Giant Wimshurst Machine - could use some guidance / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Giant Wimshurst Machine - could use some guidance

1 2 3 4 next last

Move Thread

LAN_403

AwesomeMatt

Sat May 10 2014, 09:55AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Hey all, I've been brought in on a team building a big (4' wheel) Wimshurst machine for a science center place.

I'm somewhat comfortable with my knowledge of high voltage (say, "normal" stuff at 10kv, not 100x that) at mains frequency, but not as comfortable with my grasp of static electricity or RF issues. Plus, I know you guys are the people to go to for suggestions on parts.

I try to stay interactive whenever I'm pestering people for advice, since it's the only payoff they get for helping. So, hopefully I'll document along the way.

For the curious, this is a (random) Wimshurst machine:

300x300

<-- Here is an MIT video showing how they work.

A Wimshurst is basically two counter-rotating wheels with brushes and cones to generate HV static, which charges Leyden jars (caps), and then eventually reaches high enough to bridge the spark gap. It sparks impressively then resets.

Here are the rough specs they'd like:

- 4-5 foot wheel
- 18" sparks hopefully (will be 10' off the ground)
- 5-15 seconds between sparks (don't want it too distracting from other exhibits)
- Handle-cranked by the public to turn the wheel

So, right off the bat, I know I'm targeting about 1,350,000 volts. Probably more for safety margin on the caps (any suggestions? I was figuring 30% extra for safety would be okay).

I'm not really that interested in building Leyden jars (historical accuracy is not necessary) or glass plate caps. I'd like to either buy some giant caps, or, build a massive MMC array.

Cost is not really too much of a concern, it's corporate funded and compared to wages, parts is pretty nil.

I've questions:

1 - What types of caps should I use? (Polypropylene was going to be my choice, but, let me know why that could be wrong).

2 - I know I'm building a much more massive MMC than typical coilers will, because it's not the primary stage. It's the full voltage. Any specific suggestions for parts? I figure I'm going to end up chaining 1000+ 1-2kv caps, but if there's a large source of affordable higher voltage caps that would cut down on soldering time and complexity.

I'm sure I'll have a lot more along the way.

Thanks in advance.

Sulaiman

Sat May 10 2014, 01:00PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

For elegance of construction I suggest 'doorknob' ceramic capacitors, e.g.

however, you should consider lower capacitance / stored energy,

If you used these capacitors to make two towers of 25 series capacitors each
the stored energy would be > 30 Joules ... a fairly impressive bang

Wastrel

Sat May 10 2014, 01:40PM

Registered Member #4095 Joined: Thu Sept 15 2011, 03:19PM
Location: England.
Posts: 122

A Leyden jar is just a glass container coated on the inside (and for lower voltages, optionally on the outside) with metal. Your machines capacitance needs to be tiny. Hopefully someone can chime in with actual values for what spark energy is safe without ear defenders. Lose the MMC idea.

I don't know if a 4 ft device will reach the voltages you expect, school models tend to be unimpressive next to van der graaf machines, 75kv versus 250kv. Voltage may scale linearly but lots of problems, capacitance, forces will scale squared. A million volt wheel pushing 100ua might be quite difficult to turn.

With sharpish ends you shouldn't need a million volts but don't rule out an enclosed gap at lower pressure to get the effect at much lower voltage. A sign that says "If using this machine without the enclosed spark gap, ear defenders MUST be worn." might be helpful even if the spark is unimpressive in room air. There is a degree of showmanship to these things.

Build some smaller versions first to get an idea what materials work. Whimshurst machines are temperamental and can fail to work at all if the air is slightly too humid or the moon is in the wrong place...

Dr. Slack

Sat May 10 2014, 02:06PM

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

I've seen really big Wimshurst machines simply use a large sphere on each electrode as the capacitance to ground (annoyingly can't see anything like it in the first few pages of 'google images wimshurst'). At >1MV, only a few pFs is needed to make a healthy bang. Air has a lot to recommend it as a dielectric. Google for sphere to ground capacitance formulae.

Antonio

Sat May 10 2014, 03:36PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

For 18" sparks in a Wimshurst machine you need disks with three or four times this in diameter, maybe a bit less with good insulation and small sectors. Assuming that sparks between sphere terminals only occur with length up to 4 times the diameter of the spheres (see my avatar image), 18" sparks would require 4.5" spheres. The breakdown voltage between two 4.5" spheres separated by 18" can be calculated as 298 kV at sea level. A spark gap with the positive ball smaller can produce this same spark length with less voltage, 151 kV with one 2.25" sphere.
ItÂ´s better to use double balls, as frequently seen in these machines. So, the required voltage is not so high. Leyden jar capacitors can be safely used, and are very easy to make from plastic containers. For greater insulation two can be connected in series at each terminal. The jars can have less than 100 pF of capacitance.

AwesomeMatt

Sat May 10 2014, 09:29PM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Thanks for everyone's feedback so far.

Bear with me here, I have a rough understanding of most of these things, but not solid enough to make decisions based on them.

Wastrel wrote ...
I don't know if a 4 ft device will reach the voltages you expect, school models tend to be unimpressive next to van der graaf machines,

What I keep hearing is that you can get a spark gap 1/3 the size of the wheels. So, 18" gap = 54" wheels (4 and a half feet).

Sadly, the goal isn't simply "high voltage sparks", they for whatever reason chose a Wimshurst. They do like the contraptiony feel of the spinning wheels, it's flashy and gives the visitors something to do between sparks (sparks can't be generated all the time, too distracting from other exhibits).

wrote ...
75kv versus 250kv. Voltage may scale linearly but lots of problems, capacitance, forces will scale squared. A million volt wheel pushing 100ua might be quite difficult to turn.

Interesting. I knew there were going to be non-linear issues, I just wasn't sure what they were. Could you elaborate?

Well, obviously one would be that the diameter of the wheel is 4x 1 foot, but the area of the wheels is now 16x the 1 foot model. But I'm more concerned about the electrical considerations.

In terms of force, I can gear the machine however I want. If they want sparks every 5-15 seconds, I have to figure what kind of average effort is being put into the machine, and size the capacitors appropriately. If it's too hard to turn, I'd need less capacitance.

wrote ...
A sign that says "If using this machine without the enclosed spark gap, ear defenders MUST be worn." might be helpful even if the spark is unimpressive in room air.

Well it'll be in a hall I presume, so, ear protection isn't something they could expect visitors to wear. I'll have to consider that as an issue. I've not heard how loud discharges get.

wrote ...
Build some smaller versions first to get an idea what materials work. Whimshurst machines are temperamental and can fail to work at all if the air is slightly too humid or the moon is in the wrong place...

They've already been playing with a tabletop 12" machine. And yeah, you're not the first to suggest they're twitchy buggers.

Dr. Slack wrote ...
]I've seen really big Wimshurst machines simply use a large sphere on each electrode as the capacitance to ground

Can you explain the context of "ground" in this sense?

Obviously it's not a mains powered device so there's no safety issues with ground in that respect.
Obviously the machine will have one side or the other that I'll refer to as "ground" for the sake of grounding things, enclosure if any, whatnot.

... is there another issue with the actual ground (as in, floor/dirt)? I'm peripherally aware of capacitive losses and such, but not really in a tangible "how does this effect design" kind of way.

Antonio wrote ...
Assuming that sparks between sphere terminals only occur with length up to 4 times the diameter of the spheres (see my avatar image), 18" sparks would require 4.5" spheres. The breakdown voltage between two 4.5" spheres separated by 18" can be calculated as 298 kV at sea level. A spark gap with the positive ball smaller can produce this same spark length with less voltage, 151 kV with one 2.25" sphere.

I'm a bit lost. I thought breakdown voltage of air was about 75kv/inch. How would you get 151kV to jump 18"?

I understand air plasma has about 2% the resistance of air gas, but, that requires it to be arced already.

When you say "spheres" is that hollow spheres acting as capacitors? Designs I've seen so far just use solid brass spheres much smaller.

It's also been suggested to me that sharp needles, rather than smooth spheres, would greatly reduce the voltage needed to initiate a spark because corona is already going to tend to leach in that direction.

wrote ...
ItÂ´s better to use double balls, as frequently seen in these machines. So, the required voltage is not so high. Leyden jar capacitors can be safely used, and are very easy to make from plastic containers. For greater insulation two can be connected in series at each terminal. The jars can have less than 100 pF of capacitance.

The issue with Leyden jars, that I've worried about, is maintenance and gradual degradation/catastrophic runaway once perforations start appearing. If it was for me, I'd say no problem, I know the symptoms and I'll just repair it now and then if it needs it. But this is an exhibit. It has a large budget to build it well the first time, after which I think it's expected to work reliably for a long period of time without hiring contractors to repair it.

I am not set in stone on any of this, I"m just going by my best understanding of the material so far, so by all means feel free to criticize and educate me. I'm not here to show off or brag.

Antonio

Sat May 10 2014, 11:24PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

The figure of 75 kV/inch is for breakdown between parallel surfaces. Between spheres the breakdown voltage depends on the geometry.

The spheres are the spark gap spheres. The longest capacitor-discharge spark possible between spheres is about 4 times the diameter of the spheres, and about two times this if one of the spheres is much larger than the other and the smaller is positive.
Points don't produce long sparks because they leak the charge generated by the machine at low voltage and the capacitors are not charged. You get just corona. The same if the terminal balls are too small.
Leyden jars fail completely at the first perforation and have to be replaced. But they last forever if sufficiently thick from the start.
You can use large metal balls as capacitors, as in this machine with 2.15 m disks:

Note the large terminals.
Expect frequent maintenance necessary. These machines accumulate dust fast and have to be cleaned at every few days of use. Enclosing everything in a sealed box kept clean and dry inside may help. Think carefully about a solid construction, since it's a big machine and its weight and the force necessary to turn it are not low.

AwesomeMatt

Sun May 11 2014, 12:14AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Antonio wrote ...

The figure of 75 kV/inch is for breakdown between parallel surfaces. Between spheres the breakdown voltage depends on the geometry.

I thought I was doing okay, and then I saw all those formulas and not a lot of examples or context for them. My brain fuzzes over on that, so, that'll take me a few days to work through.

Interesting though, had no idea there was a difference between parallel surfaces and spheres. That completely changes the numbers, I'm glad you mentioned it.

wrote ...
The longest capacitor-discharge spark possible between spheres is about 4 times the diameter of the spheres, and about two times this if one of the spheres is much larger than the other and the smaller is positive.

I don't understand that 4x diameter stuff. Electricity will jump between the shortest points, it's not going to hold up to infinite voltage just because the spheres aren't large enough, no?

Also, correct me if I'm wrong, but I don't think there's a way of determining which way the Wimshurst charges. Polarity will be a cointoss, so, I won't be able to cheat with a small ball sadly.

wrote ...
You can use large metal balls as capacitors, as in this machine with 2.15 m disks:

Note the large terminals.

Not sure I want to go with that route, but, good to know there's another option.

wrote ...
Expect frequent maintenance necessary. These machines accumulate dust fast and have to be cleaned at every few days of use. Enclosing everything in a sealed box kept clean and dry inside may help.

For safety's sake I'm starting to lean towards pressuring them to putting it in a clear enclosure. Keeps moisture out too.

wrote ...
Think carefully about a solid construction, since it's a big machine and its weight and the force necessary to turn it are not low.

Yeah, I'm looking at probably half inch drive shafts, standard pulleys, that kind of thing. I'm not as worried about that.

Thanks for the advice so far, it's been very helpful.

Eleccentric

Sun May 11 2014, 12:18AM

Registered Member #33460 Joined: Tue Aug 27 2013, 06:23PM
Location: Seattle
Posts: 46

AwesomeMatt wrote ...
What I keep hearing is that you can get a spark gap 1/3 the size of the wheels. So, 18" gap = 54" wheels (4 and a half feet).

That is my understanding as well, though this applies only to well-designed machines. I built four of them before I finally had a design that met my ambitious expectations. With sectored disks is manages about 1/3 the diameter, and with sectorless discs I get about 2/3 the diameter. However, forgoing sectors means that it is not self-starting, and makes it almost impossible to start in humid weather, so it wouldn't be a good idea on a museum installation. You'd need some sort of retractable charge spraying brush and the whole thing would have to be kept in a climate-controlled case.

AwesomeMatt wrote ...
In terms of force, I can gear the machine however I want. If they want sparks every 5-15 seconds, I have to figure what kind of average effort is being put into the machine, and size the capacitors appropriately. If it's too hard to turn, I'd need less capacitance.

Reducing capacitance isn't going to be too difficult with Leyden jars, depending on how you construct them. Perhaps you could put them in place without attaching the "nice" outer metal electrode first, just wrap it in cheap aluminum foil for testing and remove it in strips until you get a good value of spark rate vs difficulty of cranking. Then you can install the correct amount of beautiful copper or brass shim stock. The inner electrode doesn't have to be adjusted, and it will only add significant capacitance in areas covered by the outer electrode.

AwesomeMatt wrote ...
Well it'll be in a hall I presume, so, ear protection isn't something they could expect visitors to wear. I'll have to consider that as an issue. I've not heard how loud discharges get.

An 18" discharge isn't necessarily deafening, but the more capacitance behind that spark, the louder it will be. The goal should be to use the least amount needed to achieve your desired spark.

AwesomeMatt wrote ...
They've already been playing with a tabletop 12" machine. And yeah, you're not the first to suggest they're twitchy buggers.

As a general rule of good practice, do two things: Avoid hygroscopic materials for your insulators, and avoid small radii on all conductors. This means absolutely no sharp points or bends in your HV conductors (except your collection combs), and though this is less critical for the neutralizers, you don't want sharp points on anything anywhere near your lovely gently curving mirror-smooth conductors.

AwesomeMatt wrote ...

Dr. Slack wrote ...
]I've seen really big Wimshurst machines simply use a large sphere on each electrode as the capacitance to ground

Ground in this sense is going to be halfway in potential between the two sides. If you want to attach it to earth ground, do so to the outer foil of both Leyden jars. This will equalize the voltage on each side of the machine, reducing corona and ion wind discharges to surrounding objects. There should be very little current moving into or out of the earth ground, so it doesn't have to be low-resistance like one used for a large Tesla coil. I don't use an earth ground on my Wimshurst machines, but I do attach the outer foil of each Leyden jar to one another.

Listen to Antonio over anything I say. A lot of what I know about electrostatic machines came from studying his website. His mentioning of dust collection is very important, and perhaps is another argument in favor of an enclosed display. Also, avoid using rubber belts, or anything that will be degraded by ozone or nitric oxides.

The shape of a conductor directly influences the gradient of the electric field around it. The larger the gradient, the more likely that space will break down into a spark or corona. Sharp bends or points create sharp gradients. Theoretically, two perfect spheres of infinite size in a perfect vacuum could hold off infinite voltage. Also, the type of charge, positive or negative, affects the likelihood of charge leaping off the conductor. Hence asymmetric balls to take advantage of this asymmetry.

AwesomeMatt wrote ...
Yeah, I'm looking at probably half inch drive shafts, standard pulleys, that kind of thing. I'm not as worried about that.

Thanks for the advice so far, it's been very helpful.

Consider using chain drives or enclosed gear drives, as most belt materials will break down into a fine dust that will stick to everything, especially in the presence of electrical discharges. Ozone, UV, various nitric oxides and the numerous exotic high oxidation number creations of plasma will be produced by the machine, IIRC.

Antonio

Sun May 11 2014, 02:14AM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

AwesomeMatt wrote ...

I don't understand that 4x diameter stuff. Electricity will jump between the shortest points, it's not going to hold up to infinite voltage just because the spheres aren't large enough, no?

Also, correct me if I'm wrong, but I don't think there's a way of determining which way the Wimshurst charges. Polarity will be a cointoss, so, I won't be able to cheat with a small ball sadly.

The problem is that the spark is a capacitor discharge. The capacitors are charged slowly by the very low current that these machines produce, and the terminals must hold the charge until sufficient voltage for breakdown is reached and enough charge is stored to feed a spark. Small terminals or points leak at low voltage and do not form sparks. In systems that charge the terminals fast or with with high current, as Marx generators and Tesla coils, even points produce long sparks. But a Wimshurst machine produces just hundreds of uA at most and charges the terminals slowly.
The polarity of the machine is really random, unless some form of precharge is used. Terminals with two balls allow adjustment of the terminals so the positive terminal inclinated in the direction of the negative works as an asymmetrical gap.

1 2 3 4 next last

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