DIY MRI scanner

A 'proof of concept' idea. Instead of superconducting magnets, I'm using a pair of Lorenz coils, operating at around 5Mhz - hopefully.

Similar projects worth following
I have an electronics background and I used to have a job in an R&D dept designing NMR instruments, so what else could I build!. It will be quite small, about the size of a shoe box I estimate (excluding power supplies). And rather than having to move in order to house 3T superconducting magnets, I'm using a pair of 10cm Lorenz coils with an operating frequency of 5Mhz (ish). If it works, I may upgrade to a pair of 20-30 cm coils.

I will be making as much as I can, so I will have to make the power supplies, signal generators, detector, mechanics, etc. At the moment I am making the coils, which is not simple, as the first thing I needed to build was a winding counter. In order to achieve the best field homogeneity, they need to be wound as accurately as possible, so the winding mechanism needs to be robust. Possible issues will be heat dissipation (40-50 watts each), mechanical strain (attractive force between coils), rigidity.

Lorentz coils are proving to be trickier than expected. In order to operate at 5Mhz, the field needs to be apprx 117mT. To achieve this I need a coil of 1000ampere turns and 10mm in width. This is going to be difficult to implement. So rather than spend a great deal of time and resources, I have decided to lower the operating frequency to 1Mhz. The coils and power supply will be much easier to put together. It will incur a lower signal to noise, but at least I will be able to get something going and hopefully i may be able to upgrade it at a later date (if it works at all).

  • 1 × Coil former 21/2" plastic drainpipe.

  • Magnet power supply

    jebullard07/23/2015 at 11:20 0 comments

    The magnet power supply needs to supply a constant current (1-2amps) and must be stable to within about 10ppm as this defines the resolution (I haven't actually worked out the required resolution yet, but I have to start somewhere). The magnet power supply is contains an op amp, so I need a power supply for that. Just finishing a +/- 15v power supply for it. (as I said before, I'm making as much as I can myself)

    While I was doing it, it occurred to me that I could use a Maxwell coil. This uses three coils and provides an improved field homogeneity. The geometry is simple enough and has the added benefit that I can use my existing coils and because they will need to be placed further apart, it will give me a larger sample space. Also, if I have better homogeneity, I may not need golay coils to trim the field. Lastly, because I would be using three coils I should get a greater field strength which in turn increases the operating frequency and hence improve the signal to noise. (note. I guess it should be possible to create a four coil system, but I can't find any details about that, and the 'n coil' scheme requires greater engineering capability than I possess).

    The only drawback is that I will need to wind a coil on a former that is apprx 120-130mm dia. Search as I may, I can find nothing that size. So what I will do is contact someone locally that has a 3D printer and see if they can make me one. If that works, I might get them to make the field gradient coils too.

    I've also been looking at ultra low noise tuned amplifiers and it looks promising. I'm thinking of using a four wire 'twisted pair', which should have a lower characteristic impedance and hopefully reduce the noise figure. However, it is likely to make the matching network a bit fiddly.

    It will take me a while to get the 3D former sorted, so in the meantime I will carry on with the 2 coil system in order to see what results I get with the tuned amp - just how bad the signal to noise is (might need to stick the whole thing in a Faraday cage).

    So, by my next entry I should have acquired a basic resonance signal.

    PS. If anybody can throw some light on field strength formula for 3 or 4 coil systems, I will be grateful.

View project log

Enjoy this project?



Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates