You are here

Magnetostrictive transducer to build

Hi Dr. Chris Cheers,

I'm an Italian student at Politecnico di Milano.

For my final work (thesis) I am investigating the relationship between frequency and quality of ultrasonic welding of metals.

To do this I've to build a magnetostrictive transducer+sonotrode to be tested at different frequencies.

Do you know where I can find informations about dimensioning the different parts of the transducer (diameter of wire, current, dimensions of the pieces of magnetostrictive (Fe-Co) matherial) and sonotrode (something like an old project for a tranducer)?

Forgive my poor English.

Thanks.
Valerio

Hi Valerio,

I think Etrema are the only people still manufacturing magnetostrictive transducers - they might be willing to give you some information on transducer design, or at least material properties that you could use to estimate dimensions and turns required:
http://www.etrema-usa.com/

For length of transducer and sonotrode calculate sound velocity (square root (E/rho)), then wavelength at your chosen resonant frequency (sound velocity/frequency), then sonotrode length is half a wavelength. I'll have a calculator up on the web site to do this soon.

Wire turns and current are based on the alternating field you're trying to produce - look for the linear section on the magnetostriction graph for the material. Obviously you'll have a range of options - higher current with fewer turns or lower current and more turns. I'd go for a relatively small diameter wire (so it bends around the corners) and as much current as it can possibly take! You must use high-temperature insulation or it will melt... Also remember you also have to apply a biasing magnetic field to keep the material working in the linear response section, either as a d.c. current in the coil or using permanent magnets.

Having said all that, I don't think a single transducer / sonotrode will help you with metal welding over a range of frequencies. Welding demands significant power, which a magnetostrictive transducer won't produce off resonance (they're inefficient enough at resonance!). So to test different frequencies you'd need a number of different transducers and sonotrodes, each resonant at the frequency you want to use. And of course a higher-frequency transducer being smaller will have lower power capacity, which is likely to affect the weld quality more than the frequency change itself... Welcome to the wonderful world of power ultrasonics!

Good luck with the project. Let me know if I can help, and in any case I'd be interested to hear how you get on.

Regards
Chris Cheers

Thanks Chris.

At the moment I should have a good magnetostrictive material: Vacoflux 50 (50%Fe 50%Co) and the italian vendor gave me some informations and properties.

In your opinion this material is good for my purpose?

I'll try to contact Etrema to have informations about dimemsions.

[quote]For length of transducer and sonotrode calculate sound velocity (square root (E/rho)), then wavelength at your chosen resonant frequency (sound velocity/frequency), then sonotrode length is half a wavelength.[/quote]

I don't understand... is sonotrode+tranducer length half a wavelength or is only sonotrode lenght half a wavelength?

[quote]You must use high-temperature insulation or it will melt... [/quote]

I was thinking to use PTFE insulated wires which resist till 200-250C. They can be good?

[quote]Having said all that, I don't think a single transducer / sonotrode will help you with metal welding over a range of frequencies. Welding demands significant power, which a magnetostrictive transducer won't produce off resonance (they're inefficient enough at resonance!). So to test different frequencies you'd need a number of different transducers and sonotrodes, each resonant at the frequency you want to use.[/quote]

So I have to use different transducers and work only in resonance... you confirm my doubts. Now I have to discuss this problem with my teacher. I start to understand why in literature there is nothing concerning the relationship between frequency and quality in ultrasonic welding.

[quote]And of course a higher-frequency transducer being smaller will have lower power capacity, which is likely to affect the weld quality more than the frequency change itself... Welcome to the wonderful world of power ultrasonics![/quote]

To "solve" this problem I thought to use constant energy... E=power*time==> if power is higher I weld for a small time, if power is lower I weld for a longer time. This way I should have more comparable results. What do you think?

[quote]Good luck with the project. Let me know if I can help, and in any case I'd be interested to hear how you get on.[/quote]

Sure!

Thank you very much for your interest and for your time.

Ciao Valerio.

Hi Valerio,

Interesting. I found the site vacuumschmelze.com - is that the vendor? (so I can add them to the suppliers list). I haven't looked in detail but it looks like the material is supplied in solid pieces. Can you also get it in laminated form to reduce eddy current losses?

For a resonant system, the length of the transducer would be a half wavelength based on the transducer material properties, and the length of the sonotrode would also be a half wavelength, a different dimension based on the sonotrode properties.

PTFE insulated wires are probably the way to go. You might also want a temperature sensor and automatic cut-off. And good safety guards. My experience with these things is that they get very hot!

Constant energy is a reasonable approach, which is why it's often used in plastic welding to allow for variations in joint quality / friction / contaminants etc. I'm not sure how meaningful it would be in the context of a significant frequency difference. You should also be thinking about amplitude, but I don't think you could realistically expect to obtain the same amplitude at different frequencies. I guess it would be interesting to see results covering all the available adjustments (so also perhaps contact force and tip area) at each frequency and see what trends emerge...

Regards
Chris

Hi Chris.

Vacuumschmelze is the manufacturer of Vacoflux 50 and of lots of other materials. In Milan (for Italy) we found a retailer of vacuumschmelze:
Sisram s.r.l. [url]http://www.sisram.it[/url].

Another possible manufacturer is specialmetals [url]http://www.specialmetals.com/[/url] but they sell above all nichel allloys, a good one should be Nilo 48 or Nilo 50.

They sell it also in laminated form with a wide range of thicknesses. They told us that for our purpose a possible thickness is 0.3mm

I have just gone to dicuss with my teacher about the work, I told him the difficults of which we discussed today and probably we will
make a little change to the purpose of my work (or better, our work, because I am working with another student: Nicoletta).

I'll write to you soon about these changes.

Ciao Valerio.

P.S:
I have to disturb you with another question.
It could be very useful to me to read also the parts of your thesis (chapters 2 to 6) which there are not in your web site. Is this possible?

Thanks Valerio.

It would be interesting to get some follow-up on this as you go on.

I mailed you the download address for my thesis privately. Generally I'm giving this information only on request, free to academic users but for commercial / industrial users there will be a small fee...

Regards
Chris

Hi Chris,
forgive my delay, but I had a lot of problems to solve concernig my thesis.
I briefly summarize what happened to my work.
With my teacher, we decided that It was not possible for us to build many different vibrating groups in order to use different welding frequencies. So we decided to change the purpose of the thesis. We bought a commercial ultrasonic metal welder to study the welding process with a statistical approach in order to evaluate the possibility of applying it to aluminium car bodies.
Now we are looking for an aluminium alloy currently used for car bodies, and we are also searching informations about solid state diffusion which is the physical phenomenon at the base of the ultrasonic metal welding process.
We are going to perform a series of experiments (using the statistical method of design and analysis of experiments) to understand what kind of parameters really influence the resistance of welded joints.
Then we are going to analyze the section of joints using a metallographic microscope in order to oserve what happens to cristalline granes and to the layer of oxide which always coats aluminium sheets.

We need a book that explains what happens during the welding process at the interface of aluminuim sheets; do you know where we can find such a book?

During the experiments we have to misure the amplitude at the tip of the sonotrode, what kind of instrument do you suggest us?

We have to also to misure the force that sonotrode applies to the specimen, but we do not know where is the best location for the load cell.
We think that the best location should be between the knurled plate and the anvil because this way we could use only one load cell, but we fear that vibrations may destroy it (particularly if it is a piezoceramic one). What do you think about this problem?

Thank you very much.
Best regards,
Valerio.

Hi Valerio,

Good to hear you're making progress. I think the change to commercial equipment is a wise decision. It's always an attractive idea to start from scratch but in practice you'd spend such a long time trying to catch up with the current state of the technology that you probably wouldn't get to any new work.

As you're now looking specifically at welding of aluminium car bodies, you presumably know about the big US research project (ATP) now going on with Ford, EWI, Sonobond and AmTech? More info [url=http://jazz.nist.gov/atpcf/prjbriefs/prjbrief.cfm?ProjectNumber=00-00-55.... Perhaps there's some way you could link your research with theirs?

Sorry I can't help you with books on the solid-state bonding process. I did recently find Edgar de Vries' 2004 [url=http://www.ohiolink.edu/etd/view.cgi?osu1078415529]dissertation[/url] which contains a good discussion of this and many references to papers detailing the theory. This work looks to me very close to your research as now described - I'd suggest you review all his work if you haven't done so already.

Re. amplitude measurement - the above also describes use of the [url=http://www.mtiinstruments.com/gaging/products/mti-2000.html]Fotonic sensor[/url], an optical system from MTI that seems ideal for out-of-plane ultrasonic vibrations. Not cheap though - what's your budget?

Per your e-mail request I've put up this equipment diagram you sent.
[img]http://www.powerultrasonics.com/images/LoadCell.png[/img]

I think your fears about putting a load cell under the knurled plate are very reasonable - the whole point of ultrasonic metal welding equipment is to generate shear stresses from the lateral vibrations and I wonder how well the load cell would cope with these (perhaps one to ask the manufacturer?). If you just want a static or average force then placing the load cell under the anvil would probably be ok. You could perhaps support the anvil at 3 points, one being the load cell, and calibrate to allow for load sharing. Then the vibration forces would come through as direct loads so you'd need to specify the load cell to allow for much higher forces than you expect, and filter out these dynamic load signals. But if you want to measure the dynamic forces during the welding process that's even more difficult...

Sorry if this all sounds a bit negative - I'm just trying to offer the best advice I can. Feel free to come back with more questions. I hope the research works out for you and please let me know how you go.

Regards
Chris