Fixing the alien translators mistake

Now the alien translator wrote the PMTs I am using needed 500 V. But deciphering the documents that came with the PMTs revealed those need 1750 V.

The avid reader of this newsgroup will have noticed I made a nice 500 V design, stabilised, and all that, tested too, used little power, see thread little 'audio transformer'...

So could I 'push up' that design to say 2 kV? (seems a safe value if it is OK at 2 kV it should work at 1750 V).

So I tried a few things yesterday, I could get reliably to 1kV with the existing design, as published, but at about 1200 V the transformer flashed over, and I had to wind a new one. Insulated it better, but now there was a much bigger problem. Those alien PMTs need a voltage divider that cannot have a higher Ri than 1.1 MOhm... Now 1.1 MOhm at 500 V is 250 mW. But 1.1 MOhm at 2 kV = 4 W! the design was to be able to run an battery - for a reasonable amount of time, say a 12 hour day -.

Shattered dreams, went to sleep, to 'sleep on it'. This morning after waking up it occurred to me, after thinking about zeners, transistors with resistors, and all that sort of stuff to make all the voltages and replace the resistor divider for the PMT, and get power consumption down, to use a multi-winding transformer. Such a transformer would have a winding for each anode of the PMT, and a diode, and a capacitor, and all that in series. I was faced with a bigger core (for space, and insulation), and a multi section (10 actually) coil former, all needed to be specially made, isolation problems at the transformer connections.. 2 kV is not nothing.

So that sort of looked like the end of this design adventure. Now I gave up, and put the problem back into the subconscious area of the brain to have it work out the impossible. Engineering is after all making new things from the things you HAVE. I did not have such a transformer, and the prospects of getting one at a reasonable price seemed too dim. Call it lateral thinking (I think somebody did), brain did a scan of what we DID have, and showed me a picture in my mind of all those nice 33 uH coils I got for almost nothing from ebay (a bag full, bought them for switchers). I had used things like those before with a few extra turns of wire added for feedback in this project:

formatting link
So, and these were very nicely insulated, brain then pictured those with more turns on it... primaries in parallel. So I took one, put 10 turns on it, fed it from the MOSFET in the old design, to figure out the number of turns it had.

33 uH, 30 turns! perfect! I got 60 V on it with almost zero drive, so 2 volt / turn! That would mean for 100 volt 50 turns. So I wound 100 turns on one, added a diode and capacitor, and measured a nice 200 V DC. Takes about 1 minute to wind the 100 turns by hand. So, now, putting 9 of these coils + diode +cap, each with the appropriate number of turns to get the next PMT voltage, in series, solves all problems, first standard components, second I already have them, third power consumption was almost zero, (no resistor load! in fact I had to add some load to be able to make any measurements - 10M probe), very cheap parts (coils were 5 Euro for 20 IIRC), and it solves the insulation problem in a beautiful way, the caps can be low voltage caps (250 V type), and small, these diodes are SMD I already had, and the total board space is rather small. There will be some more RFI with these open coils than with the closed E cores, but it will be in a metal box, so that does not matter that much. Test setup
formatting link
I think I corrected the Alien translators mistake.

PMT ... . ... . \ / k . ----------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a1. ---------------------- . . |-------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a2. ---------------------- . . |--------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a3. ---------------------- . . |-------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a4. ---------------------- . . |--------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a5. ---------------------- . . |-------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a6. ---------------------- . . |--------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a7.---------------------- . . |--------|>|-------- . . | )| ------- +5V . . === )|( . . | )| --------------A . a8 ---------------------- . . |------------------------------------- 10 M --------- voltage sense = current limiter reference to PIC comparator . . |--------|>|-------- | . . | )| ------- +5V [ ] . --- . === )|( | ....|.... | )| ------------- A /// | ------------------- | anode | | /// ---| ---| |------- PWM from PIC |---| | IRLZ34A |------------- current sense to PIC comparator [ ] 1 Ohm | ///

-- **** * * * * * * * * * * * *** * ** ***** *** * ***** ** *** **** * * ** * * * * * * * * * * * * * * * ***** * * * ***** * * * * * * * * * * * * ** * * * ** **** * ** * ****

Reply to
Jan Panteltje
Loading thread data ...

Hm..when I had to do this, I used a transformer that was made for CCFL-Tubes. It is more funny to buy one than to wind one. And they are easy to buy or to rob from some old PCB.

What is the name of the PMT? Is it a Channel-PMT? I think normal ones are working around 1kV. With 1.7kV I would expect to much amplifiering and noise.

Olaf

Reply to
Olaf Kaluza

On a sunny day (Tue, 11 Oct 2011 14:48:54 +0200) it happened Olaf Kaluza wrote in :

1750 V.

Sorry I am under NDA for the alien project, but they are really maximum 1750 V. (But you have probably seen 'mars attacks' and 'independence day'). I would probably end up operating those PMTs at around 250 to 1500, I had already a signal at 1000, and it goes up (the gain) log. The CFL transformers could work, but the power consumption problem would stay.

30 minutes later:
formatting link
I would all the coils, tested them, and did a polarity check (some of the thin wires are marked black, but that is not very visible in the picture). Note many different coils, some with 200 turns, some with 100 turns, and some with 150 turns, to get the correct PMT anode voltages. Next is the flash-over test, if flash over happens at 2 kV then I will have to put some tape under the coil for the highest in the chain, and rewind. Only a minute work, this beats any transformers.
Reply to
Jan Panteltje

Nice approach! :)

Suggest that existing toroid inductors can have a 5 V *primary* wound on them easily and quickly to minimize RFI issue.

--Winston

Reply to
Winston

On a sunny day (Tue, 11 Oct 2011 07:18:45 -0700) it happened Winston wrote in :

Thanks :-)

Dunno what you mean by that?

I put every[1] coil secondary in series, and all primaries in parallel,\ and could get away over 2 kV easily without flash-over.

formatting link

[1] I already killed 2 meters this month, so after reaching 2 kV ++ I left some coils out of circuit, also to protect my scope. Probing more than 2kV with the 10x probe is not so good.

It passed the flash-over test, the stabilisation test, and the power consumption test, less than a watt altogether for the whole circuit. But I will still have to try with a diode fan cap for each coil. Do not have enough caps at the moment. Looking at the drive current in the MOSFET I think saturation will never happen, nice linear ramp, those little 33 uH coils are for 10 A switchers IIRC :-)

Reply to
Jan Panteltje

On a sunny day (Tue, 11 Oct 2011 08:49:55 -0700) it happened Winston wrote in :

OK, I get the idea. But hard to beat the one minute per coil it took winding these. Toroids with 150 turns and that small easy to obtain? I find it very hard to wind turns on very small ring cores.

Reply to
Jan Panteltje

On a sunny day (Tue, 11 Oct 2011 08:30:57 -0700) it happened John Larkin wrote in :

me:

sense = current limiter reference to PIC comparator

to PIC comparator

John:

Yes, I thought about that, but at least on this PMT the voltages are different for some anodes, focus, whatever. You would have to use the lowest common denominator (? correct English?) for the drive, and many many sections. Now I am done with 9 small coils, the voltages are interlocked by the winding ratios, NO DRIVING TRANSFORMER AT ALL, no saturation problems, no isolation problems, easy to stabilise (stabilise the lowest and they are all OK), 9 SMD diodes,

9 small caps, 9 cheapo inductors, a little bit of wire.

I like it.

Reply to
Jan Panteltje

MOhm...

a reasonable amount of time, say a 12 hour day -.

Ugh, battery-powered and PMT is a difficult combination. Yes, some kind of distributed power supply could help a lot, the resistive voltage divider is a time-honored but very inefficient way to run PMTs.

You need to watch out for noise in the distributed scheme, small fluctuations in dynode voltages have large effects on gain, and will smear out your histograms, ie. make amplitude resolution worse.

Jon

Reply to
Jon Elson

(...)

A good thing quickly is inferior to an excellent thing slowly.

Sure! Suggest visit your local hobby supplier and buy a bunch of examples to test. Start on the bottom of page 11:

formatting link

How about removing turns. That's pretty easy, yes?

Also, nothing is stopping you from going gonzo on it. for experimental purposes.

One could glue 'n' toroids together and wind a primary through all the cores, using the existing windings as secondaries. Perhaps.

--Winston

Reply to
Winston

d 1750 V.

design,

le 'audio transformer'...

existing design,

n 1.1 MOhm...

y - for a reasonable amount of time, say a 12 hour day -.

ers, transistors with resistors,

stor divider for the PMT,

diode,

section (10 actually) coil former,

onnections.. 2 kV is not nothing.

e brain to have it

reasonable price seemed too dim.

t we DID have,

or almost nothing from ebay

for feedback in this project:

more turns on it...

ign,

e number of turns to get the next PMT voltage,

ady have them,

d to add some load to be able to make any measurements - 10M probe),

lation problem in a beautiful way,

small.

cores, but it will be in a metal box, so that does not matter that much.

--
> >. =A0 =A0 a1. ----------------------
> >. =A0 =A0 =A0 . =A0 =A0|-------|>|--------
> >. =A0 =A0 =A0 . =A0 =A0| =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0)| ------- +=
5V
> >. =A0 =A0 =A0 . =A0 =3D=3D=3D =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 )|(
> >. =A0 =A0 =A0 . =A0 =A0| =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0)| ---------=
-----A
> >. =A0 =A0 a2. ----------------------
> >. =A0 =A0 =A0 . =A0 |--------|>|--------
> >. =A0 =A0 =A0 . =A0 | =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 )| ------- +5V
> >. =A0 =A0 =A0 . =A0=3D=3D=3D =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0)|(
> >. =A0 =A0 =A0 . =A0 | =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 )| -----------=
Reply to
Bill Sloman

1750 V.

design,

'audio transformer'...

existing design,

1.1 MOhm...

for a reasonable amount of time, say a 12 hour day -.

transistors with resistors,

divider for the PMT,

diode,

section (10 actually) coil former,

connections.. 2 kV is not nothing.

brain to have it

reasonable price seemed too dim.

DID have,

almost nothing from ebay

feedback in this project:

more turns on it...

number of turns to get the next PMT voltage,

have them,

add some load to be able to make any measurements - 10M probe),

insulation problem in a beautiful way,

small.

cores, but it will be in a metal box, so that does not matter that much.

sense = current limiter reference to PIC comparator

to PIC comparator

That doesn't surprise me a bit. And it doesn't surprise me that you didn't bother to google it.

formatting link

formatting link

and lots more.

Obviously, you can add resistors, or hop C-W taps, or whatever where you need to. Use a little imagination, for Pete's sake. Jan did.

It would also be interesting to build a C-W stack that was driven from two or more different AC pump supplies, taps on a transformer maybe.

John

Reply to
John Larkin

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

This type of stacked multiplier is common, as it eliminates a lot of secondary capacitance in the transformers. I would love to find the orignal or the origins of it. But so far have only found traces of low voltage implemtations.

Cheers

Reply to
Martin Riddle

need 1750 V.

V design,

ittle 'audio transformer'...

he existing design,

than 1.1 MOhm...

tery - for a reasonable amount of time, say a 12 hour day -.

zeners, transistors with resistors,

esistor divider for the PMT,

d a diode,

ti section (10 actually) coil former,

r connections.. 2 kV is not nothing.

the brain to have it

t a reasonable price seemed too dim.

what we DID have,

t for almost nothing from ebay

ded for feedback in this project:

ith more turns on it...

design,

iate number of turns to get the next PMT voltage,

lready have them,

had to add some load to be able to make any measurements - 10M probe),

nsulation problem in a beautiful way,

er small.

d E cores, but it will be in a metal box, so that does not matter that much= .

pg

+5V

------A

- +5V

--------A

+5V

------A

- +5V

--------A

+5V

------A

- +5V

---------A

+5V

------A

5V

-----A

---- voltage sense =3D current limiter reference to PIC comparator

=A0 =A0 |

+5V =A0 =A0 =A0 =A0 =A0[ ]

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|

A =A0 =A0 =A0///

=A0 |

=A0 =A0 =A0 =A0 =A0 |

=A0 =A0 =A0 =A0 =A0 =A0 ---|

=A0 =A0 =A0 =A0 =A0 =A0 =A0 ---| |------- PWM from PIC

=A0 =A0 =A0 =A0 =A0 =A0 =A0|---|

=A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A0 =A0IRLZ34A

=A0 =A0 =A0 =A0 =A0 =A0 =A0|------------- current sense to PIC comparator

=A0 =A0 =A0 =A0 =A0 =A0 [ ] 1 Ohm

=A0 =A0 =A0 =A0 =A0 =A0 =A0|

=A0 =A0 =A0 =A0 =A0 =A0 /// =A0

It wasn't intended to "surprise" anybody. It's a simple statement of fact.

Surprisingly little "surprises" you. You don't know much, and are depressinglyably gullible.

neither of which work for me.

It's the kind of solution that would appeal to the technically ingenious. It appeals less to people who have had to make their photomultiplier power supplies very quiet to avoid modulating the the photomultiplier gain.

The crucial features of most photomultiplier power supplies are stability and adjustability - the photomultiplier gain is typically sensitive to about the eighth power of the overall supply voltage (depending on the number of dynodes and the secondary electron multiply surface on them) and you need fine control of the voltage to get the gain you need, and high stability to keep the gain at the desired value. The Cockroft-Walton multiplier isn't a wildly attractive voltage source in this kind of context.

Interesting perhap, but definitely stupid.

-- Bill Sloman, Nijmegen

Reply to
Bill Sloman

jpg

+5V

-------A

-- +5V

---------A

+5V

-------A

-- +5V

---------A

+5V

-------A

-- +5V

----------A

+5V

-------A

+5V

------A

----- =A0

=A0 =A0 |

+5V =A0 =A0 =A0 =A0 =A0[ ]

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0|

A =A0 =A0 =A0///

=A0 =A0 |

=A0 =A0 =A0 =A0 =A0 |

=A0 =A0 =A0 =A0 =A0 =A0 ---|

=A0 =A0 =A0 =A0 =A0 =A0 =A0 ---| |------- PWM

=A0 =A0 =A0 =A0 =A0 =A0 =A0|---|

=A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A0 =A0IRLZ34A

=A0 =A0 =A0 =A0 =A0 =A0 =A0|------------- =A0

=A0 =A0 =A0 =A0 =A0 =A0 [ ] 1 Ohm

=A0 =A0 =A0 =A0 =A0 =A0 =A0|

=A0 =A0 =A0 =A0 =A0 =A0 ///

It seems to have been invented in 1919 by Heinrich Greinacher. Cockroft and Walton made it famous.

formatting link

-- Bill Sloman, Nijmegen

Reply to
Bill Sloman

On a sunny day (Tue, 11 Oct 2011 12:52:25 -0700) it happened Winston wrote in :

Well, this is next to a perfect thing, so it cannot be beaten:-) I do not see why a toroid would help in any way except reduce external magnetic fields, and those are not very relevant in this application.

Not if it is a few hundred.

Time.

You need space for isolation.

Reply to
Jan Panteltje

On a sunny day (Tue, 11 Oct 2011 14:48:45 -0500) it happened Jon Elson wrote in :

MOhm...

a reasonable amount of time, say a 12 hour day

Exactly, I considered that, as in the normal resistor divider chain you always see something like this: A7 A8 A9 | | |

-1750 - R - R - R - ... --- R --- R --- R - GND | | | === === === | | | /// /// /// So the last 3 anodes are usually decoupled, peak currents.. but also the whole thing forms a RC RC RC lowpass to make the last anode the cleanest.

So in my application I will add some RC lowpasses from the last (lowest) power feeds to the last anodes. Still needs only a few small relative low voltage caps, and uses no power.

PMT

A5 -----------< +500 V

A6 ---- R ----< +400 V | === | ///

A7 ---- R ----< +300 V | === | ///

A8 ---- R ----< +150 V | === | ///

Reply to
Jan Panteltje

On a sunny day (Tue, 11 Oct 2011 16:18:15 -0700) it happened John Larkin wrote in :

Thanks, but a whole lot of diodes and capacitors :-)

As they mention, the PMT's response to light becomes more linear than with a resistive divider. That one seems to have all the voltages the same.

In my case the 'granularity' is 50 V, so without extra dividers as in your fist pdf example, for 1750 V I would require 35 stages.

70 capacitors and diodes.... board space then becomes an issue..

I think there is plenty of opportunity for experimentation. but also I have to deliver the goods so to speak, so there is a compromise and decision to be made. I have also plans to add some extra RC filters for the last anodes, see my other reply in this thread with a diagram. That only requires a few extra capacitors and resistors. Not even sure if it is needed with these huge signals, will have to test that.

Reply to
Jan Panteltje

(...)

You can lower the e-field emission, too by making half the coil a 'return winding'.

formatting link

(...)

It might be just a few, or none. Who knows?

Here, just wind them from scratch. Takes less than 10 minutes even for large coils (Evil Grin):

formatting link

Aw, come on. You work faster than *that*. :)

That is why God made fish paper washers.

formatting link

Besides, you now have a short arc path through your paralleled primaries in your solenoid - wound transformers.

:)

--Winston

Reply to
Winston

It seems to have been invented in 1919 by Heinrich Greinacher. Cockroft and Walton made it famous.

formatting link

No, not the Walton generator. But the Stacked Primaries on the transformers.

Cheers

Reply to
Martin Riddle

1750 V.

design,

little 'audio transformer'...

existing design,

1.1 MOhm...

for a reasonable amount of time, say a 12 hour day -.

zeners, transistors with resistors,

resistor divider for the PMT,

diode,

section (10 actually) coil former,

connections.. 2 kV is not nothing.

brain to have it

reasonable price seemed too dim.

we DID have,

for almost nothing from ebay

for feedback in this project:

more turns on it...

design,

number of turns to get the next PMT voltage,

already have them,

to add some load to be able to make any measurements - 10M probe),

insulation problem in a beautiful way,

small.

cores, but it will be in a metal box, so that does not matter that much.

sense = current limiter reference to PIC comparator

PIC

sense to PIC comparator

Well, "work" is not your favorite activity.

All ideas seem stupid to you, unless you find them in peer-reviewed journals.

When I was a kid, 931As cost 99 cents each from Fair Radio Sales. I did all sorts of stuff with them. When I was a junior in high school, I won the state science fair in Physics with a scintillator project, and senior year I went to the national science fair in Baltimore with another PMT-based project. (I hung out with Amory Lovins, now the AGW guru. He was a little strange even then.) I generally used a 60 Hz power transformer, a string of VR tubes for regulation, and a resistive divider. In retrospect, I could have been much cleverer, like using the VR string to power the dynodes directly.

John

Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.