Question about vacuum chuck forces:

Originally posted by twhitehead
And that is where you are going wrong. There is no such thing as a vacuum force. The vacuum does not suck.

The flat object has three different forces acting on it:
1. The atmosphere on one side - proportional to the area.
2. The solid support which gives a static non-moving force that keeps the flat object stationary but doesn't take part in our calc ...[text shortened]... exposed to the vacuum.

So, total force required to move the object is force 1. minus force 3.

I'm not sure why you are disagreeing with him here.

Look at a force balance on the plate. Call Normal to side 2 ( acting as though a force would be pushing on side 2) positive.

->+ ΣF = 0

-P_atm*A1 + (P_atm - P_pump)*A2 + P_atm*(A1 - A2) + N = 0

-P_atm*A1 + P_atm*A2 - P_pump*A2 + P_atm*A1 - P_atm*A2 = -N

(-P_atm*A1+P_atm*A1) + (P_atm*A2 - P_atm*A2) - P_pump*A2 = -N

(0) + (0) - P_pump*A2 = N

N = P_pump*A2

Originally posted by sonhouse
That is semantics. Of course the resultant is A minus B.

My post has no A or B in it, so I don't know if you are agreeing with me or not.

It is just convenient to refer to it as a force.
Which in this case is a mistake.

The force of the atmosphere on top V the lower pressure on bottom results in a downward force on the substrate.
And the force of the atmosphere applies everywhere on top of the plate, regardless of whether or not it has holes to a vacuum behind it.

I understand that. But it is ONLY the area exposed to the lower pressure that causes the downward pressure not the total surface area of the substrate.
Sorry, but that is nonsense. How does the atmosphere know to only press where there is vacuum on the other side?

Originally posted by joe shmo
I'm not sure why you are disagreeing with him here.

Look at a force balance on the plate. Call Normal to side 2 ( acting as though a force would be pushing on side 2) positive.

Can you simply explain it without lots of figures that I have to decipher? Sonhouse keeps doing the same, trying to give hundreds of figures while ignoring the model.

Assuming the force from the near vacuum side is negligible, the force on the plate is from atmospheric pressure and is proportional to the size of the plate. Sonhouse is essentially ignoring that and instead treating the vacuum as a suction force.

Originally posted by joe shmo
As a consequence they necessarily intersect ( the pump is pulling and the system resisting). The point of intersection is the operating point (pressure and flow) of the system.

My understanding of the system was that there was no flow at all and the pump is irrelevant. Maybe I have misunderstood the scenario. Perhaps you could explain it better than sonhouse has?

Originally posted by twhitehead
Can you simply explain it without lots of figures that I have to decipher? Sonhouse keeps doing the same, trying to give hundreds of figures while ignoring the model.

Assuming the force from the near vacuum side is negligible, the force on the plate is from atmospheric pressure and is proportional to the size of the plate. Sonhouse is essentially ignoring that and instead treating the vacuum as a suction force.

"Can you simply explain it without lots of figures that I have to decipher? Sonhouse keeps doing the same, trying to give hundreds of figures while ignoring the model."

The figures are the explanation, thats the language of physics. But ill break it down as best I know how.

I'm going to add the pulling force F_p into the analysis, maybe that is where you are hung up.

Side 1:
On the Blanked side we consider "2" forces acting on the object.

Force 1:
The resultant of our constant pressure gradient integrated over the entire side1: P_atm*A1 ( it acts to push on side 1)

Force 2:
Our pulling Force on side 1 (F_p)

Side 2: ( the side with a portion of negitive gauge pressure acting on a discrete area)

There are "3" Forces acting normal to this surface ( acting in the same direction as the pulling force ( F_p)

Force 1:
The resultant of net constant pressure gradient integrated over an Area (A2). Its magnitude is: (P_atm - P_pump)* A2

Force2:
The resultant of atmospheric pressure acting over the remainimg area of side 2. Its magnitude is: P_atm(A1-A2)

Force 3:
The Normal contact force ( N )

When you add all of these forces up ( this is what I did above) and set the normal force ( N ) to Zero (the plate is just leaving contact, but the body is remaining in static equilibrium ΣF = 0) the magnitude of the force F_p is:

F_p = P_pump*A2

Originally posted by twhitehead
My understanding of the system was that there was no flow at all and the pump is irrelevant. Maybe I have misunderstood the scenario. Perhaps you could explain it better than sonhouse has?

I added that information in regards to leakage and sealing that you guys were discussing. It is the mechanical reality of the pump/system relationship. In reality things are less than ideal becuase there is a "real system" coupled to a "real machine". They each have behavioral characteristics to consider for the model. I was also bringing that up because Sonhouse was using Ideal ( probably inaccurate) figures (perfect vacuum) in his computations. I was trying to point out that the amount of vacuum a pump can provide is specific to the pump and the system it is coupled to. He would have to find the performace curves of the pump, and model the system to be able to anticipate the closer to "actual" force (via pressure) provided by the chuck (or he could just put a pressure guage in the chamber).

Originally posted by joe shmo
I added that information in regards to leakage and sealing that you guys were discussing. It is the mechanical reality of the pump/system relationship. In reality things are less than ideal becuase there is a "real system" coupled to a "real machine". They each have behavioral characteristics to consider for the model. I was also bringing that up becaus ...[text shortened]... rce (via pressure) provided by the chuck (or he could just put a pressure guage in the chamber).

I specifically used 50% vacuum, not perfect vacuum because that is what I read on other vacuum chucks, where perfect in terms of physical pressures and such is 28 inches of mercury and so 50% is 14 inches of mercury or about 7 PSI which I converted to metric. That's all I did, convert to metric, I just wanted to show my work.

If my numbers are wrong please show me. It's just that the numbers I generated roughly agree with what we feel when we try to separate the substrate from the chuck.

I am bringing this up because I designed a chuck with an o ring instead of channels so the force would not depend on surface defects like scratches and such that would reduce the force due to leakage.

I did find the surfaces to be very smooth, one a very smooth alumina substrate about 210 by 120 mm and I took pains to make the chuck surface smooth and scratch free by using a fine diamond grit polishing tool.

I can see how good that is by the amount of photo resist fluid getting sucked under the substrate. At first go, before polishing, it was terrible, photo resist getting into the controlling vacuum solenoid which jammed up the little magnetic plunger that opens to vacuum and closes. That after a filter that should capture most of the photo resist getting into the vacuum tubing.

After polishing the leakage went way down, it was obvious when we turned off the vacuum and pulled the substrate loose for the next step in processing, the amount of photo resist under the substrate was at least 90% less. It got gradually worse after a number of substrates were processed presumably because the alumina is much harder than the aluminum chuck and would inevitably scratch that surface and thus reduce the effective holding force and contribute to solenoid contamination, which requires disassembly and cleaning with acetone to put it back in operation.

My O ring will stop that completely or at least let us run ten times longer without solenoid contamination.

At least putting out these questions illustrated where I was going wrong calculating the downwards clamping force.

At first I didn't understand why I was only seeing maybe a kg of upwards force on the substrate to cause separation, and it needs a good deal of force when the substrate is spinning at 3000 or more RPM. When the vacuum gives way due to leakage or whatever, bad pump, the substrate flies off the chuck with great force, enough to penetrate the metal outer shell which looks like it has been hit twenty times with shrapnel!

So now I understand the surface area of the channels times the percentage of vacuum compared to atmosphere is what gives the downwards holding force. The channels are only a mm wide and maybe 100 mm long so that would be about 100 mm square to calculate the downwards force.

When I install the new chuck with the o ring the area will be much greater since it will be around 115 X 95 mm or so, giving a force of more like 100 kg holding to the chuck which will mean the substrate will never fly off again even if the vacuum goes to 10% or so instead of my estimate of 50%.

The o ring will also ensure the two surfaces will be separated by a small distance, maybe 100 microns or so which will mean small scratches will not effect the overall holding force.

My only concern there is if there is too much separation the downwards force will break the substrate which as I said is alumina but only 1 mm thick and of course extremely brittle.

The question I have to answer is if the separation forced by the distance inherent in an o ring is sufficient to cause a bending force sufficient to break the alumina substrate.

I have a back up plan for that if needed though.

Originally posted by joe shmo
(or he could just put a pressure guage in the chamber).

He almost certainly has, so the whole talk about a pump is irrelevant as far as I can see and unnecessarily confusing.

Originally posted by joe shmo
Force2:
The resultant of atmospheric pressure acting over the remainimg area of side 2. Its magnitude is: P_atm(A1-A2)

That is where you are going wrong:

You are assuming that there is atmosphere on the plate where it is in contact with a solid surface. Clearly that cannot be the case.

Sonhouse is making essentially the same error.

Originally posted by twhitehead
That is where you are going wrong:

You are assuming that there is atmosphere on the plate where it is in contact with a solid surface. Clearly that cannot be the case.

Sonhouse is making essentially the same error.

What do you mean there is NO atmosphere on top of the plate, the alumina substrate I am talking about?

When you have atmospheric pressure on a plate and underneath is a channel leading to a partial vacuum, there IS a downward force on the plate, in my case the alumina substrate.

What is it you think we have wrong? There is an obvious force pulling the plate down to the chuck so what are you talking about? If it were say a window on the space station and my substrate was on the surface of the window and we could suddenly get rid of the window but still leave enough to hold the 'plate', you think there would be no force holding the plate to the window?

That WOULD be the atmosphere pressure inside the ISS pushing on the plate with whatever PSI is in there. With essentially zero vacuum on the outside (something like 1E-9 torr or better) then the force would be the surface area times the pressure.

You have a problem with that?

Originally posted by twhitehead
That is where you are going wrong:

You are assuming that there is atmosphere on the plate where it is in contact with a solid surface. Clearly that cannot be the case.

Sonhouse is making essentially the same error.

I see what you are saying, but you are choosing to add some parts of a model at a certain level of complexity and ignore others, in such a way as to fit your theory. When ALL assumptions are combined they should confirm Sonhouse's observations. (if you want I'll revise the model with added complexity to show you this), but I think this can be resolved by discussion.

Lets start by placing a "very thin - perfect" seal perfectly around the perimeter of the orifice of A. The magnitude of the force needed to pull the plate is:

F_p = P_pump*A_orifice

This is the model that MATCHES OBSERVATIONS ( it is the one I described earlier ). You may cry foul and say there is no such "very thin - perfect" seal in the setup, just two metal plates coming in contact with each other over their entire surface ( less the area of the orifice) "perfectly" under your assumption.

It is the case "in theory"- as you point out, that if they are "perfectly sealed" over the entire area of contact that the pulling force necessary grows linearly with A1.

(P_atm - P_pump)*A1 - P_atm*A1 + N + F_p = 0

As N goes to Zero:

F_p = P_pump*A1 ( so the in this case you are correct, the larger the plate the larger the force )

However, reality suggests that the seal is NOT perfect over the entire area of contact. So since you bring it up, exactly how NOT perfect is it; you tell me? And since the seal is not perfect there is necessarily leakage. Now we have to dive into the case of pump performance, system modeling, etc...

So what is going on? It is of my opinion that the seal over the entire area of these two mating plates is actually very imperfect. Again, Observations confirm this.

So the next logical thing to do is ask yourself why reality is much more like situation 1 as opposed to the situation you propose:

I'll give my opinion on this. The assumption of an absolutely rigid body is the one that is most likely when removed supports reality. The body deflects locally around the orifice effectively making the "very thin - perfect" seal around the perimeter of the orifice. which leads to:

F_p = P_pump*A_orifice << P_pump*A1

What do you think?

Originally posted by sonhouse
What do you mean there is NO atmosphere on top of the plate, the alumina substrate I am talking about?

No, that is not what I said.
I said there is NO atmosphere on the other side of the plate where it is in contact with whatever it is in contact with.

When you have atmospheric pressure on a plate and underneath is a channel leading to a partial vacuum, there IS a downward force on the plate, in my case the alumina substrate.
And that downward pressure is still there even when there is solid surface under the plate rather than a vacuum. So the downward force is proportional to the surface area of the top of the plate and not the channels leading to a vacuum.

What is it you think we have wrong?
I have said it over and over. You are ignoring the area where the plate is in contact with a solid support. Your calculations should treat that as equivalent to a perfect vacuum.

Originally posted by joe shmo
I see what you are saying, but you are choosing to add some parts of a model at a certain level of complexity and ignore others, in such a way as to fit your theory.

No, I am choosing to include only parts of the model relevant to the calculations sonhouse is doing to show where he is going wrong.

When ALL assumptions are combined they should confirm Sonhouse's observations.
They wont because in reality air slips in from the sides which is why it is relatively easy to lift the plate, but we have no formula for this and almost certainly wont be able to get one.

What do you think?
I think that assuming leakage you are certainly closer to the actual situation, but I think you are going to far in assuming that sonhouse has accurately measure everything and that his figures match your calculation. Also sonhouse is not assuming leakage so you disagree with him there.
In addition he is talking about improving the seal which he claims will make no difference, but your description should have it making a significant difference.

Originally posted by twhitehead
No, I am choosing to include only parts of the model relevant to the calculations sonhouse is doing to show where he is going wrong.

[b]When ALL assumptions are combined they should confirm Sonhouse's observations.
They wont because in reality air slips in from the sides which is why it is relatively easy to lift the plate, but we have no formul ...[text shortened]... ms will make no difference, but your description should have it making a significant difference.[/b]

The substrate (plate) is 1mm thick, aluminum. (a ductile metal). The simplest assumption to remove is that of rigid body, because it is a BAD assumption. It is not at all a rigid body ( its barely thicker than aluminum foil ). it is without a doubt deflecting locally around the orifice "effectively" creating a "very thin-perfect" seal around the perimeter of the orifice.

F_p = P_pump*A_orifice is the obvious model from the reality of the situation.

if the substrate was thicker plate steel this may not be the case...As it stands, if the information sonhouse provided is truthful it aligns with the model described above.

Originally posted by joe shmo
it is without a doubt deflecting locally around the orifice "effectively" creating a "very thin-perfect" seal around the perimeter of the orifice.

Except for the fact that the 'orifice' consists of multiple slot whose exact arrangement we don't know.
But otherwise I generally agree with your analysis.
What I don't quite understand is where sonhouse plans to put a seal and how this will change the situation. If I am correct and he places a ring around the edge of the flat objects so that it now has a much better seal, then he is mistaken in thinking it will have no effect on the difficulty in removing it.