how to make a 1mv input

[Bill Tyler]

Hey guys, I am trying to figure out how to get a 1 millivolt signal to inject into a device to test and calibrate the recording of the signal.

Would I use a signal generator or a function generator for that?

[Akarsh Simha]

What's the difference between a signal generator and a function generator?

I hope your signal is at a fairly high frequency, because the low
frequencies are typically dominated by noise in any system.

I think it would be hard to get rms noise ~1mV so that you can
actually test 1mV, so ideally, your signal should be a sharp peak at a
frequency that is high enough.

Regards
Akarsh

[Riley]

What's the signal supposed to look like? Analog like audio or eeg/biosensor feeds or digital as in clock/data/etc?

-
R

[Bill Tyler]

Thanks guys, let me give a little more detail, sorry it was an all nighted trying to get something figured out.  I am working on tuning some software for displaying ecg collected from my product.

The device works ok, but there is no filtering so the signal is usually unusable unless you are sitting perfectly still.

I am trying to calibrate the software and get to a known good display, and then start working on some filters for signal scrubbing and beat detection.

My thought was that I could hook up a signal generator of some sort, to take the human out of the loop and calibrate the display this way.

So a 1mv flat line to get the gain calibrated such that is is displayed as 1mv on the display, then various wave forms for verification that what I display matches what is injected.

Then back to human connection to determine where to go next.

This is somewhat new to me so please correct me where I am wrong.  And thanks for the help.

--
You received this message because you are subscribed to the Google Groups "ATXHS Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to atxhs-discus...@googlegroups.com.
To post to this group, send email to atxhs-...@googlegroups.com.
Visit this group at http://groups.google.com/group/atxhs-discuss.
For more options, visit https://groups.google.com/groups/opt_out.

[Danny Miller]

Many ways to generate a signal.  You must specify details for any sort of meaningful answer.

The Rigol scope has a 1v P-P output, IIRC a square wave.  With a resistor bridge you can divide it out.  You can verify the amplitude with the scope.  But a) it's a square wave, which may not be appropriate, and b) if this is to measure THD, then you have to know the THD of your input signal first.  And c) the frequency of course has to be appropriate.

Danny

On 1/22/2014 11:01 AM, Bill Tyler wrote:

Hey guys, I am trying to figure out how to get a 1 millivolt signal to inject into a device to test and calibrate the recording of the signal.

Would I use a signal generator or a function generator for that?

[Akarsh Simha]

I'm rather inexperienced too -- I've never built a circuit that
accepts anything below 10mV. One of my "low noise" circuits had 4mV of
rms noise when seen on an oscilloscope.

You could try and produce a 1mV signal from a function generator, but
in my limited experience, it is quite likely that you might see some
sort of noise instead.

How are you inputing the signal to the computer? Do you have some
information on the noise in the channel that you are using to input
the signal?

Regards
Akarsh

[Akarsh Simha]

In principle, you could convert a square wave into a near-sinusoid
with passive RC filters.

Regards
Akarsh

[Danny Miller]

It would be much simpler and yield a better quality signal to just make
a sinusoidal oscillator.

Danny

[Robert Ristroph]

I built a ECG as a lab project in college.

A 1mv sign or square wave is fine. If your signal generator doesn't
do anything that low, pass it through a voltage divider (two resistors
in series) to get the range you want.

I would build the filter first, though. When you can see the signal
well on an oscilloscope, then start working on the software display.
This is because your filter may change the range of the signal in some
way in addition to cleaning things up.

Finally, please do not connect anything connected to mains power to
your body, especially to leads placed so that they cross your heart,
like an ECG leads should be. The ECG should be able to be powered by
a small battery (9v rectangular batteries are good), and either save
the data to a SD card or something to transfer it, or communicate it
via radio or optical isolators. Regardless of your confidence in your
circuit and components, it's easy enough not to. Ask for advice here
if if you are not sure.

--Rob

> --
> You received this message because you are subscribed to the Google Groups
> "ATXHS Discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to atxhs-discus...@googlegroups.com.
> To post to this group, send email to atxhs-...@googlegroups.com.
> Visit this group at http://groups.google.com/group/atxhs-discuss.
> For more options, visit https://groups.google.com/groups/opt_out.

--
http://rgr.freeshell.org/

[Danny Miller]

But your amplifier will likely have its own filters. As such, a square
wave will not only be distorted towards a sinusoid, but its amplitude
will be screwed with- defeating the point of calibrating the amp gain.

Danny

[Bill Tyler]

Thanks guys!

The device is built so my filters have to be in the software for this
version. Which is why I am looking to have calibrated input.
I think I am just going to have to go with an ecg simulator. I tried to
use the function generator at tech shop this afternoon and could not get
it register with my device.

[stevek...@austin.rr.com]

An Audio Precision Portable One Plus could possible do what you need.  It will output a very low distortion (0.002%) sine wave from 10Hz to 120KHz from 26Vrms down to 200uVrms.  Depending on input impedance setting (40 to 600 ohms), the noise floor on the analyzer portion is anywhere from 1uV to 19uV with a 22Hz - 22KHz BPF setting.  Would that work for you?  I can bring it by the space sometime if you wish but cannot leave it as I use it at work daily.

 

Steve