Need info about freeze clamping

[maartenlecleir]

Does anyone can provide me with information about freeze clamping?

Thx, Maarten

[mxtras]

Are you referring to shrink fitting???

Not exactly sure what you are asking here....sorry.

Scott

[home-tek]

no, he is referring to holding the part by freezing it with water and very low temperature (ice) while performing the machining.

We have tried this very long time ago and the result was not very good, may be the problem was the machining process we tried to do...

[maartenlecleir]

Hi Scott,
What I need is a way to keep very small objects in place on a CNC machine. Vacuum clamping wasn't a choice because the schapes are too irregular, but I noticed that freeze clamping was a good option. I know a chuck is cooled down to 3 °C, then some water is sprayed onto the clamping surface and the part is positioned. Then it's further cooled down to -10°C so it stays in place while manufacturing. But I would like to know more about this proces. I even think it's possible to do this with sending compressed air through a venture to get the cooling effect, but how exactly I'm still figuring out...so more information would be welcome.

[mxtras]

Sorry to ask - it wasn't clear and I thought it could be a language thing as it is such an uncommon approach. I have only seen one attempt and it was not all that impressive. I am sure it has a place but not in my shop.

Please keep us informed if you try it - always interested to hear about this kind of thing!!

Scott

[maartenlecleir]

I think this method should work with micro parts, but if someone has already bad experiences with freeze clamping micro parts, let me know.

[BillPSu]

You could use a low melting point metal such as MCP 70. Melting point is 70-75 degrees C. the following is a link to their site.

http://www.mcp-group.com/alloys/lmpa_workh.html

[Mcgyver]

I saw this done years ago in the jewelry manufacturing biz to finish those flattened gold chains. The chain was basically frozen to a 12-18" cylinder as a helix and a skim cut was taken...yup machining gold in a lathe. This was the opposite of trying to hold a micro part - the overall surface area was huge and a very light cut was being taken. maybe you can network to someone in that industry

[psychomill]

I've done this for thin wall brass parts and some other small aluminum stuff. Have used this method off and on for 9 or 10 yrs. We call it 'Ice Chucking" around here. A few key things: Freeze the subplate or fixture first. I've also used some refrigerant products to freeze just a "top plate" to minimize distortion. However, some of this stuff is illegal now ( and not sold in the US anymore to the general public ), or very hard to come by. Either way, OSHA would have a fit. Then pour a layer of water down, freeze the entire thing and machine away. Gotta be fairly quick on the cycle or you'll have to "refreeze". Hence the use of refrigerants. Dry machine, use no coolant.

Have been trying some other media (besides water) to freeze in order to get more clamping life. Haven't come across anything significant though. I don't do this on a daily basis (ice chucking that is). And I will seek for alternate methods first before going that route.

[Geof]

Your comment about putting compressed air through a venturi is almost correct; actually it is a joule thompson nozzle. If you Google on 'joule thompson effect' you will find many references to obtaining low temperatures by expanding a compressed gas. You might want to consider using compressed (actually liquified at around 1800 psi) carbon dioxide and allowing it to expand through a small orifice and then pass through a labyrinth machined in an aluminum plate. You will get temperatures well below 0 Celcius.

[maartenlecleir]

Last weeks I've been quite busy with finding solutions for clamping those micro parts and it's almost certain that ice chucking is the most desirable solution for this kind of application. Knowing this, it was just a question how to make a cold plate.

There are, in my eyes, different possibilities to develope this cold plate, like using technology based on refrigerants, vortex- or venturi-effects.
Looking at sites who use extreme cooling for cpu's gave me the idea to use another technologie based on THE PELTIER-effect, or TEC's (thermo electric cooler's). An application based on a peltier can be found on the next link, and it looks like the ice chuck I need to make myself. http://www.melcor.com/mpa.html
Principle:
DC current flows through the TEC causing heat to be tranferred from one side of the TEC to the other, creating a cold and hot side. The sides can be inverted when you change the direction of the DC current, so it's easy to change from "cold-mode" to "heat-mode", and temperatures are easy adjustable by changing the amount of current through the peltier cell.
Other advantages of this technologie are that it takes little space to implement a peltier-cell (I've got room for a plate of 100mm x 100mm) and it's cheap technologie. This almost sounds like a commercial but is just a big introduction to my next problem.

I find it rather difficult to chose the right peltier cell. The danger excist in chosing the wrong cell, which causes to heat your chuck and not cooling it, because the hot sides really becomes hot
For example, I could use a high powered peltier cell so cooling will happen quite fast, but if you can't get lost of all the watts the peltier cell creates himself, the hot side becomes to hot and you'll end up heating everything and your cell will be ready for the dump.

So I'm searching someone who has experience with peltier technologie.

The goal is to reach a temperature of maximum -15°C at the cold side of the peltier, so a piece of aluminium with dimensions of 100mm x 100mm x 10mm shall cool down (between -10°C and -15°C) in a maximum of 10 minutes.
The maximum ambient temperature Ta will be a 30°C and a heat sink with a ventilator is prefered. (I know I ask much, sorry for that

Perhaps knowing this, someone with the experience of peltier could tell me if it will be possible to cool down with a single stage unit, or multistage has to be applied because the temperature difference is to big, or the time is to short. Or if it will be possible at all to use a peltier...

I already said that a heat sink with a ventilator is prefered, watercooling of the heatsink is perhaps a solution if there is realy no where else to go. I would rather wait another extra 10 minutes, then use watercooling.
Cooling down the heatsink by injecting the coolants who are already available in the CNC machine (95% water) to improve the efficienty of the heat sink sounds like a good idea to me.

Here are some calculations I tried myself, but I changed the order in which they usually happen, tell me if I'm wrong somewhere:

Ta = 30°C
Tc = -15°C
Th = Ta + (P+Qc) (°C/W)

P: the power of the module
°C/W: the efficiënty of the heatsink, or thermal resistance
Ta: the ambient temperature
Th: the hot side temperature of the module
Tc: the cold side temperature of the module
Qc: the heat input to (or heat pumped by) the module in watt

the most difficult part here is to estimate Qc. There is no active load, only a passive load. I used a program from some site to estimate which load there has to be transported. It gave me about 20 watts, but I would like to see some calculations about this Qc.

Knowing most the elements have a dTmax of 65°C I estimated Th at max 50°C
deltaT = Th - Tc
65°C = Th - (-15°C) => Th = 50°C

Th = Ta + (P+Qc) (°C/W) => 50°C = 30°C + (P+20W) (°C/W)

I think it should be possible to get an heat sink of 0,20°C/W, so
20°C = (P+20W) 0,20°C/W => 100W = P+20W => P=80W


Following calculation I made to estimate the time it takes to cool down the piece of aluminium.
The amount of energie needed to transport, to assure a piece of aluminium will cool down from example 30°C to -15°C is according to me:
Q = m.c.dT where
c = 960 J/kgK
dT = 45K
m= 2710 kg/m³ . 0,0001m³ (100mm x 100mm x 10mm)
= 0,271 kg
=> Q=0,271kg . 960 J/kgK. 45 K = 11,7 kJ

so if my peltier with power 80 W has a Qmax of 50W, I would take 11700 J/ 50 W= 234 s to cool down the piece of aluminium.

Am I right now to think that I can order a peltier element with deltaT = 65°C and a power that shouldn't necessarrily exceed the 80 W,
for example the TEC1-12705T125 with
Umax = 15,2
Imax = 5 (=>P max=76W)
dT=67°C
Qmax = 52,1W

Hopefully my assumations where right and I didn't make to much faults, because I'm new peltier. A lot of information I found on sites for computermodding, but they always calculate with their known active load of the cpu which I have to estimate.
I have to thank all that have had the courage to read this post and those who will reply with any kind of information.

[Idle hands]

Interesting stuff here. I have only limited experience with peltiers, but I am wondering if the vibrations caused by the light machining might cause stress to the TEC components over time?

You mentioned building a cold plate and I was wondering if you had considered using a second hand freezer and running the refrigerant lines through the cold plate, or possibly making the cold plate have channels with tapped ends into which you could couple your refrigerant lines to thus foregoing having to bond the lines to the cold plate.

You might also consider having some rigid/stable material that is a poor thermal conductor between the bottom of your cold plate and the table/bed of your mill and possibly the sides of the cold plate as well? This would help keep the 'cold energy' from transferring to your mill table and help concentrate it on your work.

Anyway just a few thoughts, good luck with it.