Re: For those who say the Merkin education system is a bunch of crap....

On May 12, 12:24 pm, Scotty <nob...@xxxxxxxx> wrote:
On Mon, 11 May 2009 03:31:17 -0700 (PDT), HardySpicer



<gyansor...@xxxxxxxxx> wrote:
On May 10, 12:28 pm, "deemsb...@xxxxxxx" <deemsb...@xxxxxxx> wrote:
   http://instruct1.cit.cornell.edu/courses/ee476/FinalProjects/s2009/ra...

Yes, the calculation for cutoff freq of the simple high-pass filter
should be

1/2piRC and not 1/RC and the Bode plot is dodgy - is that supposed to
be a dB scale?

Prototype wiring is appalling but no worse than most students of that
age-group. A few dodgy bits in the op-amp circuits as well but overall
a good try and no worse than other countries. I expect their software
is good. Younger people excel at software and are crap at hardware.

Hardy

I'll tell you what few people below the age of forty can do and that is
analogue design. The formula given is for time constant. The frequency of
the filter is out by a factor of 12, 6 for the 2pi and 2 for the next
comment.

It's impossible to tell whether it was intentional but they have two
high-pass filters, one at the input and one at the output both with the
same design fault. The 2 X 1 k ohm resistors are in parallel because the
power rail is at "signal ground" which doubles the cutoff frequency.

There is also a superfluous 1k5 resistor in parallel with the voltage
divider at the output. If it was needed to adjust the DC level into the ADC
then they should have recalculated the upper resistor.

There is also a superfluous 1k ohm resistor in the feedback circuit of the
loud speaker amplifier.

There is also another high pass filter around the feedback network although
the cut-off frequency is much lower.

I would like to see a spectrum analysis of a fart, I'm sure that the 100Hz
cutoff is too high, especially for those  low frequency flappy farts one
experiences after a spicy curry...

The construction is about the level of a hobbyist and not surprising for
high school students but disappointing for a Cornell student.

But all that is trivial when the response time of the transducer might be
24s, a bit slow for the application unless they are farting God Save the
Queen...

The scale is probably logarithmic as it appears to be asymptotic to a
straight line and should have a slope of 12 dB per octave and 18dB per
octave below abt 10 Hz.

They also seem to have drawn quite heavily on established software
libraries.

Still, top marks for presentation and amusement...

Yep I spotted the resistor divider and the high pass filter preceeding
it. A few clamgers. It's good to know that Cornell don't produce
anything better than average like anywhere else. Analogue circuit
designers are becoming a rarity. You can command 200K US as an IC
designer for analogue or mixed signal design.


Hardy
.

Loading