A. B. S. E. ???
I've also thought about piezo transformers-they are interesting
devices although driving them is no picnic!
I've wondered about using a simple flash unit from a disposable camera
to generate high voltage and then using that high voltage to power a
linear output stage. A freerunning 100 khz oscillator would supply the
sinewave input and the output would be regulated by varying the the
output of the oscillator. The cameara flah unit draws BIG power, but
it charges an energy storage cap, so it would not have to run full
time....maybe this would be quick, dirty and cheap?
I've studied CCFL inverters too, it's a shame they don't/can't operate
about 1 khz or so....and that they all have fault detectors that don't
have variable thresholds. They all expect 1 or 2 ma of load current,
otherwise they think the lamp is defective and they shut down! Oh
well!
So, tell me more please! And thankyou to all who made comments and
contributions to this thread!
Regards,
Art
John Woodgate
<to...@ledelec.demon.co.uk> wrote (in <4b4c980...@ledelec.demon.co.
uk>) about 'low power high voltage PSU for PMT', on Wed, 26 Jun 2002:
>In explaining the Baxandall oscillator, Thomas
> Roddam remarked on the key features of the
> transistor drive-timing....
Do you have a reference for that? I though the pseudonym 'Thomas Roddam'
was unique to WW, but I don't recall any such article on PB's
oscillator.
--
Regards, John Woodgate, OOO - Own Opinions Only. http://www.jmwa.demon.co.uk
Interested in professional sound reinforcement and distribution? Then go to
http://www.isce.org.uk
PLEASE do NOT copy news posts to me by E-MAIL!
John Larkin
Art,
I posted the jpeg to alt.binaries.schematics.electronic, which is
where the guys in thie group generally post binaries. That might be a
starting point for what you want to do.
There are some tricks associated with the inductor-only flyback thing,
a fight between stored energy in the inductor versus how much energy
it takes to pump up the fet drain capacitance. A resonant converter,
well, resonates out the parasitic capacitance, so is maybe easier to
do if you can get the right transformer. Knowing Hamamatsu (and I do!)
they probably spent 20 man-years (all their engineers *are* men)
developing theirs.
A CCFL or camera transformer would be worth looking into.
Maybe I'll post an old resonant Tektronics CRT supply, too.
John
Bill Sloman
The other thing I had in mind was that the inductor in the classic
Baxandall parallel-tuned circuit doesn't deliver its current at the
right time. (IIRR Baxandall's paper also talks about the series-tuned
circuit as well, so he may be said to own both ....)
To compensate for the resistive losses in the tank circuit and the load,
the transformer centre-tap should be fed with a current which is more or
less in-phase with the voltage from centre-tap to ground, and to get
that, the voltage across the inductor should lead this current by 90
degrees. In fact, the voltage across the inductor is 180 degrees out of
phase with the current you want, which accounts for the weird waveform
you get out of a heavily loaded class-D oscillator.
It strikes me that you could use an extra pair of switches to PWM the
voltage at the driven end of the inductor to produce the roughly 90
degree phase lead that you want - something pretty rough, just aimed at
getting the third harmonic about right (make that a divide by 12
counter?) could give a significantly nicer looking sine wave under load.
The tank circuit will kill the higher harmonics pretty well anyway.
You lose some of the voltage step-up, of course.
----
Bill Sloman, Nijmegen
David Reeve
Ok, I see your problem. I was thinking of the CW socket they make for their
1&1/8 side window tubes. The circuit I designed has been used for quite a
range of tubes, large and small side window devices, as well as a 1" linear
focussed endwindow device. In all cases we are talking of DC coupling
straight into an I to F convertor situated right at the anode pin. I have
also used a variation on the theme to power the dynode chain of a high speed
(8ns resolution) two tube coincidence circuit for a liquid scintillation
counter. However, the need for a relatively low impedance dynode chain hits
the power consumption, and a 3KV supply draws some 200mA at 5V, if I
remember correctly.
How does it work? Flyback on a single inductor steps 5V to better than
800V. Switch-off speed approx 40ns. Switch is a standard MOSFET with a bit
of stuff on the gate to handle the gate capacitance on switch-off.
Regulation is a LinearTech skip pulse regulator running at a nominal 20KHz.
Positive supplies require just the single miniature inductor, negative
supplies for a CW PMT supply require the same inductor to be bifilar wound
with a secondary to provide the inversion. Miniaturisation and careful
design reduces stray inductance low enough you don't need snubbing. All SM
components throughout. Final build has to be encapsulated to prevent
arc-over and tracking.
Art, if this is a non commercial project then contact me directly and I will
get you the details, however, my time is strictly limited, and you would
want to know your way around this type of circuit. If this is a commercial
project, then I can assist in getting the design into production, but would
need to be paid for my time.
regards
Dave
Tony Williams
John Woodgate <j...@jmwa.demon.contraspam.yuk> wrote:
> Do you have a reference for that? I though the pseudonym
> 'Thomas Roddam' was unique to WW, but I don't recall any
> such article on PB's oscillator.
A book, "Transistor Inverters and Converters",
by Thomas Roddam. Published by Illiffe Books
for WW in 1963, (42 shillings net).
Part of TR's chapter on sinewave inverters
quotes from two papers that were apparently
offered to the IEE at the same time.
1. Yarrow CJ, Proc IEE, 106, B, 1320 (1959).
2. Baxandall, PJ, Proc IEE, 106, B, 748 (1959).
They were very similar designs, but it became
known as the Baxandall oscillator. In fact
though Yarrow's paper would be more interesting
to this thread, because he designed a 16KV
converter, where the secondary winding appears
to have been resonated+loaded solely by the
input impedance of the voltage doubler.
I only have TR's description, not the original
papers.
--
Tony Williams.
Tony Williams
Bill Sloman <bill....@ieee.org> wrote:
> The other thing I had in mind was that the inductor in the classic
> Baxandall parallel-tuned circuit doesn't deliver its current at the
> right time. (IIRR Baxandall's paper also talks about the series-tuned
> circuit as well, so he may be said to own both ....)
> To compensate for the resistive losses in the tank circuit and the load,
> the transformer centre-tap should be fed with a current which is more or
> less in-phase with the voltage from centre-tap to ground, and to get
> that, the voltage across the inductor should lead this current by 90
> degrees. In fact, the voltage across the inductor is 180 degrees out of
> phase with the current you want, which accounts for the weird waveform
> you get out of a heavily loaded class-D oscillator.
I must admit, I haven't thought about it in that depth.
The few times I've used the Baxandall circuit have
all had reasonable sinewaves at the load (afair).
But that may be because of the sums quoted by TR.
PJB estimated the third harmonic distortion at the
load as 100/8Q (%), where Q= Rload/2.pi.F.Lsec.
PJB's L= Rl(collector-collector)/2.pi.F.
With a Q of 10 (the max suggested, otherwise
squegging could occur) this makes the supply
inductance about 5x L-(collector-collector).
> It strikes me that you could use an extra pair of switches to PWM the
> voltage at the driven end of the inductor to produce the roughly 90
> degree phase lead that you want - something pretty rough, just aimed at
> getting the third harmonic about right (make that a divide by 12
> counter?) could give a significantly nicer looking sine wave under load.
> The tank circuit will kill the higher harmonics pretty well anyway.
TR speculated in the direction of replacing
the single inductor with a series of band-
pass filters, at the even harmonics. He ended
up with a (speculated) m-derived low-pass filter.
TR has a final paragraph associated with this
which I have never really understood... below;
"Another way of thinking about the circuit which
is useful is to regard the system as one in
which the collector centre-tap voltage is made
up of all the even terms and which then operates
as a switching modulator. Then third-harmonic
terms not accepted by the load are pushed back
into the modulator and remixed with fundamental
to produce 2nd and 4th harmonics which attempt
to escape along the supply line: these are also
driven back and in the ideal form the second and
third mix to produce a fundamental, which can
reach the load while all other terms modulate
and re-modulate each other until either the zero
or first order terms are produced."
--
Tony Williams.
Frihtiof Andreas Jensen
> If Frihtiof Andreas Jensen had read the rest of my post, he might have
> woken up to the fact that his "current-fed Royer" is actually the
> classic Baxandall class-D oscillator.
It's not "mine" - I stole the design right out of Pressmann: "Switching
Power Supply Design" ;-)
The Baxandall design should fix the capacitive load problem: Since the
voltage is sinusoidal, the current will be sinusoidal too, it's peak value
being limited by the slew-rate of the sine voltage. (Or the impedance of the
capacitive load at the operating frequency, for those who prefer that).
I did reach your post - thanks for the links BTW.
Tony Williams
Tony Williams <to...@ledelec.demon.co.uk> wrote:
I'm just using one of my posts to give what
looks like a seriously useful reference.
Electronics World, October 1996. The article
entitled "Stepping Out" by Ian Hegglun.
Pages 795 to 797, plus pages 805 to 806.
The article describes a new multiplier technique,
(with reduced component count) which the author
named the Pentupler.
Two example designs are given; A 300W converter,
13.8V to +/-50V, and a ?V to +/- 160KV (yes KV),
where a tv line transformer (sec resonated at
30KHz by the multiplier) generated 16KV AC, and
into a pair of Pentuplers.
The idea was the subject of a patent application,
but in the artcle the author gives permission for
experimenters to use the circuits provided it is
for non-commercial applications.
--
Tony Williams.
John Woodgate
uk>) about 'low power high voltage PSU for PMT', on Thu, 27 Jun 2002:
>TR has a final paragraph associated with this
> which I have never really understood... below;
He did have a habit of becoming obscure at times, unlike 'Cathode Ray'
(M G Scroggie) who was, I think consistently and utterly lucid.
John Woodgate
uk>) about 'low power high voltage PSU for PMT', on Thu, 27 Jun 2002:
but I can't find it now.
I expect the book is long out of print. But IEE can supply the original
papers, I feel sure. Class D in 1959?
John Woodgate
uk>) about 'low power high voltage PSU for PMT', on Thu, 27 Jun 2002:
>Electronics World, October 1996. The article
> entitled "Stepping Out" by Ian Hegglun.
> Pages 795 to 797, plus pages 805 to 806.
The technique in Figure 4 seems to be derived from the parallel-fed
voltage multiplier, which does indeed have advantages over the series-
fed or Cockroft-Walton multiplier. But the article suffers from the EW
figure bugs. Fig 2 a) certainly isn't what the caption says; it's
another version of Fig.3.
Bill Sloman
Tony Williams wrote:
>
> In article <3D1A44E1...@ieee.org>,
> Bill Sloman <bill....@ieee.org> wrote:
<snip>
> TR has a final paragraph associated with this
> which I have never really understood... below;
>
> "Another way of thinking about the circuit which
> is useful is to regard the system as one in
> which the collector centre-tap voltage is made
> up of all the even terms and which then operates
> as a switching modulator. Then third-harmonic
> terms not accepted by the load are pushed back
> into the modulator and remixed with fundamental
> to produce 2nd and 4th harmonics which attempt
> to escape along the supply line: these are also
> driven back and in the ideal form the second and
> third mix to produce a fundamental, which can
> reach the load while all other terms modulate
> and re-modulate each other until either the zero
> or first order terms are produced."
I had a colleague who used to talk like that. It was fine as long as he
was talking about stuff where I'd had six months to immerse myself in
the same technical details that were now engaging his (not
inconsiderable) intelligence, but for the rest of the time it was just
random noise.
Good technical writing is tricky.
----
Bill Sloman, Nijmegen
Bill Sloman
Oops. I read the relevant bit of Pressman, and you are right - and there
is no mention of a tank circuit anywhere. When I was playing with Royer
inverters, we used to add resistance in series with the base drive in
various places to get the "current fed" effect, which means that the
switching period is set by transistor saturation rather than By core
saturation, and I've always thought of that as an intrinsic part of the
Royer bag of tricks. Wrong! Another thing I thought I knew that wasn't
so.
Hmm. I wonder if I'm cleaning up my intellectual baggage as fast as I'm
adding new errors?
-----
Bill Sloman, Nijmegen
Steven Arnold
>
> Art
My apologies if this has been mentioned before, but have you looked at
the supplies from Burle? They have some which run on 1-5 or 1-10 VDC,
as I recall. I don't remember the current requirements. They are small
potted units, but just the supply, not the socket, pmt housing, etc.
like the Hammamatsu units.
Steven Arnold
Smile, it can only get WORSE
>
>My apologies if this has been mentioned before, but have you looked at
>the supplies from Burle? They have some which run on 1-5 or 1-10 VDC,
>as I recall. I don't remember the current requirements. They are small
>potted units, but just the supply, not the socket, pmt housing, etc.
>like the Hammamatsu units.
>
>Steven Arnold
Hi Steve,
I have been to Burle before and even contacted them by phone a few
years ago. At that time, they didn't even know what 'low power' meant,
I left and vowed never to call them again.
After receiving your message, I went to their website.. After
reviewing their website, it appears that nothing has changed at Burle.
I see some small psu's, but there is nothing in the low power
class...all of them are the conventional power sucking type that rely
on resistive voltage dividers (rather than the more preferable
cockcroft-walton voltage multiplier type).
If you have knowledge of other Burle products that might meet my
needs, I'd appreciate it if you could give me a model number or part
number so that I can contact them directly.
To the best of my knowledge, Hamamatsu and Matsuada are the only
vendors supplying cockcroft-walton psu's for pm tubes.
Hamamatsu sells only potted units that include a pmt-it is useless
unless yuou need that specific tube--nothing in the 'building block
line'.
Matsuada has only recently brought a www site online and they do not
list any cockcroft-walton based power supply units in their current
product listing.
If anyone can suggest a source for cockcroft-walton based pm tube
psu's, please contact me or post on usenet.
Thanks,
Art
Christopher R. Carlen
> unless yuou need that specific tube--nothing in the 'building block
> line'.
>
> Matsuada has only recently brought a www site online and they do not
> list any cockcroft-walton based power supply units in their current
> product listing.
>
> If anyone can suggest a source for cockcroft-walton based pm tube
> psu's, please contact me or post on usenet.
Is an "active voltage divider" PSU unacceptable to you? The Hamamatsu
C6270 HV PSU socket contains such, and is solidly linear to >100uA.
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crc...@sandia.gov
John McMillan
Smile, it can only get WORSE
<crc...@sandia.gov> wrote:
But it draws 150 ma, even when the tube sees darkness! The resistors
provide a constant load, and dissipate power in close proximityu to
the tube.
I'm looking for a low power alternative, although I am not opposed to
resistive dividers. I've never seen a resistive divider that allowed
full power to the tube without placing a big load on the supply
voltage line.
Any alternative suggestions appreciated.
Art
Chris Carlen
> <crc...@sandia.gov> wrote:
>
> But it draws 150 ma, even when the tube sees darkness! The resistors
> provide a constant load, and dissipate power in close proximityu to
> the tube.
>
> I'm looking for a low power alternative, although I am not opposed to
> resistive dividers. I've never seen a resistive divider that allowed
> full power to the tube without placing a big load on the supply
> voltage line.
>
> Any alternative suggestions appreciated.
>
> Art
I seem to recall this was talked about a bit already, but if Hamamatsu
can put the ideal cockcroft-walton PS into the little tiny boxes that
they sell as modules H6780-xx I seem to recall, then would it bee so
difficult to build one from scratch, even if it were a little bigger and
less efficient? It still would likely beat the power waste of the C6270
socket.
Do you have painful size requirements as well?
--
_____________________
Christopher R. Carlen
cr...@earthlink.net
Suse 7.3 Linux 2.4.10
Smile, it can only get WORSE
No, my size requirement is not nearly as painful adn the power
requirement.
I have the full spec sheets for the HC-120 Hamamatsu (not the
shortened version available on the web).
In the detailed sheet, they give some info...
Frequency of operation is 140 khz, it's a single transistor LINEAR
free running oscillator that is based on an autotransformer wound on a
ferrite pot core. It draws less than 8 ma from a 15 volt supply (with
the tube in darkness).
If they would sell me the supply separately, I'd be a happy camper,
but they do not offer it as a building block component--only as a
complete unit with tube, shielded case etc.
I actually tried to duplicate the circuit several times and couldn't
do it..... The oscillator either stops OR it clips. This is
understandable becasue the supply needs to have a BIG +++ dynamic
range (from darkness to full illumination) and power oscillators don't
like variable loads.
Anyway, I've never managed to make anything on the bench run-which is
why I posted here.
Has anyone ever tried to take an HC-120 out of its case to reverse
engineer it???
Keep those cards and letters coming...and thanks.