Human Olfactometer in Action

Today was a day of putting the new Alicat MFC powered human olfactometer through its paces.

We tested:

1. Onset Latency (NResearch final valve)
2. Concentration steps via flow dilution (Alicat MFC)
3. Concentration waveforms (Alicat MFC)
4. Odor sampling by human subject (YBS)

Here are pictures of our setup:

Olfactometer Test Setup 1: Kinetics.  The PID (photoionization detector) is sampling directly from  the nose port.  The scope is displaying PID output and flow from the flow sensor (below).  Scope is being triggered on final valve onset.

Closeup of nose port (Nasal Ranger) connected to our teflon odor injector (right top) and then Honeywell flow meter at exhaust (bottom).  Here the PID is sampling from the exhaust.

Test 1: Onset latency

Tektronix snapshot of odor onset.  Rise time: 50 msec, valve delay 50 msec.

We tested the onset latency by simply preparing a high concentration of Isoamyl Acetate (IA) and triggering the final valve.  We allowed 1 sec for odor flush time.  The odor (0.3 LPM) was first picked up by a 2 LPM  carrier stream and then a 20 LPM carrier stream for delivery to the PID.

Test 2:  Concentration steps

We next tested how the olfactometer produces steps of concentration.  We set the odor flow rate down from 0.3 to 0.15 and then 0.07 LPM.  The result is quite satisfactory:

Concentration steps from 300 to 70 ml/min.  Note fast return to baseline.

Test 3: Odor waveforms

We then tested how well MFC flow rates translated to odor output. We tested sinusoids and square waves at different frequencies (0.4, 0.8, & 1.6 Hz) while presenting odor for 5 sec (sort of an upper range of a single human inhalation).  While the olfactometer presented reasonable sinusoids at all frequencies, the high frequencies suffered from significant DC shift due to the filtering properties of the delivery tubing.  Below is the olfactometer output for the different sinusoids:

Frequency = 0.4 Hz

Frequency = 0.8 Hz

Frequency = 1.6 Hz

And here are the results for square waves:

Freq = 0.4 Hz; Duty: 50%

Freq = 0.8 Hz; Duty: 50%

Freq = 1.6 Hz; Duty: 50%

Freq = 0.4 Hz; Duty: 25%

Freq = 0.8 Hz; Duty: 25%

Freq = 1.6 Hz; Duty: 25%

The above responses are pretty good and show that it is possible to get a series of concentration bursts with a frequency of up to 1.6 Hz.  The rise time is especially impressive.  The biggest problem is the slow decay in odor probably due to sorption in the tubing.

Test 4: Human sampling

We then wanted to know how good the system was when being sampled by human subjects. Specifically we were concerned that while we are able to produce odor waveforms at the output of the olfactometer, humans would actually inhale somewhat distorted signals due to the volume of the odor port.  We therefore positioned the PID at the exhaust of the nose port and covered the port with a rubber membrane (substituting for the nose) so that all the flow is diverted to exhaust (and the PID).

Human-like test setup (worst case scenario).  PID is measuring at exhaust port (below) while nose port is covered by rubber membrane (glove).  Any signals measured here have travelled through the full volume of the nose port as well as the exhaust tubing (i.e. the Honeywell flow sensor).

Olfactometer output at exhaust with port covered by glove (above).  Input was a square wave (Freq = 1 Hz, Duty = 25%).

We then replaced the glove by a real human nose (mine) and tested again.  When I did not inhale the output was identical to that with the glove.  However, with an inhale, the PID no longer picked up anything at all!  Meaning I was inhaling the whole odor waveform!

This is me with my nose stuck into the odor port.  Stylish!

Odor at exhaust when I'm not inhaling.  The blue line is the output of the flow sensor.  It goes up when I'm inhaling.

Odor at exhaust when I'm inhaling.  NOTE: NO ODOR.  The flow sensor is registering a prolonged increase in flow proportional to my inhalation. 

So this is nice, we are getting ready to do some science!