# Re: Cv and flow

```On Thu, 23 Sep 1999, Dave Gomberg wrote:

> For gases:
>
> If you are trying to use a valve to drop regulator output from a stable
> 10psi to the 1psi you need at the bottom of a two foot deep tank to make a
> bubble, and the gas is CO2 and the temperature is 75F, this means:
>
> CFM=7.2*Cv; Cv=CFM/7.2
>
> If we want 1 L (about a quart) per day, that is about 1/40700 CFM.

Experimental results from my own experience, as well as George and others,
shows that the rate can be adjusted quite nicely with the Nupro valves
(and not so nicely with others such as the ARO-1). So where's the
disconnect between the prediction and the results?

So first, how do we know it's 1 liter per day and not 10?  A 1mm
radius bubble per second comes out to 0.36 l per day.  A 2mm comes out to
8 times that... In my larger tanks I was going 2 or more bubbles per
second. There's an order of magnitude potential variance there.  Perhaps
this is part of the answer.

> Cv=1/7.2*40793=.0000034

= 3.4 x 10-6.

Nupro valve Cv = 2.5 x 10-4 at two turns, this means at 1/4 turn you
might get the Cv down to 2.5 x 10-5, which is now only 1 order of
magnitude off your calculation for full flow.  Can I get finer precision
than 1/4 turn on the nupro valve?  You betcha!  There's at least another
order or magnitude here.

> That is for full flow.  To be able to regulate to +-10%, you need to be
> able to modify the Cv by:
>
> .00000034

I submit that we don't need to regulate +/- 10%; that our reactors are not
that efficient (given loss to the atmosphere and such) that such changes
make significant difference. We are adjusting for factors of two or more.
1 bubble per second, 2 bubbles per second, etc.

- Erik

--
Erik Olson
erik at thekrib dot com

```