Teensy in an MRI

Status
Not open for further replies.
S

spectasaurus

Active member
Hi,

I have a research project that involves placing a Teensy 3.1 in a 3 Tesla MRI. Unfortunately it seems that this does not work and the Teensy becomes unresponsive. I have no experience with such things, but does anyone know why the Teensy becomes unresponsive? I am curious if there are any components on the PCB that might be susceptible to high magnetic field? Any help is greatly appreciated.
 
There is general concern about loops of wire having induced currents from the ever changing magnetic field. It's also a pretty heavy RF environment. You can't help the loops that are on the PCB, but you can help the loops and inherent antennas that are part of the typical crap on a breadboard prototype So... Step one is to get a virgin no pins teensy of your desire, load it up with blink, and give it the static 3T. Still blinking? Start running sequences. Still blinking? Try adding peripherals piecemeal.

Stops blinking? You might try different clock speeds, using only the teensy 3.1 5V tolerant pins, or switching to the Atmel based 5V Teensy for a deferent architecture. You might try shielding things in a non-ferrous metal box with non-magnetic screws.

Good luck! Hope this helps. Let us know what works in the end.
 
If you have "no experience with such things" you have no business doing experiments without the guidance of someone knowledgeable and responsible. Volunteers on the internet might be knowledgeable but certainly aren't accountable for whatever you break or whomever you injure. Try googling for 'MRI accident' if you aren't convinced.
 
pictographer said:
If you have "no experience with such things" you have no business doing experiments without the guidance of someone knowledgeable and responsible. Volunteers on the internet might be knowledgeable but certainly aren't accountable for whatever you break or whomever you injure. Try googling for 'MRI accident' if you aren't convinced.
Click to expand...

I have no experience with placing microcontrollers in an MRI. If you do, please offer suggestions. If you don't, then please don't provide advice on irrelevant topics like MRI safety. I am well aware of the possible effects of a 3T MR, and you don't just get access to such things as 3T MRI's without being aware of these things and having a certain degree of expertise. I know several world-class MRI physicists and while they have been helpful in this project, they to have "no experience with such things" either. But we are willing to try to figure this out.

Who knows, maybe the life we save one day with our research project will be yours.
 
Thanks drjohn. This is the approach we are going to take. The Teensy is already housed in a carbon fiber box, with no ferromagnetic components other than what might be on the PCB (ie, nickel plated SMD's?). I am thinking that perhaps the 3T is erasing the flash memory of the Teensy and essentially erasing the program or HalfKay. We are going to start placing it at various distances from the MR while running a sketch and see when it starts to fail. Thanks again for the suggestions.
 
The schematic for the Teensy 3.1 includes an inductor. A previous poster discovered you can remove the inductor and still have a working Teensy. How are you powering the Teensy?
 
When MCUs outright fail, one of the top reasons is the power is out of spec, either dropping out or spiking. That's why I'm wondering about the power. Does the USB power supply have a transformer in the field? If you have MI-safe batteries, you might try powering it that way with very short wires. A 3.3v or even a bit lower will work for an experiment.
 
Sorry nothing to add to this discussion. But I do have a question: Spectrasurus is only talking about ferromagnetic components. I was wondering about induction current in the groundplane. Ok the groundplane is small compared to the wavelength, but then again, this might be compensated by the enormous field strength.
 
Any conductor can be subject to rather interesting behavior near MRI machines.

Hate to be a kill-joy, but if for use with professional medical equipment where the device is around a patient and not in research environment, please consult the medical engineers at your local NRTL (CSA, ETL, MET, TUV, UL, etc). Or sit down and read (approx 5k pages) of IEC60601-1, IEC60601-1-2, IEC60601-2-23, ISO14971, ISO13485, and others that cannot be remembered by this old and aging mind.

Did some power converters for use in MRI when most machines were only 1.5T, and that was challenging, but interesting. The problem with most of these machines is that the equipotential references has enough current flowing such that the larger bus bars were not 'equi' potential. So reference nodes much be carefully designed.
 
Unfortunately I am familiar with those references. I designed some beam steering electronics for proton therapy. It was about 5 % design effort and 95 % of the time spent on generating evidence, writing qualification reports and collecting lots of signatures.
 
Yes, this goes without saying. We have no intention of creating a medical device for use in humans. Understandably, there are a lot of regulations that need to be followed for medical devices. This is not the issue here. We are basic scientists working on an idea at his point. You have to start somewhere.
 
Quick update: It appears the 3T magnetic field isn't the culprit, but the RF is. This is probably not surprising. It appears we are getting large 10V spike in the USB power in, that is frying the Teensy. Not sure why it's not frying the PC, but it may be more protected against these things. At any rate, we are going to disconnect the USB power and try a filtered external source.
 
Can you put the 'external source' (battery ?) and the Teensy in an aluminium box ? If you do have external wires, thread them through a common-mode choke -- if you can't use magnetic materials, you'll be challenged, but just looping the wires over an air-core is better than not.
 
Status
Not open for further replies.
Back
Top