New things to do...

We need more payload.

Do some calculations.

What is the current payload of the envelope when it is full?

What is the volume of the envelope?  Make appropriate approximations.  The website says "5 cubic feet" does this match what you measure?

What is the diameter of the envelope? 

If the envelope is filled with 90% pure helium and helium has a density of .18 kg/m3.  Air has a density of 1.2 kg/m3.  Given the volume of the envelope, what should the theoretical payload be? 

The mass of the envelope is about 15 grams.  Is our measured payload even close?

I can get a 52" x 36" envelope here in about a week or so.  I only want to if it will solve our problems. 

On other topics:

Write some code to read the ultrasonic sensor and then have the blimp hover exactly 1 meter above the ground.  You know how to read the US and control the motors and you have already tied them together.  Now you need to tie them together in a meaningful way so that feedback from the us is controlling the blimps altitude.  

The most common scheme for doing this is called PID (proportional integral derivative) control. This sounds fancy, but it is pretty simple.  For our purposes I don't think we will need the integral portion of it, since the balloon will oscillate a LOT.  

Read this overview of control systems 

Proportional control just means that your output is propotional to the difference between the desired altitute and the current altitude. IE if we are really far from where we want to be, we should spin the motors fast to get there.  If we are close, we should spin the motors slowly so that we don't overshoot. 

In practice we define and error as 

error = desired value - current value.

The proportional part just means that the output to the motors is proportional to the error!

Output = k1 * error    Where k1 is some constant which is tuned manually.  (When you take a control theory class you will learn lots more about this but for now, just try different numbers until it works.)

Read this section of my friend alex's web page: Proportional Control

The derivative part of proportional derivative control is as follows.  We wnat to be traveling parallel to the floor.  If the rate of change of the error is large then we are moving toward or away from it and should provide counter thrust to prevent overshoot.   

Calculate the rate of change of the error by storing error in a variable and then taking the difference.   If we were really worried about things, we would look at the elapsed time for each loop and divide, but that is close enough to constant that it shouldn't matter. 

Now read this part on Derivative Control.

We will use the derivative to modify the output as follows:

Output = k1 * error - k2 * DError

The DError term is opposing the error term.  

For putting this together see: Proportional Derivative Control

I need to talk to Alex about moving that stuff to a friendlier site.  

Ok, let me know how your progress goes.  I will try to get in and drop of the IR remote so we can play with that.