Maker Pro
Maker Pro

Variac nonsense

J

Jon Slaughter

Jan 1, 1970
0
Heres a simple explanation why the contact doesn't interpolate... no doubt
you still won't get it. It doesn't involve any math.

Voltage A
+-----+-----+-----+
| | | |
R R R R
| | | |
V---L1--+--L--+--L--+--L--+---L2---GND
V1 V2

How the hell can the voltage A a be any different no matter what L is(the
inductance of one turn) and what R is? Your telling me as you move the
contact it "interpolates"? So if we move the contact to the left it now
somehow interpolates between the voltages?

The above is identical to

Voltage A
+-----+-----+-----+
| | | |
R R R R
| | | |
V---L1--+--L--+--L--+--L--+--L--+--L---L2---GND
V1 V2

Where now the contact as moved to the left. It hasn't interpolated between
any voltages. It has simply been shifted and the voltage reduced. It would
be exactly(almost anyways) the same as if the contact had 20 points of
contact or just 1.

A is basically The average voltage across the contacts and this average
voltage shifts exactly by the same amount as if it were just one point of
contact.

YOU CAN ONLY CHANGE IN DISCRETE STEPS!!!!. But when you move the contact
you do not add any current paths!

But it doesn't matter. All the resistors have the high side at the same
fucking potential. As you move the contact one turn to the left or right you
have increased or decreased that by a constant discrete amount. Each time
you do it you move it in a discrete step. There is no damn interpolating.


Yet I guess that is even too hard for you to understand?

How can the contact be at different potentials?

Now for all the fucking genius that believe the contact interpolates heres
the fucking reason for a contact - It makes contact wtih more than 1 turn so
that the contact will always have a voltage on it. It can't get stuck
between two turns and not output a voltage. It also smooths voltage as you
turn. (that is not interpolating)

e.g., As you turn you might add a turn and loose a turn but because the
turns are not spaced perfectly you might add a turn and have the voltage
jump up before the other turn is added. (basically changing the contact size
as it connects to N then N+1 then back to N causing the voltage to change)

In any case you all are morons... ever single person that replied and agreed
with larkin was wrong. Thats what you get for jumping on the bandwaggon.
Maybe learn to think for yourselfs next time... assholes.


The reason for the resistance is obvious... It's so A won't be at different
potentials. This is basic electronics...
Maybe these expert engineers are not so expert?


Of course I expect the usual responses...
 
J

Jon Slaughter

Jan 1, 1970
0
Just so it's clear the reason why it isn't interpolation is because you move
in discrete steps. If V is the voltage at some position then V + dV is the
voltage when you turn one step... you can't get anything lower than dV.
There is no interpolation... has nothing to do with the contact or anything.
How do you expect to get dV/2?

Note V is linear w.r.t to the turns so dV is constant. This means as you
turn the rotor you increase V by discrete steps. Simple as that. Now how's
the moron? Should name names?
 
F

Fred_Bartoli

Jan 1, 1970
0
Jon Slaughter said:
Just so it's clear the reason why it isn't interpolation is because you
move in discrete steps. If V is the voltage at some position then V + dV
is the voltage when you turn one step... you can't get anything lower than
dV. There is no interpolation... has nothing to do with the contact or
anything. How do you expect to get dV/2?

Note V is linear w.r.t to the turns so dV is constant. This means as you
turn the rotor you increase V by discrete steps. Simple as that. Now how's
the moron? Should name names?

The one small thing your narrow mind don't get is that the Rs change,
depending on the driving points position.

Should you have checked, since you have one, you'd have noticed continuity
and (maybe) thought before spouting all this non sense.
 
S

Sylvia Else

Jan 1, 1970
0
Jon said:
Heres a simple explanation why the contact doesn't interpolate... no doubt
you still won't get it. It doesn't involve any math.

Voltage A
+-----+-----+-----+
| | | |
R R R R
| | | |
V---L1--+--L--+--L--+--L--+---L2---GND
V1 V2

How the hell can the voltage A a be any different no matter what L is(the
inductance of one turn) and what R is? Your telling me as you move the
contact it "interpolates"? So if we move the contact to the left it now
somehow interpolates between the voltages?

Your model contains non-interpolation as an assumption, because of the
way you've placed one resistor per turn.

Model the brush with twice the number of resistors per turn, and allow
it to be moved a half turn at a time, and in three consecutive positions
you get

Voltage A
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+
| | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and

Voltage B
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+ +--+--+ +--+--+ +--+
| | | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and
Voltage C
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+
| | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

It should be reasonably clear that Voltage B will lie midway between
Voltage A and Voltage C, so the number of discrete voltages has been
doubled (ignoring end effects).

Model with three resistors per turn, and the number of discrete voltages
is tripled.

The closer the model gets to being representative of a continuous
resistive contact, the closer you get to having a continuously variable
voltage.

Sylvia.
 
J

Jon Slaughter

Jan 1, 1970
0
Sylvia Else said:
Your model contains non-interpolation as an assumption, because of the way
you've placed one resistor per turn.

Model the brush with twice the number of resistors per turn, and allow it
to be moved a half turn at a time, and in three consecutive positions you
get

Voltage A
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+
| | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and

Voltage B
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+ +--+--+ +--+--+ +--+
| | | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and
Voltage C
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+
| | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

It should be reasonably clear that Voltage B will lie midway between
Voltage A and Voltage C, so the number of discrete voltages has been
doubled (ignoring end effects).

Model with three resistors per turn, and the number of discrete voltages
is tripled.

The closer the model gets to being representative of a continuous
resistive contact, the closer you get to having a continuously variable
voltage.

Funny, so your contact changes size as you turn the rotor?
 
S

Sylvia Else

Jan 1, 1970
0
Jon said:
Funny, so your contact changes size as you turn the rotor?


Hilariously, that's not what the model implies.

If it makes it clearer, think of the horizontal lines beneath the
resistors as representing the surface of the wires, thus


Voltage A
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+ +--+--+
| | | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and

Voltage B
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+ +--+--+
| | | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND
V1 V2 V3 V4

and
Voltage C
+-----+-----+-----+-----+-----+
| | | | | |
R R R R R R
| | | | | |
+--+--+ +--+--+ +--+--+ +--+--+
| | | |
V------L1--+-----L-----+-----L-----+-----L-----+---L2---GND


Sylvia.
 
E

ehsjr

Jan 1, 1970
0
Jon said:
Funny, so your contact changes size as you turn the rotor?

Maybe this will help you see what Sylvia is saying:

The contact area between the wiper and the winding changes.

| Wiper } |
----------------------------}--------
__________ ___________ ______}
Winding N Winding N+1 Winding N+3

Move it a little:

| Wiper } |
----------------------------}--------
__________ ___________ ______}
Winding N Winding N+1 Winding N+3

Move it a little more:


| Wiper } |
----------------------------}--------
__________ ___________ ______}
Winding N Winding N+1 Winding N+3


Ed
 
A

Ancient_Hacker

Jan 1, 1970
0
Way back in ancient times, philosophers used to gather at the local
Greek pub and vehemently argue about important questions of the day,
such as how many teeth a horse had.

Euripedes would argue that it was self-evident-- just as there were
6 main deities, each with four limbs, a horse must have 24 teeth. No
argument possible there. Everyone else must be a moron. Right
after that, Bellerophon belched loudly and went into a rant as to how
everyone other than him was a moron, because it was obvious that if a
lizard has 72 teeth, and a chicken has none, you can with straightedge
and chalk construct a 3/4/5 triangle whose perpendicular quadriceptor
gives you a projection of length 28.

Silly them, if they'd just have looked at the butt of the slave boy
they were passing around, he'd been bitten by a horse and you could
count the teeth marks. But nobody did.


------------------
If you don't get the drift, if you're going to use a model, it has to
be a realistic model. Or you could look at the real thing.

Your basic carbon contact is not a bunch of R's standing on end, it's
a bunch of vertical and horizontal R's. If you try that model, you'll
see the vertical resistance near the edges is higher, as there are
half as many horizontal R's paralleling the vertical ones. So there
is going to be a continuous sweep of voltages averaged by the contact.
 
A

Archimedes' Lever

Jan 1, 1970
0
Way back in ancient times, philosophers used to gather at the local
Greek pub and vehemently argue about important questions of the day,
such as how many teeth a horse had.

Euripedes would argue that it was self-evident-- just as there were
6 main deities, each with four limbs, a horse must have 24 teeth. No
argument possible there. Everyone else must be a moron. Right
after that, Bellerophon belched loudly and went into a rant as to how
everyone other than him was a moron, because it was obvious that if a
lizard has 72 teeth, and a chicken has none, you can with straightedge
and chalk construct a 3/4/5 triangle whose perpendicular quadriceptor
gives you a projection of length 28.

Silly them, if they'd just have looked at the butt of the slave boy
they were passing around, he'd been bitten by a horse and you could
count the teeth marks. But nobody did.


------------------
If you don't get the drift, if you're going to use a model, it has to
be a realistic model. Or you could look at the real thing.

Your basic carbon contact is not a bunch of R's standing on end, it's
a bunch of vertical and horizontal R's. If you try that model, you'll
see the vertical resistance near the edges is higher, as there are
half as many horizontal R's paralleling the vertical ones. So there
is going to be a continuous sweep of voltages averaged by the contact.


Edges make no difference. Your model should be about the line of
tangency that CONTACT is defined by. THAT mating surface, and the
electron flow interface into or out of it is all that matters. The area
near the corner or face of the brush rarely even gets utilized, much less
plays into the physics behind the operation of a carbon based high
amperage brush against a copper face. The mating face is rarely a 'full
face' contact, so think about a railroad wheel and the rail. All that
weight rests on a strip of steel as narrow as a dime. A quarter inch
wide brush likely has less than an eighth inch of strip used as a
calculation for its capacity by its designers.
 
A

Archimedes' Lever

Jan 1, 1970
0
More important, the tops of the wires are often machined flat, so they
don't act like point contacts.

Nope. More often, they are flattened without removing any of their
cross sectional area, then lightly surface finished to yield a coplanar
face. They *still* do act like a point contact as full face mating is
never possible due to they way the brush "wears". One gets about half
the "thickness" of the brush as a 'mating line'.
You could actually look inside a Variac and see this,

Or, you could look, and thinking that you know all, guess at how it is
made, and even assume that all are made 'your way', as you have done
here.
instead of
theorizing.

Yes, John... exactly that, fuckhead.
 
E

ehsjr

Jan 1, 1970
0
John said:
Yup, that's how it works. The windings aren't little cylinders, they
are machined flat. The ratio of exposed contact width to gap is
roughly what you've drawn... more copper than gap.

The brush contacts, at most, the full width of three wires; at least,
two.

The brushes on mine are also rounded on the ends, so they sort of
creep up on a contact...




instead of making a hard on/off sort of connection. I suspect they
naturally wear in this shape, rounded on the corners.

I have read that the brush is made of anisotropically conductive
graphite, with the better conduction along the vertical axis. This
reduces the direct-turns-shorting resistance.

Neat - I wonder how they do that.

I like your diagram - much better than mine. Also I noticed
that I made an error - should be winding N+2, not N+3.
Still, the concept remains the same.

I've observed no change in primary idle curent as the wiper is moved,
so the turns-shorting transition is very smooth.

Pretty clever, overall.

Yes! :)

Ed
 
G

Glen Walpert

Jan 1, 1970
0
Neat - I wonder how they do that.

Graphite is an inherently anisotropic material having far different
properties parallel to the strongly bonded hexagonal planes than
perpindicular to them in the weakly bonded inter-plane direction.
Electrical resistivity within the hexagonal planes is about 2.5 x
10**-6 to 5.0 x 10**-6 ohm*meter, and about 3000 X 10**-6 ohm*meter in
the perpindicular direction.

Real polychrystaline graphite used as brushes is only partially
ordered (else it would shear too easily between the planes) so the
ratio of conductivities is probably less than 10:1. (Thermal
conductivity varies similarly, with about a 300:1 ratio in single
crystal graphite and much less in practical semi-ordered
polycrystaline graphite.) The best ordered readily available graphite
is graffoil (tm), but it is much too weak to use as electrical
brushes. Graphite flakes pressed into sheets tend to order themselves
with the strong planes roughly in the plane of the sheet if the flakes
are large enough relative to sheet thickness, and the sheets can be
stacked and pressed at high pressure and temp to form semi-ordered
blocks suitable for brushes.

Ref. "Handbook of Carbon, Graphite, Diamond and Fullerenes" by Hugh O.
Pierson
 
J

JosephKK

Jan 1, 1970
0
Yup, that's how it works. The windings aren't little cylinders, they
are machined flat. The ratio of exposed contact width to gap is
roughly what you've drawn... more copper than gap.

The brush contacts, at most, the full width of three wires; at least,
two.

The brushes on mine are also rounded on the ends, so they sort of
creep up on a contact...


instead of making a hard on/off sort of connection. I suspect they
naturally wear in this shape, rounded on the corners.

I have read that the brush is made of anisotropically conductive
graphite, with the better conduction along the vertical axis. This
reduces the direct-turns-shorting resistance.

I've observed no change in primary idle curent as the wiper is moved,
so the turns-shorting transition is very smooth.

Pretty clever, overall.

John

Since you have a unit to hand, is the contact area of the wiper a bit
skew to the contact area of the windings?


slaughter could save himself a lot of embarrassment just by using a
search engine and reading a few pages.
 
slaughter could save himself a lot of embarrassment just by using a
search engine and reading a few pages.

Yeah but then what would I read on a rainy afternoon?
Mit der Dummheit kaempfen Goetter selbst vergebens.
--Schiller (from "Die Jungfrau von Orleans")

And they hadn't even met Slaughter yet.
 
P

Paul E. Schoen

Jan 1, 1970
0
Glen Walpert said:
Graphite is an inherently anisotropic material having far different
properties parallel to the strongly bonded hexagonal planes than
perpindicular to them in the weakly bonded inter-plane direction.
Electrical resistivity within the hexagonal planes is about 2.5 x
10**-6 to 5.0 x 10**-6 ohm*meter, and about 3000 X 10**-6 ohm*meter in
the perpindicular direction.

Real polychrystaline graphite used as brushes is only partially
ordered (else it would shear too easily between the planes) so the
ratio of conductivities is probably less than 10:1. (Thermal
conductivity varies similarly, with about a 300:1 ratio in single
crystal graphite and much less in practical semi-ordered
polycrystaline graphite.) The best ordered readily available graphite
is graffoil (tm), but it is much too weak to use as electrical
brushes. Graphite flakes pressed into sheets tend to order themselves
with the strong planes roughly in the plane of the sheet if the flakes
are large enough relative to sheet thickness, and the sheets can be
stacked and pressed at high pressure and temp to form semi-ordered
blocks suitable for brushes.

Ref. "Handbook of Carbon, Graphite, Diamond and Fullerenes" by Hugh O.
Pierson

I can verify that the brushes are made from grain-oriented graphite, at
least in the units made by Staco. BTW, Variac is a trademark of General
Radio:
http://www.ietlabs.com/Variac/Variac.html

Superior Electric uses the trademark Powerstat, and I think GE had a
"Volt-Pac". Staco's units are properly called variable autotransformers.
There are models that have a two-layer insulated winding which provides
isolation, at the cost of larger size for the same kVA.
http://www.superiorelectric.com/
http://www.stacoenergy.com/variable_transformers.htm
http://archive.jenkins.com/jenkins/transformers/variable_trans.htm

We had problems with one of the Staco models, a 1520, which is nominally
rated 9.5 amps at 240 VAC. We use them at currents up to 10x their rating
for short bursts, and although they are rated for such use, we had problems
with the brushes and windings failing catastrophically. Upon testing, we
saw a lot of arcing at the higher levels, and also noticed that the brush
got very hot at moderate levels of overload, where we might need to draw
power for up to 3 minutes. We noticed that this model had a very thin, flat
brush, rather than the wider tapered version illustrated above. It seems
that the thin brushes overheated and cracked to the point that the brush
holder landed on the windings and shorted them out, causing a major
overload and burned windings. We reduced the problem by substituting a
tapered brush assembly, and also adding a stronger spring to maintain more
pressure on the windings.

There is or was a company in Italy that makes a variable autotransformer
with a carbon roller for a brush, but I have never used them. I would think
it might cause more problems than the usual design.

There is also the unique Peschel variable transformer made by Hipotronics.
It uses a sliding copper brush and diodes to eliminate shorted turns and
provide higher power capability. They have units up to 1.6 MW! I worked
with an engineer from Hipotronics and he showed me some of the units. But
I'm not convinced they are all that much better than the toroidal design.
http://www.hipotronics.com/pdf/PVT.pdf

What I really want to do is design a solid state "variac" using PWM
technology.

Paul
 
P

Paul E. Schoen

Jan 1, 1970
0
Yeah but then what would I read on a rainy afternoon?


And they hadn't even met Slaughter yet.

What he has been saying is funny, as fits a name that can be rearranged to
make Jons Laughter...

Paul
 
A

Archimedes' Lever

Jan 1, 1970
0
Doesn't look like it. But that would be another way to smooth the
interpolation. Or use a round or diamond-shaped contact. Or use two
offset contacts, as one pic in abse hints at.

John

Whatever happened to "if you short a turn, it will fry..." declaration,
John? Make up your mind, boy.
 
A

Archimedes' Lever

Jan 1, 1970
0
Yup, that's how it works. The windings aren't little cylinders, they
are machined flat.

Wrong. Machining would reduce the ampacity of each turn.

The wire faces are pressed flat and minimal machining follows.

That way, the cross sectional area of the wire is not diminished
appreciably, which with machining, it would be.
 
Top