Tuesday, November 01, 2011

Front End Turns PC Sound Card into High-Speed Sampling Oscilloscope

Various software packages enable PC sound cards to provide oscilloscope-like displays, but the low-sample-rate, high-resolution ADCs and ac-coupled front ends have limited bandwidths. For repetitive waveforms, a sampling front-end stretches the time axis, allowing the PC to be used as a high-speed sampling scope. This article describes a front end and probe that provide an appropriate adaptation.


At 5:05 PM, Anonymous Anonymous said...

All sound-cards include a 20 Hz. high-pass filter that blocks DC. Your sampler passes DC to the sound-card - which then blocks it.

A useful 'scope should extend to DC.
Your's seems to be AC-coupled always.

At 8:24 AM, Blogger Analog Dialogue said...

You are correct, anonymous; just about all sound cards are ac coupled, but you can’t do anything about that. In some Internet hacks, people have removed the ac-coupling caps (these are external in some models) from cheap USB sound adapters, adding an op amp to level shift the 1.5 V dc common mode on the input of the single-supply ADC to provide bipolar input voltage swings. But this is not something average students want to do to their laptop computers.

The intent of the article was to introduce the concept of sampling--specifically undersampling--as a means to frequency translate signals, thereby stretching them in time.

Having the ac/dc coupling option at the front end is useful to block the large dc component of a signal having a small ac component from the input S/H. DC coupling the input might seem unnecessary when using this with an ac-coupled sound card, but other examples of low cost, low bandwidth USB hobby scopes could benefit from this front end. This one, for example, is available for $49.00: http://www.gabotronics.com/development-boards/xmega-xprotolab.htm. One of these plus my front end gives you a 2-channel 50-MHz scope for less than $100.

At 8:02 PM, Blogger Oskar said...

hope this blog is still alive ...
AC/DC issue: when connecting a 10x or 100x probe to an instrument, the input is expected to be connected to ground through a 1MegOhm resistor and some parasitic capacitance, which is not of interest for this issue. In the schematic shown, with AC setting, there is no resistor on the input side, only on the amplifier side; therefore the input will rise to whatever the mean value of the voltage on the probe tip is, which can be 500V or more when working with a 100x probe. Obviously biasing the input stage with (almost) infinite resistance becomes a design challenge, in order to maintain the standard 1MegOhm input resistance. I have seen this problem on many hobby type schematics. What is the correct solution?

At 10:02 AM, Blogger Analog Dialogue said...

The AD783 S/H input needs a DC path to ground to bleed off any tiny leakage
current so some part of the resistor needs to be on the device side of the AC
coupling capacitor. The reader is also correct in pointing out that if the probe does
not also include a DC path to ground on that side of the AC coupling capacitor
the DC value could reach a very high value and the AC coupling cap would need
to be able to stand off this high DC voltage.

The solution would be to include 1/2 of the total 1 Meg Ohm input resistance on each
side of the coupling cap as two 2 Meg Ohm resistors, one on each end of the cap
to ground. The AC parallel impedance would still be 1 Meg so the AC gain would be
right but the DC gain would be wrong but who cares in an AC coupled situation.

The intent of including the AC coupling cap in this application was to remove the
small DC component of say a 0 to 5 V logic signal and better center it in the
+/- 3 V input range of the AD783 S/H input.

At 12:41 PM, Blogger Sasuke Uciha said...

Good Article, agen souvenir pernikahan murah. thanks for your article. :)

At 6:56 PM, Blogger Uzumaki Naruto said...

Good Article, harga souvenir nikah. thanks for your article. :)


Post a Comment

<< Home