Copper coils for nMRI Given the lack of financial efficacy in the medical industry, I was wondering if nuclear magnetic resonance imaging would be possible with a copper/silver coil if the use is short 10 second imaging bursts.
Or is the resistance simply too disruptive for any imaging to occur?
What I had in mind is a very small machine you could stick an arm or leg through for home scanning. Either completely uncooled or a very inexpensive cooling technique like simply running faucet water through it for the brief duration in a bathroom.
 A: Several reasons why this is not a great idea: 1) signal to noise ratio scales with field strength: 1.5T is considered a reasonable number although it is possible to get diagnostic images in some situations with lower fields. 2) inductance of coil capable of producing such large fields - try computing the energy stored! 3) during ramping of the magnet, you create eddy currents. Ramp too fast and these currents will heat up other parts of the system (this is a problem even with the gradient coils which ramp in the 10's of milli Teslas per second). 4) need for homogeneous $B_0$ field: all this thermal expansion makes it impossible to shim the magnet to the required homogeneity. 5) Change in geometry because of coil heating: you need field stable to less than a few ppm (although you can tolerate worse if you don't intend to do spectroscopy).
A few months later...
I gave this some more thought. A few more problems that I can see (and these are very real):


*

*An MRI machine is dangerous - even at low field strengths. It cannot be made into a DIY scanner - if you have any metallic implants, pace makers, jewelry etc you might find yourself trapped, burnt, lacerated... Before people are allowed into an MRI suite (the vicinity of the magnet, let alone in it) they are evaluated to make sure this is safe for them.

*You don't just need the static B field: you need RF power (to stimulate the signal) and gradients (to create the image) as well. Although advances are being made for "silent" MR for certain exams, MRI machines are typically very loud - another reason why you need them in a shielded room, and why the patient has to wear good quality hearing protection. Again not something to do without supervision. The RF power (128 MHz at 3T, but scales linearly with field strength) also needs shielding, or it will jam all kinds of radio transmission and other equipment (typical power for whole body scanner 10's of kW). The voltages in the gradient run into the kV range, although it might be possible to make this lower (and accept slower / worse resolution imaging)


Having said all that - yes, you can probably make an MRI image with a scanner that has water cooled copper coils and that doesn't take out the electrical grid when you turn it on. You might be able to build it for a lot less than a superconducting scanner - maybe 100k dollars if you just want to scan extremities (arms and legs) - it has been done.
The question is - what then? Under what circumstances does such a scan give information that will help you? 
In the end, a medical diagnostic image is intended to give information to a care giver that allows him or her to select the right treatment to make the patient better.  An MRI machine can be a physics curiosity, or a medical tool; but if it is the latter, the information needs to end up in the hands of a doctor, and help this doctor to select a treatment that could not have been chosen without this information.
I don't want to sound pessimistic, but I don't think a kickstarter IPO is going to get you a billion dollars...
See also http://magnetic-resonance.org/ch/03-02.html for some more information
A: It might not actually answer your question, but to throw it into the bowl:
There are some advances in MRI using permanent magnets and even conventional electromagnets with static magnetic fields of about 0.5 Tesla. As far as I know one can do imaging with a reasonable resolution with these devices without the need for extensive cooling. They are used for MRI of the limbs and open MRI when guided intervention is needed. Because they are more easily handled then supercondicting magnets, there are devices that can turn and therefore image the patient for example in an upright position (for spine-imaging in different positions/ different states of weigth). These devices should be much cheaper than conventional MRI, too, although not cheap. But I would think they should be more cost-efficient.
However, everything Floris said is valid for these devices, too, so one will not be able to use them as self-scanners at home.
