Voltage Multiplier (CW) - some clarification / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Voltage Multiplier (CW) - some clarification

1 2 next

Move Thread

LAN_403

yroethiel

Sun Nov 29 2009, 12:20AM

Registered Member #2488 Joined: Sat Nov 28 2009, 11:26PM
Location: UK
Posts: 8

Hey everyone, been browsing around a bit and I have to say really like the feel of the community from what posts I've read! So...hi :).

Getting straight to the point, I'm constructing a full-wave CW voltage multiplier, and am studying degree-level physics, but not electronics :P. I've read the rules of the site, I've done days and days of research - what I'm hoping to do here is present my plan, and ask for verification/problem predictions/reassurance so that I can address any issues before making it! Hopefully my reading is sound and you'll be able to just say yay/nay/consider this - the friendly atmosphere I've seen will hopefully help out?

Objective: 5kV DC Power supply (negative relative to ground). Lower current is good, I just want big fields. Voltage ripple below 5% if possible. Negative DC from diagram attached (diodes reversed).

Input: 0-240V Variac, 50Hz (running off UK mains).

CW Stage equation: 2*Epk*Stage - I need 10 stages to get up to 4.8kV. Using simulator at Blaze Labs, 11 stages more likely.

As far as I understand, the components only ever see the stage voltage, so having high value diodes/capacitors not really necessary. Catering for 2x (safety margin) of input 240V should be fine. So diodes can be good (CHEAP) old 1N4007.

Capacitors are slightly more confusing. The smoother the DC out the better, so higher capacitance is better according to my reading. I have a WHOLE bunch (45+) of photoflash capacitors (salvaged from disposable cameras) with 80-100uF 330V ratings - will these do? They're not exactly designed for prolonged usage. Otherwise, these (0.15uf, 630V radial caps).

If this is all alright, could you let me know...and if not, what I need to sort out? These specs seem to correlate with some previous posts in these forums by other newbies.

Now slightly beyond the verification/reassurance of the above, some more direct questions!

Firstly, I plan on building this into a silicone oil filled project box, but want to be able to tap out of the CWVM at various points using switches on the outside of the case (and indicator LEDs to show which stages are active). I have some SPDT 250VAC 3A On-On toggle switches, but don't think these will do the job. Assuming using one of these switches to redirect stage 6's Vout to the output terminals (isolating 7-10), the switch has the full multiplied voltage relative to the ground across it. Although the P=IV (250x3) is pretty high, I assume that 250V rating is there for something...is this safe? What's the best option for using some form of switch to tap out?

In the same vein, the LED indicators. Assume stages 1-6 are being used, and 7-10 are isolated from the power supply. I want an LED related to stage 6 to light up - is this a case of having an LED between 5-6 (therefore only receiving the stage voltage), and using a 0.15uF 630V capacitor in series with the LED to limit it to about 10mA. Stick a 220ohm resistor in series to protect it from spikes. Problems I can maybe foresee - sticking another cap into the system might mess it up? My knolwedge isn't good enough to predict what will happen. If this technique is wrong, I guess I'm asking quite a big question. How do I tap out at specific stages using switches, safely, and have an active LED indicating what's active? Without an extensive knowledge of the components available easily out there, I'm just relying on what I vaguely think is right and can be found on Ebay, suggestions really, really welcome!

Turned out much longer than I hoped, urgh. Any help that you guys can offer will be really appreciated, there's a lot to this project and this power supply is a major thorn in my side because I just don't know enough about the electrics-DIY field! There's so much conflicting advice/demonstrations/tutorials/12 year olds on the internet that I've grown desperate enough to ask :P.

doctor electrons

Sun Nov 29 2009, 01:13AM

Registered Member #2390 Joined: Sat Sept 26 2009, 02:04PM
Location: Milwaukee Wisconsin
Posts: 381

Hello and welcome to the forum! I do not have a lot of answers for you as my experience with cw's is limited. BUT,
Check out this thread, maby ask a few questions there.

Sounds like what you are shooting for! There is also some discussion relating to the diodes.
Good luck with the project, hope this link steers you in the right direction!!
If not, there is a great amount of genius here that will surely have the answers you need!

yroethiel

Sun Nov 29 2009, 01:30AM

Registered Member #2488 Joined: Sat Nov 28 2009, 11:26PM
Location: UK
Posts: 8

Hey,

Thanks for the quick reply! I did have a read through of that post (first in search), but it mostly discusses ratings 20x times more than what I'm looking at. I decided not to post in there so as not to detract from the original poster's crazy work - mine's on a lot smaller scale and hopefully the discussion at a lower level!

doctor electrons

Sun Nov 29 2009, 02:01AM

Registered Member #2390 Joined: Sat Sept 26 2009, 02:04PM
Location: Milwaukee Wisconsin
Posts: 381

Sorry friend! Thought that may help, but i guess not. Be patient, someone will lend a hand who has more cw knowledge than myself.

Proud Mary

Sun Nov 29 2009, 02:27AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hi there lad,

Welcome to our little fleck of space-time where we look out on the Universe and hope we see Reason.

Been here a few years mysel - changed from one sex to another in that time - but to the quality folk in this forum that don't matter zip. Does the Man know what he bin talkin bout, or doan e? That's all that counts here.

So you wanna get right with the C&W? No problem, man. You got things tail over tits man with that scheme, Man, like something pulled out a school book.

So listen up, babe: FIRST you defines the needs of your applicashun, how many Kv at how many mA to feed its face, then you design your C&W to meet those needs.

The full-wave circuit you posted is what I'd call a school-book circuit - something that works out in class room math but will cost you the first time you switch it on in the real world.

Look at the circuit you done drawn for full-wave, sweetheart. S'pose you got yoursel a dead short down to the hull - every single one of your diodes is blown away. So you gotta stick some damper resistors in every stage, stop those bad farads downloading their selves
back down the chain into the hull in a fault condition.

You jus put 50R 10W before and after every diode and you got yoursel a much more stable C&W right away.

If you're really gonna do it, I'll help you through it with parts, circuits and all that good stuff, as I have many others over my years here, but I am too busy to help fantasists, people that haven't bothered to learn Ohms Law by heart, and sexual inadequates with odd, unhealthy ideas and interests about gender variant people like myself,

Be seeing you,

Stella X

doctor electrons

Sun Nov 29 2009, 02:35AM

Registered Member #2390 Joined: Sat Sept 26 2009, 02:04PM
Location: Milwaukee Wisconsin
Posts: 381

Thanks Stella! Once again you have made up for my lack of experience! With a southern drawl even!

See yroethiel, i knew someone would have those answers! Now you will go in the right direction!

Proud Mary

Sun Nov 29 2009, 02:54AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Mike, as you know, I'm proud of my time at MCLB Barstow-Dagget and NTC Fort Irwin, and my experiences up there in the High Desert have shaped many of my views for the better.

The very best, kindest soldiers and sailors I have ever met in my life - families and all rotated through NTC Fort Irwin - normal women with normal husbands and children - never appear in the movies, where everyone is either a cold-hearted killer or a mental case suffering out the Syndrome.

Don't know what to say after that, Mike.... voltage multiplier clarification.

Stella

Dr. Slack

Sun Nov 29 2009, 08:12AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Be very wary of photo-flash caps at anything like their printed voltage. They are that size and dead cheap for a reason. Any cap which proclaims "working voltage 330v, surge voltage 350v" is asking to be fried, check the difference between working and surge on decent caps. That said however, I tend to use them for smoothing etc, just don't take them beyond say 200 to 250v if you expect them to last. If you series several caps to make one higher voltage one, then make sure you have balancing bleeder resistors in parallel with each cap, you cannot assume the leakage currents are balanced and the least leaky cap in the string will get over-volted.

The alternative cap you quote is only 0.15uF. Comparing a 100uF and 0.15uF caps as alternatives sounds a bit strange, unless of course the thing that they have in common is that they are both free?

You say you don't need much current output. You could build a half-wave CW. A lot fewer components, but admittedly more ripple.

Are you building kit for a specific purpose? In which case Stella's specification and robustness first approach is a good idea. Or are you just getting a feel for what stuff does by lashing the contents of your junk box together? In which case all the components going bang is not really a problem as long as they don't take you with them.

Sulaiman

Sun Nov 29 2009, 10:42AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

I have a few thoughts that may be useful,

1) You MUST add sufficient resistance in series with the output to prevent a current pulse that would destroy the diodes when (not IF;) the output shorts/arcs to ground.
e.g. 5 kV / 1 kOhm = 5 A peak which 1N400x can handle.
Since a 5 kV rated resistor is expensive, use several smaller resistors in series
e.g. 10x 100 Ohm 500V etc.

2) Start with a mains line filter and fuse and a TVS such as 1.5KE400 or 440 to keep mains 'spikes' within the ratings of your multiplier.

3) For 5 kV you do not need oil immersion or potting - just reasonable tracking distances and no sharp points. Once working observe in darkness and look for any purple-ish corona.
(be careful in the dark !)

4) I use a bank of 120 uF 330 V photoflash capacitors at 3.8 kV (12 series x 8 parallel). Very good for low ripple BUT a terrifying quantity of stored energy!
I advise using no more capacitance than necessary.

5) You will need 3 photoflash capacitors per stage, in my bank I ensure voltage sharing by having 2x 160V zener diodes in series across each row of capacitors.
You can use resistors but they waste power.

6) In your diagram the central column of capacitors hold the voltage, the outer two rows 'pump up' the voltage for each stage, so if you want low ripple but don't need high current you can make the central column from photoflash capacitors and use two of the .15 uF 630 V caps in series per stage for the outer columns.
You will need zeners or resistors for the central column, the outside columns will not need balancing zeners/resistors.

7) I think that CW multipliers are good for VERY high voltages (per C-W original use), or cheap 'bug-zappers' etc. (and Photo-Multiplier Tube supplies)
If I wanted a 5 kV supply I would go for a flyback inverter because;
7.1) I feel safer using a low voltage dc power supply then one directly coupled to the mains
- in experimental systems things DO go wrong in unexpected ways,
if you look at your diagram as drawn only the bottom two outside capacitors separate the output from the mains, the diodes can easily pass a lethal current from the mains to YOU.
Your life could depend on the reliability of those capacitors !
7.2) An inverter would allow continuously variable output voltage.
7.3) Due to the higher operating frequency smaller capacitors can be used to obtain the required ripple, BUT you will need faster diodes.
The simplest arrangement would be a car (automobile) ignition coil driven by a mosfet/igbt and a 555/494 etc. (Since you need a negative supply the common tv/monitor flybacks are no use as they have inbuilt diodes for +ve eht only).

8) Another option would be; Variac - Microwave Oven Transformer - voltage doubler using microwave oven capacitors and diodes. The variac would allow continuously variable output. A series output resistor will still be required.

9) There's no way I'd touch a normal switch at 5 kV, contact-to-contact rating is irrelevant - it's contacts to earth/you that's important.
High voltage switches and relays are expensive, you'd be better off powering down and changing a tapping point.

10) Some method of safely ensuring that the capacitors are completely discharged after use is required. (trust me on this !)
My scheme of zeners for example would leave almost full charge on the capacitors for hours after power down, so if some power loss is acceptable then use balancing resistors. If not you will need a load to discharge the capacitors, worst case - discharge the bank via the resistors used as a balast at the top/output end of the multiplier.

Some food for thought !

What voltages/currents/ripple do you need ?

P.S. it's a lazy Sunday morning hence the ramblings...

yroethiel

Sun Nov 29 2009, 12:40PM

Registered Member #2488 Joined: Sat Nov 28 2009, 11:26PM
Location: UK
Posts: 8

Sorry friend!

That's why I signed up! The eagerness to help noobs is great :).

gender variant people like myself

If someone's willing to share some good knowledge with me, they've got my attention and respect right there. Anything else is diddly-squat unless someone makes a point of it. 'Nuff said on that. Texan drawl though? :P

Back on topic. I know my electronics equations and a fair bit of theory, it's the practical side I'm lacking which is why I'm here. On paper I can see the previously posted circuit providing a negative DC output, but blowing it up is exactly what I'm trying to avoid! I've got the basic practical skills from messing around with IC circuits and a whole bunch of low voltage stuff, but need to make sure I get this right as I'm sure you all understand.

Comparing a 100uF and 0.15uF caps as alternatives sounds a bit strange, unless of course the thing that they have in common is that they are both free?

Bingo! Exactly right, freebies although I don't have enough of the 0.15uFs as it stands.

What voltages/currents/ripple do you need ?

OK, maybe this'll be easier if I describe everything! I'm studying a cold cathode electron gun in terms of performance, varying several parameters (accelerating voltage, materials selection, vacuum, deflection systems, etc.), trying to optimise it for an end solution that includes an oscilloscope-like function. I'm a poor student trying to put together every component as a learning exercise; the electronics are a means to get the electron gun working, not a main focus. That said, I'm keen to learn more through actually making the thing - but simplicity a key focus.

So, I need a highly negative cathode relative to the grounded anode; electrons will be accelerated across a vacuum to the screen, as standard in a CRT.

Keeping ripple low is good so that I can actually have a good prediction of the voltage characteristics being applied to my cathode for the investigation's purpose. 5% should be good enough, if a half-wave can do that then great. Current as low as possible, primarily for safety; one milliamp or even lower. A mixture of voltages up to 5kV would be ideal; I suspected it would be making physical contacts instead of switches. Bah :P. I plan on putting a 1M ohm+ resistor at the output as a load, probably a giant series-parallel mess of a lot of 0.5W resistors for price reasons.

Mains filter/fuse definitely understood, I was going to use a 100mA rated fuse and hadn't figured out an appropriate filter part number yet. I'd rather use the 0.15uF caps, but given that I have these photoflash ones lying around, I thought I'd mention them.

if you look at your diagram as drawn

In mild defense of noobishness beyond noob, I didn't mean for that diagram to be my proposal, just referring to an "upside-down" version of the normal full-wave found on the net so that the output is negative. As you've obviously all realised by the great advice coming through, I'm now looking for a practical solution to the theory.

Regarding using an inverter, I have a Maplins 0-12V DC supply, and an old inverter salvaged from the cold cathode tube in my old computer rig - the output is 180-800V. Feeding this inverter into the CW is definitely an option, although my Variac has a built-in voltmeter making it convenient for working out the end voltage - hence the attraction to that option. It's just a bit more complicated working out what the final voltage is using the inverter.

In summary, two options as yet - the variac or the inverter. I haven't put much thought into the inverter option. It does require components of a higher rating though and I'm on quite a strict budget. If I use the variac...variac (an old power supply unit used for testing TVs, with built in circuit breaker) to mains filter to fuse to a lot of capacitors/resistors and diodes to a final resistive load and then my cathode - values to still be confirmed through more discussion.

Quick question, regarding using photoflash caps as the central column and 0.15uFs as the outer columns...why use 2 in series? Surely their voltage ratings are already high enough?

[Once I've come up with a diagram y'all approve of, any shopping suggestions welcome! If the 0.15uFs can be replaced by anything more suitable, I don't have enough anyway and these ones look a bit dead; shopping will have to be done regardless. I typically use Farnell, Rapid or Ebay, but if there are any other places (UK) recommended, money is a big issue to this university student.]

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob