25kV AC

[Winfield Hill]

Tauno Voipio wrote...

The tens of kilovolts need careful attention to electric
field strength. The wires and terminals need to pretty
bulky to keep the field strength under control.

Yes indeed. We use a nice 4.8mm dia wire, AWM 3239,
with a 40kV rating, CSA TV-40. Judd Flexrad HV.


--
Thanks,
- Win

 

On 30 Sep 2019 01:45:42 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

When I was a kid, we'd bum old neon sign transformers from the sign
companies. I had a nice 18KV unit with an insulator on each end. I
guess neons are all high frequency now, or fake-neon LED things. Maybe
there are some old trannies around.

I think neon looks better with 60 Hz drive. The high-freq stuff looks
fuzzy or something.

I recently tested a Hammond toroidal 50/60 Hz power transformer. It
worked fine to over 50 KHz. Most power transformers work fine in the
audio range.

You might make some HV DC, and drive an h-bridge to make a sine. It
wouldn't be unreasonable at your power levels.

Or some C-W multipliers where you modulate the input. Discharge could
be slow at your frequency. There are maybe possibilities for push-pull
C-W.

How about car ignition coils and rectifiers, or commercial HVDC
modules, back-to-back? Needs a little management.








--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Tauno Voipio]

On 30.9.19 11:45, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

The tens of kilovolts need careful attention to electric
field strength. The wires and terminals need to pretty
bulky to keep the field strength under control.

--

-TV

 

On Mon, 30 Sep 2019 07:53:18 -0500, "Tim Williams"
<tiwill@seventransistorlabs.com> wrote:

Two CRT flybacks + pair 6BK4C + HV optos to drive them (since Vgk ~ 100V).
:^)

One flyback needs to be heavily isolated, suggest LEDs + PV for expedient
isolation; otherwise, a custom resonant transformer will do.

Tim

Tubes are not a bad idea at all here. A fiberoptic fed cascode would
be fun.

Win, can you use differential drive into your gadget? Maybe a
full-bridge with the uppers just constant-current, so the lower grids
are near ground.

Watch out for x-rays.





--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

On 30 Sep 2019 01:45:42 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

To summarize the requirement,

Do you want a sine wave or a pulse?

What waveform/polarity/coupling?

Single ended or differential?

Load? Cable?

Rep rate/duty cycle?



--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Wond]

On Mon, 30 Sep 2019 08:56:56 -0700, John Robertson wrote:

On 2019/09/30 8:26 a.m., Wond wrote:
On Mon, 30 Sep 2019 01:45:42 -0700, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast procured on
eBay, but I fear it was thrown away in a downsizing move, damn!
With that we could make a 10kV rms AC signal, at a frequency of our
choice. And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz amplifier /
resonant-transformer for a mass spec. That was 10kV amplitude, or
20kV p-p. RIS-480 / RIS-514 was a compact 7.5kV 300kHz version.
But now we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at 24kV 400VA. I
think it's molded, rather than oil- filled. OK, at least it's not a
pole-pig. But, haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us, although it'd be
nice to experiment with a little bit higher, maybe 400Hz. Hmm, many
60Hz power transformer are happy at 400Hz. Use a class-D amplifier
to drive a 50-to-230V step-up xformer, then on to a 25kV transformer
beast someplace.

I have all types of small high-voltage pulsers / transformers, many
from Alibaba, for electrostatic precipitators, etc. Some are sold
simply to make sparks, but these are short-duration pulse types, and
we need 2-10ms pulses, too long. TV flyback, 15kHz, nah. Also,
Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type. Hmm, only 10 to
15kV. And most are big and bulky.

The old oil-furnace ignition transformers would initiate and
sustain
an arc about an inch and a half long, in air. Dunno what voltage.


Formula for spark gap KV in air:

http://www.kronjaeger.com/hv/hv/msr/spk/

John :-#)#

Nice, thanks. Looks 40kV-ish.

 

[John Robertson]

On 2019/09/30 8:26 a.m., Wond wrote:

On Mon, 30 Sep 2019 01:45:42 -0700, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast procured on eBay,
but I fear it was thrown away in a downsizing move, damn! With that we
could make a 10kV rms AC signal, at a frequency of our choice. And
maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz amplifier /
resonant-transformer for a mass spec. That was 10kV amplitude, or 20kV
p-p. RIS-480 / RIS-514 was a compact 7.5kV 300kHz version. But now we
need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at 24kV 400VA. I
think it's molded, rather than oil- filled. OK, at least it's not a
pole-pig. But, haha, we only need 10uA of current, so something much
smaller would be nice.

50 or 60Hz are probably good frequencies for us, although it'd be nice
to experiment with a little bit higher, maybe 400Hz. Hmm, many 60Hz
power transformer are happy at 400Hz. Use a class-D amplifier to drive
a 50-to-230V step-up xformer, then on to a 25kV transformer beast
someplace.

I have all types of small high-voltage pulsers / transformers, many
from Alibaba, for electrostatic precipitators, etc. Some are sold
simply to make sparks, but these are short-duration pulse types, and we
need 2-10ms pulses, too long. TV flyback, 15kHz, nah. Also,
Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type. Hmm, only 10 to 15kV.
And most are big and bulky.

The old oil-furnace ignition transformers would initiate and sustain
an arc about an inch and a half long, in air. Dunno what voltage.

Formula for spark gap KV in air:

http://www.kronjaeger.com/hv/hv/msr/spk/

John :-#)#

 

[Lasse Langwadt Christense]

mandag den 30. september 2019 kl. 12.22.07 UTC+2 skrev Winfield Hill:

Steve Wilson wrote...

On Monday, September 30, 2019 at 4:45:49 AM UTC-4, Winfield Hill wrote:
We need a 25kV rms AC signal. OK, that's 70kV p-p.=20

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

Try the ignition coil from a 4 cylinder wasted spark engine. ...

So if you can use pulses instead of a sine wave, this may
be your ticket.

Thanks, but they'd need to be able to make 5ms pulses,
that's ms, not us.

drive it with a few hundred V AC? they usually have a ratio around 1:100

 

On Monday, September 30, 2019 at 4:45:49 AM UTC-4, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.


--
Thanks,
- Win

What exactly is your output waveform? It sounds like a 70kVpp of 5-10ms pulses at 60-400Hz rep rate.

 

[Robert Baer]

Winfield Hill wrote:

Jeroen Belleman wrote...

Winfield Hill wrote:
DecadentLinuxUserNumeroUno@decadence.org wrote...
Winfield Hill <winfieldhill@yahoo.com> wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.
https://www.advancedenergy.com/products/high-voltage-power-supplies/

Yep, nice stuff, lots of stuff, but all DC not AC.

Win, I don't see how you can expect to get by with 10uA.
Parasitic capacitances alone will take more than that!

What's this for?

This is for AC-driven electrospinning. OK, at 60Hz,
with 100pF wiring, the AC load is j26 M-ohms, making
an AC current of about 1mA. At 400Hz it'd be 6.3mA,
consuming 157 watts if in phase. OK, that's a point.
Maybe a 400VA transformer isn't overkill after all.

Take a bunch of scrap AC power transformers that are identical size,
tear them apart and stack them 4-6 inches, and wind your own primary
first (near the core, so insulation is symple), and zig-zag (accordian)
wind secondary.
Winding start is ground end,naturally.
Insulation needed is layer-to-layer voltage plus some safety (2X maybe).

AC power transformers are good to about 100KC depending on the drive.

Above is for continuous use; can be much smaller if used in pulse mode.
Experiment with an AC transformer soaked in transformer oil for a few
daze, input pulses of (say) 1KV..

 

[Bill Sloman]

On Monday, September 30, 2019 at 11:27:01 PM UTC+10, Bill Sloman wrote:

On Monday, September 30, 2019 at 6:45:49 PM UTC+10, Winfield Hill wrote:
We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

My thought would be an extended ferrite core - made up of perhaps two of these sets, with the two U-sections separated by the two I-sections - to make a toroid with enough space in the middle to make room for a useful number of turns of 25kV-capable cable.

http://www.farnell.com/datasheets/2138476.pdf

You could extend it further with more I-sections

http://www.farnell.com/datasheets/1595819.pdf

None of them are cheap.

25kV for 5msec is a lot of volt-seconds. Even with gaps, you might end up needing a very long path through the ferrite to avoid saturation.

It might give you a way of setting up Piotr Wyderski's array of lots of transformers with enough room for cables that wouldn't arc through.

There's nothing like sleeping on a problem to drag out the other half of the solution.

The high voltage windings would pancake wound, as stack of printed circuit boards.

Each board would be about 120mm long by about 60mm wide, and would have two square holes about 29mmx30mm to accommodate the (almost square) legs of the extended core.

Each board would have - say - 500 turns (250 around each leg) which (at around 1turn per volt) would develop about 500V. There would be fifty of them to generate your 25kV.

You'd connect each board to the next with strip of berylium copper.
For 1.6mm thick boards, the spacing between board would be about 5mm when you were soldering on the strip. You then flip the first board onto the second, solder on another Be/Cu strip, and add the next, Z-stacking them until you have your fifty.

Then thread the ferrite cores through the stack - which is going to end up 80mm high (so two of the U-cores would give you enough winding space).

The energising/driver winding is left as an exercise for the reader.

The whole thing would go into a metal box with two 25kV coax sockets sticking out either side, and you'd fill it up with transformer oil or perhaps a high-voltage soft potting compound.

An obvious variant would be to use a pair of C-cores to make a toroid, feeding a similar z-stack of boards (with only one central hole) to create the high voltage winding.

--
Bill Sloman, Sydney

 

Wond <lost@the.ether.net> wrote in news:qmt6rt$j4j$1@dont-email.me:

On Mon, 30 Sep 2019 01:45:42 -0700, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast procured on
eBay, but I fear it was thrown away in a downsizing move, damn!
With that we could make a 10kV rms AC signal, at a frequency of
our choice. And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz amplifier /
resonant-transformer for a mass spec. That was 10kV amplitude,
or 20kV p-p. RIS-480 / RIS-514 was a compact 7.5kV 300kHz
version. But now we need more voltage, and at much lower
frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at 24kV 400VA.
I think it's molded, rather than oil- filled. OK, at least it's
not a pole-pig. But, haha, we only need 10uA of current, so
something much smaller would be nice.

50 or 60Hz are probably good frequencies for us, although it'd
be nice to experiment with a little bit higher, maybe 400Hz.
Hmm, many 60Hz power transformer are happy at 400Hz. Use a
class-D amplifier to drive a 50-to-230V step-up xformer, then on
to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers / transformers,
many from Alibaba, for electrostatic precipitators, etc. Some
are sold simply to make sparks, but these are short-duration
pulse types, and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type. Hmm, only 10 to
15kV. And most are big and bulky.

The old oil-furnace ignition transformers would initiate and
sustain
an arc about an inch and a half long, in air. Dunno what voltage.

They are like 25kV @ 1mA. But any loading, even the 1mA drops
(clamps) it way down.

Cause I hit myself with one way back in the '70s when I used one
for a jacobs ladder. Then I touched it. Once. Good thing I am
still here.

 

[Winfield Hill]

bloggs.fredbloggs.fred@gmail.com wrote...

On September 30, 2019, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

What exactly is your output waveform? It sounds like
a 70kVpp of 5-10ms pulses at 60-400Hz rep rate.

The industry standard is 50Hz sine wave, often 24kV rms.


--
Thanks,
- Win

 

Winfield Hill <winfieldhill@yahoo.com> wrote in
news:qmsl3202udc@drn.newsguy.com:

Piotr Wyderski wrote...

Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

~100 small 230V/12V transformers with primaries connected
in series should do the job. This is 10x10 grid of them,
doable and reasonably compact.

Whoa!! Huge arcs from secondary to primary
destroys the transformers, sorry.

Not really. As long as they are all isolated from each other, the
primary to secondary voltage differential will be the same.

Xfmr #10 simply cannot be allowed to be proximal to Xfmr #1

Actually, one wants full isolation of all of them from any of their
bretheren. Then, there will be no arcs.

Take a 2:1 for starts.

# 1 120 in 240 out Pri to Sec separation: 120 Volts
# 2 240 in 480 out Pri to Sec separation: 240 Volts
# 3 480 in 960 out Pri to Sec separation: 480 Volts
# 4 960 in 1920 out Pri to Sec separation: 960 Volts

This is where I would start to worry about winding to winding arcing
(in air).

So, immerse remaining set in fluorinert dielectric fluid or xfmr
oil.

# 5 1920 in 3840 out Pri to Sec separation: 1920 Volts
# 6 3840 in 7680 out Pri to Sec separation: 3840 Volts

Possible arcing:
# 7 7680 in 15360 out Pri to Sec separation: 7680 Volts
# 8 15360 in 30720 out Pri to Sec separation: 15360 Volts

Definite arcing potential even in the oil or fluorinert
# 9 30720 in 61440 out Pri to Sec separation: 30720 Volts
#10 in out Pri to Sec separation: Volts

Transformers can be made with increased primary to secondary
isolation capacity, but then efficiency will start to suffer at some
point.

Yeah, this is not a good method...

 

Winfield Hill <winfieldhill@yahoo.com> wrote in
news:qmsl0t02ual@drn.newsguy.com:

DecadentLinuxUserNumeroUno@decadence.org wrote...

Winfield Hill <winfieldhill@yahoo.com> wrote in
news:qmsfbm02m7s@drn.newsguy.com:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

https://www.advancedenergy.com/products/high-voltage-power-supplie
s/

Yep, nice stuff, lots of stuff, but all DC not AC.

Dag nabbit! Did my old eyes miss that itty bitty tid bit?

Sorry 'bout that.

Yes, a neon sign xformer should do it. If you want, you can
isolate it (read insulate: lucite box) and use a higher input voltage
on it to get where you need to be. Or insulate both from ground and
step up into it.

IOW I am saying it can handle the higher voltage, one just has to
be mindful of any exposed nodes, which I am sure you already know
about.

Another source might be looking into a couple popular Tesla coiler
forums. They often have AC gear sources.

 

[Winfield Hill]

Tauno Voipio wrote...

On 30.9.19 18:17, Winfield Hill wrote:

We use a nice 4.8mm dia wire, AWM 3239, with
a 40kV rating, CSA TV-40. Judd Flexrad HV.


Please be careful with radii of wire ends and
terminations, so that your kilovolts won't
sizzle into thin air as corona.

I use spherical hadware-store fishing weights.


--
Thanks,
- Win

 

[bitrex]

On 9/30/19 9:35 PM, bitrex wrote:

On 9/30/19 8:19 PM, Winfield Hill wrote:
bloggs.fredbloggs.fred@gmail.com wrote...

On September 30, 2019, Winfield Hill wrote:

We need a 25kV rms AC signal.  OK, that's 70kV p-p.

What exactly is your output waveform?  It sounds like
a 70kVpp of 5-10ms pulses at 60-400Hz rep rate.

  The industry standard is 50Hz sine wave, often 24kV rms.



what about the way the 25kV AC electrified railroad traction motors
return power to the grid, low voltage AC to ~1-2k DC by PWM boost-type
converter, DC to 3 phase AC "modified sine" with thyristors, then boost
to 25k with 3 phase transformer.

But in miniature...

You could use something like the 2D21 thyratron for the second stage,
probably! they can take 1300 PIV, switch half an amp and are very fast.

 

[bitrex]

On 9/30/19 8:19 PM, Winfield Hill wrote:

bloggs.fredbloggs.fred@gmail.com wrote...

On September 30, 2019, Winfield Hill wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

What exactly is your output waveform? It sounds like
a 70kVpp of 5-10ms pulses at 60-400Hz rep rate.

The industry standard is 50Hz sine wave, often 24kV rms.

what about the way the 25kV AC electrified railroad traction motors
return power to the grid, low voltage AC to ~1-2k DC by PWM boost-type
converter, DC to 3 phase AC "modified sine" with thyristors, then boost
to 25k with 3 phase transformer.

But in miniature...

 

Martin Riddle <martin_ridd@verizon.net> wrote in
news:5ga5ped2ntd8jqondg6umr64smqtg5gij3@4ax.com:

On 30 Sep 2019 01:45:42 -0700, Winfield Hill
winfieldhill@yahoo.com> wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

Used old portable 60hz X-Ray unit? Take the tube and rectifier out
and you have a 100kv transformer with 120vac input.

Cheers

Old ECG airport X-ray unit has dual 40kV units if one follows your
same instructions (remove multipliers).

The voltage multiplier units alone are worth some bucks.

 

[Martin Riddle]

On 30 Sep 2019 01:45:42 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

We need a 25kV rms AC signal. OK, that's 70kV p-p.

I had a 30kV Trek model 678 amplifier, a huge beast
procured on eBay, but I fear it was thrown away in
a downsizing move, damn! With that we could make a
10kV rms AC signal, at a frequency of our choice.
And maybe transform it up to 25kV.

Years ago I made RIS-333 / RIS-470, a 10kV 600kHz
amplifier / resonant-transformer for a mass spec.
That was 10kV amplitude, or 20kV p-p. RIS-480 /
RIS-514 was a compact 7.5kV 300kHz version. But now
we need more voltage, and at much lower frequencies.

A company, Transzvill, in Budapest, sells the FM-24.
It's a huge 300x363mm AC-line transformer, rated at
24kV 400VA. I think it's molded, rather than oil-
filled. OK, at least it's not a pole-pig. But,
haha, we only need 10uA of current, so something
much smaller would be nice.

50 or 60Hz are probably good frequencies for us,
although it'd be nice to experiment with a little
bit higher, maybe 400Hz. Hmm, many 60Hz power
transformer are happy at 400Hz. Use a class-D
amplifier to drive a 50-to-230V step-up xformer,
then on to a 25kV transformer beast someplace.

I have all types of small high-voltage pulsers /
transformers, many from Alibaba, for electrostatic
precipitators, etc. Some are sold simply to make
sparks, but these are short-duration pulse types,
and we need 2-10ms pulses, too long. TV flyback,
15kHz, nah. Also, Tesla-coil types need not apply.

Ah, maybe a neon-sign transformer of some type.
Hmm, only 10 to 15kV. And most are big and bulky.

Used old portable 60hz X-Ray unit? Take the tube and rectifier out and
you have a 100kv transformer with 120vac input.

Cheers