-
-
Skip navigation
(Access key: S).
Contact information.
The project requires designing an MRI-compatible small animal ventilator using a stepper-motor system for controlled gas delivery. The design will deliver small volumes of an oxygen and hyperpolarized helium mix in a study using hyperpolarized helium Magnetic Resonance Imaging.
This project was continued from last semester. To see its website, please click Animal Ventilator Fall 2006.
We worked hard over last semester to design and build a new and improved ventilator. We needed to make a number of changes that would improve the functionality of the device. First, we needed new placement of the syringes, which we placed at the front for easy access for interchanging. Second, we wanted to scale the entire device down because our original model was unnecessarily large. Third, we wanted to make our device capable of makeing repeatable volume deliveries. Our original model had some flex and unwanted movement in a number of parts. For our new design we corrected these with either better machining or we redesigned the mechanism.
In the end, were very proud of the ventilator that we produced. It performs just how we had designed it to and it was a great improvement over the original model. If this semester taught us one thing, its that it doesnt matter if your first model doesnt work perfect. Its a prototype and thats what its for, to pinpoint problems. You need to target those problems and redesign to make a better one. Engineering is constant improvement, and this semester, were happy to say that our new ventilator is a successful design. Please read on to learn about the project in more detail, and some of the obstacles that we needed to overcome to get to where we are now.
Background
Medical imaging systems have proven their capabilities in diagnostic means and physiological verification. Magnetic Resonance Imaging (MRI) is superior in detecting soft tissue contrast. Another advantage with MRI, unlike Computed Tomography (CT), is the ability to image in oblique planes. This is very advantageous in a number of different applications, such as interventional MRI and the diagnosis of abnormal tissues.
The conventional clinical MRI detects signal from the H1 protons in the body which resonate at 42.58 MHz/T. There are other potential sources of signal, like 31P (17.24 MHz/T) and 13C (10.71 MHz/T). A newer technique in MRI is the use of hyperpolarized Helium (He3) as the source of signal. In order for Helium to have the magnetic properties that allow it to be imaged with MRI it needs to be hyperpolarized (He3). The hyperpolarization gives the Helium a heightened spin state that is required to give an MR signal. Although there is no natural He3 in the human body, it can be safely inhaled to achieve signal in the respiratory tract and lung systems, which are often difficult areas to image with conventional MRI because of the air interfaces. Oxygen has a paramagnetic effect that destroys the polarization of He3. Therefore, great care is taken to avoid mixing the two gases prior to the scanning.
He3 MRI utilizes the same scanner as except instead of tuning the scanner to detect the signal from H1, it is tuned to Helium which resonates at 32.43 MHz/T. Therefore, the scanner detects the signal specifically and solely from He3. The He3 gas is typically inhaled while the MR scanner acquires the signal. Conventional image reconstruction techniques can be performed with the acquired data to yield images of the signal. During the onset of He3 inhalation, it is possible to view in the time-resolved image set the signal traveling down the trachea in humans much like the bolus of contrast in Contrast Enhanced MRI (CE-MRI). Thus, He3 MRI can reveal physiological information about both the structure and therefore function of the respiratory system. Current animal studies are being performed to test the ability and efficacy of He3 MRI in diagnosing respiratory diseases. The He3 MRI interpretation of a respiratory disease can be compared to both Positron Emission Tomography (PET) images and histological examination.
Below are two examples of the image quality achieved with He3 MR imaging.
(Left) Image acquired using a three-dimensional (3D) imaging sequence. The spatial resolution is exemplified by the surface-rendered volume of the 3D data. The ribs and tracheal rings are visible, as is a ventilation defect in the left lung (arrow). (Right) He3 MRI in a healthy volunteer. Branching to the fifth generation of airways is visible. Recall, these are not blood vessels, these are airways that contain He3 at the time of the scan. These visible "bronchioles" reveal the shape and efficiency of the lungs.
Past Setup Used by Client
Below are pictures of the past setup. The stepper motor allows He3 air to be drawn into the syringe from a holding bag. The air is then injected into the gas tubes leading to the small animal. The client’s setup with the stepper motor is unnecessarily large. Our goal is to create a smaller motor driven syringe gas delivery system that mixes Oxygen and He3 at the precise time it is inhaled by the small animal. This will cut down scan time from 8 min to possibly 2 min. (place cursor over picture for caption)
First Generation Prototype Designed by Our Team
The following pictures represent the original prototype that we had built during our first semester to satisfy the client’s needs.
Summer Work
We performed trials over a range of motor steps, from 250-400 steps. We attached the output of the device to one end of a manometer (tubing and a meter stick) and recorded the beginning water level. After one "breath" or movement of the specified number of steps, we recorded the end water level. Calculating the difference in water level and dividing by the number of steps gave us "inches/step". We first converted inches to mL volume by using the cross sectional area of the inner diameter of the tube and the displacement height of the water (calculated by hand to be 0.201101 mL/inch). Inverting this ratio we got our desired step/mL. We then calculated the mL output given by the specified number of steps.
Problems that Occurred:
When we found this equation that described the output with a given input, and entered it into the appropriate vi file, we then attempted to verify that indeed a desired output volume was actually outputed. However, this was not the case. Everytime we performed a verification trial, the output was consistently off by no more than .3 mL. Although precise, the actual output is not yet accurate. Other trials were done, and equations tried, with no better accuracy. The calibration procedure may need to be modified to achieve better accuracy. Also, the new model with significant structural and mechanical changes will significantly improve our results.
This is where our second semester (Fall) began...
The Fall semester we focused on two main goals: 1) fix valve timing and 2) develop new improved model.
Goal 1: Valve Timing
During testing this summer, we used a monometer to watch and measure the output volume of gas that the device produced given a specified input. While we were watching the water level breath after breath, we noticed that every fourth breath (when the software switched to our device for the He3 breath) there seemed to be an overshoot of gas. We spent hours trying to pinpoint why this error occurred. It turned out that our device was performing as expected but the pneumatic valves that our client was using were at fault. After determining the order of valve firing, opening and closing, we diagnosed the problem occurs in the software. One of the valves opens too early and causes an "oxygen chaser" as we’ve learned to call it. This oxygen chaser increases the output every fourth breath. This was a problem that has been occurring before our group began the project. Our client was pleased that we found and diagnosed this problem, something he didn’t even know was present.
The valves function off separate counters in LabView, so it was a matter of changing the timing of the counter so that the valve will open at the appropriate time.
Goal 2: New Improved Model
There were several improvements we wanted to make from the original prototype. These will be highlighted in reference to the Solidworks pictures we used to aid in the designing process.
Size
The original model had 5 inches of travel for the sliders, and thus for the syringe plunger as well. However, the syringes currently being used in the study have only a 3 inch travel space for the plunger, so we decreased the size of the overall design since the length was not needed. The sliders and aluminum rods are also smaller because the extra size of the original was not needed.
Syringe Placement
Since our client wishes to exchange the large syringes out for small ones when imaging mice, we needed a better, more accessible position for the syringes. We decided to position the syringes off to the side of the sliders (see pictures below). This gives more accessibility and lead to an improvement of syringe attachment.
Syringe Attachment
Our original mechanism of attaching the slider motion to the syringe plunger was beginning to show signs of wear and tear. Therefore we wanted to focus on a new method of attaching the syringes to the sliders. We designed a plate that will be attached to the back of the sliders and will pinch the back of the syringe’s plunger (see picture below). Screws will maintain the force needed to maintain the connection.
Below are more pictures of the new improved model we designed during the Fall semester:
This is where our final semester (Spring 2007)began...
Because our client needed a device for his continuing studies, we needed to try to optimize our first prototype for him while we continued to build our new model. To do this, we installed aluminum U-shims to the back of our sliders as shown below. This deterred any flexion of the syringe clip that was evident in prior experiments. After the shim was installed, no flexion was observed.
Shown below is the data we measured for calibration of our device after the shims were installed. An ideal relationship is 1:1 for volume requested and volume delivered. Our results almost fit the line perfectly.
The final model that our team designed is shown in the Solidworks renderings below. From these models we took 2D CAD drawings and built each piece with precision.
Below (left) is a comparison between the model in Solidworks and (right) the physical model we built.
Here is another comparison of our design vs what we built:
Results
The device was tested using the developed software with He-3 MR imaging. A small bag (2x2 in) was ventilated with a 02/He-3 gas mixture. Both gases began in their respective reservoirs and were pumped by their respective syringes at the appropriate 20% 02 : 80% He-3 volumes dictated by a 2mL tidal volume. An 80 breath per minute breathing rate was used and passive exhalation was utilized. Scan time was ~9 minutes.
The MRI unit (1.5T General Electric, Waukesha, WI) was properly tuned to pick up the He-3 signal and a conventional protocol was performed for small animal He-3 MR imaging. Projection MR imaging was performed over the time course of a breath and a half and over many trials to get adequate Signal to Noise (SNR) and temporal resolution. The final time data set of images allowed viewing of individual slices of the bag over the time taken for one and a half breaths. Inflow of signal was visible down the delivery tube and entry into the bag as shown below. The time resolved data set demonstrates the adequate He-3 delivery by the developed ventilator.
| Week | Reporting Period Beginning | Activities |
|---|---|---|
| 1 | January 26 | Planned out timeline of semester, goals |
| 2 | February 2 | Testing at Waisman, performed validation trials for new calibration curve for old device |
| 3 | February 9 | Machining of parts designed on Solidworks |
| 4 | February 16 | Waisman testing of old model for client study, LabView changes in automatic handshaking |
| 5 | February 23 | Reslotted syringe block holders, modeled axle support |
| 6 | March 2 | Machined last of motor housing and axle support |
| 7 | March 9 | Assembled entire model and prepared for midsemester presentation |
| 8 | March 16 | Midsemester Presentation, plan out calibration |
| 9 | March 23 | Designed small syringe blocks and planned out LabVew code hierarchy. |
| 10 | March 30 | Updated Client’s Labview and began writing code to run both syringes simultaneously. |
| 11 | April 6 | Designed new way to calibrate syringes, continued LabView code |
| 12 | April 13 | Continued LabView code, began writing paper for IEEE |
| 13 | April 20 | Added to and edited paper, began final poster, finished LabView code |
| 14 | April 27 | Final testing with MRI, confirmed proper functionality of device |
| 15 | May 4 | Practiced for final presentation and printed poster |
 |
Solidworks Design Concept Movie (Mar 8 2007, 4602 kb) |
 |
Midsemester Presentation (Mar 8 2007, 5840 kb) |
 |
Product Design Specification (Mar 12 2007, 13 kb) |
|
Final Poster (May 6 2007, 283 kb) |
|
Small Bag Ventilation Scan Results (May 6 2007, 226 kb) |
|
Final Paper (in IEEE requested format) (May 10 2007, 496 kb) |
