Antialiasing filter strength?

Post any questions you may have about our current range of oscilloscopes
Post Reply
Ben321
User
User
Posts: 5
Joined: Tue May 23, 2017 8:09 am

Antialiasing filter strength?

Post by Ben321 » Sat Mar 10, 2018 9:01 pm

I have a Picoscope 2204A, and discovered that it can alias signals that are higher than the nyquist frequency. That is, when sampling at its maximum sample rate of 1MHz (maximum sample rate for fast streaming, which is needed for my application), the maximum frequency it should receive (nyquist frequency) is 500kHz. However, I discovered, when I connected a long wire to the input to act as an antenna, AM radio broadcasts are showing up within the 500kHz bandwidth of the scope. The FCC designated AM radio broadcast band runs from 530kHz to 1700kHz, so these signals should be outside of the bandwidth of this scope. But a number of these signals are appearing within the 500kHz bandwidth of the scope. This is what is known as aliasing, and requires a fairly strong, multi-stage, lowpass filter to prevent this from happening. Is there any command I can send to my Picoscope unit to increase the strength of the antialiasing filter, such as commands that activate relays that add more filter stages into the signal path? I was hoping to use my Picoscope as a LW (longwave) radio receiver, that receives all signals BELOW the AM broadcast band (I think some special utility, and even government or military signals can be found in this band), and this frequency space is getting cluttered with aliases of AM broadcast signals.

Gerry
PICO STAFF
PICO STAFF
Posts: 261
Joined: Mon Aug 11, 2014 11:14 am

Re: Antialiasing filter strength?

Post by Gerry » Mon Mar 12, 2018 12:28 pm

Hi Ben321,

You may be placing to much emphasis on aliasing.

As explained why, on a previous post, our PicoScopes don't have anti-aliasing filters. So, if you don't want aliasing you need to limit the frequency content of your signal and/or provide a large enough frequency range between the non-attenuated and fully attenuated signals that you will be capturing.

Bandwidth is a function of the input, not the digitizer (ADC) or sample rate. So, if you are not managing the gap between your signal bandwidth and sample rate (i.e. 1:5 ratio), for your intended application, by either limiting the high frequency content of your signal, or selecting a high enough sample rate (by sampling fast enough or using a faster PicoScope), your pass-band can indeed be higher than 500kHz at a 1MS/s sample rate (the bandwidth of the PicoScope 2204A is actually 10MHz).

If you need to stream the data faster from the PicoScope then you need to consider using the Software Development Kit (SDK) instead of PicoScope 6 (which limits the maximum streaming rate of the PicoScope to 1MS/s). By using the SDK, you can successfully implement some of the applications you have been discussing on this forum, including the Real-time video decoding, as mentioned by bennog on the previous post, and Real-time broadband radio capture as mentioned here: topic10231.html, although you may get better results from amplifying the signal received directly from an antenna. Once captured you could look at the frequency content to see what reception frequencies you've received.

If you wanted to listen to the broadcast content (radio reception) you would clearly need to first tune to the specific carrier frequency that you're interested in (using a tuning circuit), and demodulate the captured signal (which you could do in the SDK), although you may not be able to do this in real-time.

Regards,

Gerry
Gerry
Technical Specialist

Ben321
User
User
Posts: 5
Joined: Tue May 23, 2017 8:09 am

Re: Antialiasing filter strength?

Post by Ben321 » Mon Mar 12, 2018 7:06 pm

Gerry wrote:Hi Ben321,

You may be placing to much emphasis on aliasing.

As explained why, on a previous post, our PicoScopes don't have anti-aliasing filters. So, if you don't want aliasing you need to limit the frequency content of your signal and/or provide a large enough frequency range between the non-attenuated and fully attenuated signals that you will be capturing.

Bandwidth is a function of the input, not the digitizer (ADC) or sample rate. So, if you are not managing the gap between your signal bandwidth and sample rate (i.e. 1:5 ratio), for your intended application, by either limiting the high frequency content of your signal, or selecting a high enough sample rate (by sampling fast enough or using a faster PicoScope), your pass-band can indeed be higher than 500kHz at a 1MS/s sample rate (the bandwidth of the PicoScope 2204A is actually 10MHz).

If you need to stream the data faster from the PicoScope then you need to consider using the Software Development Kit (SDK) instead of PicoScope 6 (which limits the maximum streaming rate of the PicoScope to 1MS/s). By using the SDK, you can successfully implement some of the applications you have been discussing on this forum, including the Real-time video decoding, as mentioned by bennog on the previous post, and Real-time broadband radio capture as mentioned here: topic10231.html, although you may get better results from amplifying the signal received directly from an antenna. Once captured you could look at the frequency content to see what reception frequencies you've received.

If you wanted to listen to the broadcast content (radio reception) you would clearly need to first tune to the specific carrier frequency that you're interested in (using a tuning circuit), and demodulate the captured signal (which you could do in the SDK), although you may not be able to do this in real-time.

Regards,

Gerry
Thanks for the info, however the Picoscope 2204A is not capable of doing any USB streaming faster than 1MSPS, even with the SDK. I specifically stated that the device I was using was the 2204A, so your statement "If you need to stream the data faster from the PicoScope then you need to consider using the Software Development Kit" does not apply to me, in terms of getting a faster streaming rate. And yes, I am using the SDK, but only to get access to the raw samples, not to boost the streaming sample rate.

My info on the maximum possible sample rate for the Picoscope 2204A is from 2 pages on your website. One page is https://www.picotech.com/library/oscill ... aming-mode which has a table of maximum streaming sample rates possible using the SDK with different scopes. The maximum possible sample rate for the Picoscope 2204A is shown as being 1MSPS. The other page is the specs page for the Picoscope 2000 series https://www.picotech.com/oscilloscope/2 ... ifications which clearly shows that even when using the SDK, the Picoscope 2204A can't run at a sample rate greater than 1MSPS.

Are you suggesting that the actual specs of the 2204A have changed (due to either new firmware or hardware design changes) since the publication of the specs on this website?

Gerry
PICO STAFF
PICO STAFF
Posts: 261
Joined: Mon Aug 11, 2014 11:14 am

Re: Antialiasing filter strength?

Post by Gerry » Tue Mar 13, 2018 11:16 am

Hi Ben321,

No the specs haven't changed.

It is possible to set the sample rate to 6.25MS/s in streaming mode with the API for the 2204A. The issue is that, as it only has a 8kS buffer any interruption of the USB stream will not be buffered for long enough, so you would then get gaps in your data. So with an uninterrupted USB stream you can get some contiguous data.

As this is something that is a bit more hit and miss we don't publish it in our specifications. But for your specific application it is certainly something worth persevering with.

Regards,

Gerry
Gerry
Technical Specialist

Ben321
User
User
Posts: 5
Joined: Tue May 23, 2017 8:09 am

Re: Antialiasing filter strength?

Post by Ben321 » Tue Mar 13, 2018 9:57 pm

Ben321 wrote:
Gerry wrote:Hi Ben321,

You may be placing to much emphasis on aliasing.

As explained why, on a previous post, our PicoScopes don't have anti-aliasing filters. So, if you don't want aliasing you need to limit the frequency content of your signal and/or provide a large enough frequency range between the non-attenuated and fully attenuated signals that you will be capturing.

Bandwidth is a function of the input, not the digitizer (ADC) or sample rate. So, if you are not managing the gap between your signal bandwidth and sample rate (i.e. 1:5 ratio), for your intended application, by either limiting the high frequency content of your signal, or selecting a high enough sample rate (by sampling fast enough or using a faster PicoScope), your pass-band can indeed be higher than 500kHz at a 1MS/s sample rate (the bandwidth of the PicoScope 2204A is actually 10MHz).

If you need to stream the data faster from the PicoScope then you need to consider using the Software Development Kit (SDK) instead of PicoScope 6 (which limits the maximum streaming rate of the PicoScope to 1MS/s). By using the SDK, you can successfully implement some of the applications you have been discussing on this forum, including the Real-time video decoding, as mentioned by bennog on the previous post, and Real-time broadband radio capture as mentioned here: topic10231.html, although you may get better results from amplifying the signal received directly from an antenna. Once captured you could look at the frequency content to see what reception frequencies you've received.

If you wanted to listen to the broadcast content (radio reception) you would clearly need to first tune to the specific carrier frequency that you're interested in (using a tuning circuit), and demodulate the captured signal (which you could do in the SDK), although you may not be able to do this in real-time.

Regards,

Gerry
Thanks for the info, however the Picoscope 2204A is not capable of doing any USB streaming faster than 1MSPS, even with the SDK. I specifically stated that the device I was using was the 2204A, so your statement "If you need to stream the data faster from the PicoScope then you need to consider using the Software Development Kit" does not apply to me, in terms of getting a faster streaming rate. And yes, I am using the SDK, but only to get access to the raw samples, not to boost the streaming sample rate.

My info on the maximum possible sample rate for the Picoscope 2204A is from 2 pages on your website. One page is https://www.picotech.com/library/oscill ... aming-mode which has a table of maximum streaming sample rates possible using the SDK with different scopes. The maximum possible sample rate for the Picoscope 2204A is shown as being 1MSPS. The other page is the specs page for the Picoscope 2000 series https://www.picotech.com/oscilloscope/2 ... ifications which clearly shows that even when using the SDK, the Picoscope 2204A can't run at a sample rate greater than 1MSPS.

Are you suggesting that the actual specs of the 2204A have changed (due to either new firmware or hardware design changes) since the publication of the specs on this website?
Thanks for the info. That is very useful. It would be nice if a table of absolute maximum sample rates like this, for each scope you produce, would be posted on your website. That way, when buying a new scope I would have a better idea of each scope's maximum capability. I do at some point plan to buy an even better scope than the 2204A, and knowing each scope's absolute maximum streaming frame rate would be very useful, in helping me decide what scope to get next.

Gerry
PICO STAFF
PICO STAFF
Posts: 261
Joined: Mon Aug 11, 2014 11:14 am

Re: Antialiasing filter strength?

Post by Gerry » Mon Mar 19, 2018 10:14 am

Hi Ben321,

This scenario only exists for the 2204A and 2205A, and as I mentioned it opens the door to unreliable operation for applications that need consistent streaming with no interruptions (which is most applications). So it's not something we want to generally recommend.

Regards,

Gerry
Gerry
Technical Specialist

Post Reply