For a long time now I've liked the idea of buying a scope and with the improvement in soundcards it seems more and more reasonable to be able to make good quality measurements on a PC based scope, without having to spend a lot on a digital scope for the required FFT/spectral analysis.
Typical measurements required for audio far extend the audible range so, at first sight, the limited bandwidth of most soundcards is one limiting factor of a PC scope.
My idea is to use two ADCs per input - the first of which will be limited by an analogue LP filter and the Nyquist frequency limits as usual, but the second ADC will be fed from a passband signal that resides between the Nyquist frequency and the sampling frequency - this will then be folded down into the usual passband by the sampling theorem and an FFT can subsequently be performed.
Now here come the problems...
1. Firstly the FFT that is performed on the second DAC output will be reflected about Fs/2 and as such will need to be mirrored on the spectral plot - this is simple enough.
2. To maintain fidelity and as there is a finite width transition region of the analogue filters to be used the second DAC must be clocked at a lower frequency, such that when the spectral content of the passband signal is folded down there is no hole in the spectral plot.
3. I'm not an expert by any means, but I suspect it is possible to modulate the folded down spectum of the second ADC back to where it would normally reside and therefore construct a frequency plot that is (almost) double that of a single ADC.
I will be posting some pictures as I'm fairly tired and don't think my explanation is as clear as my thinking at this point, but if anyone has any ideas about this let me know.
(Also, I'm not claiming anything original, just wondering if the idea is passable within the constraints of a soundcard/PC)
Typical measurements required for audio far extend the audible range so, at first sight, the limited bandwidth of most soundcards is one limiting factor of a PC scope.
My idea is to use two ADCs per input - the first of which will be limited by an analogue LP filter and the Nyquist frequency limits as usual, but the second ADC will be fed from a passband signal that resides between the Nyquist frequency and the sampling frequency - this will then be folded down into the usual passband by the sampling theorem and an FFT can subsequently be performed.
Now here come the problems...
1. Firstly the FFT that is performed on the second DAC output will be reflected about Fs/2 and as such will need to be mirrored on the spectral plot - this is simple enough.
2. To maintain fidelity and as there is a finite width transition region of the analogue filters to be used the second DAC must be clocked at a lower frequency, such that when the spectral content of the passband signal is folded down there is no hole in the spectral plot.
3. I'm not an expert by any means, but I suspect it is possible to modulate the folded down spectum of the second ADC back to where it would normally reside and therefore construct a frequency plot that is (almost) double that of a single ADC.
I will be posting some pictures as I'm fairly tired and don't think my explanation is as clear as my thinking at this point, but if anyone has any ideas about this let me know.
(Also, I'm not claiming anything original, just wondering if the idea is passable within the constraints of a soundcard/PC)
As I say, at the moment this is only an idea - there may be many limits that stop this being done with simply a PC soundcard.
I think most soundcards use a fixed (relatively low frequency ~20KHz ish) input filter that could make this a problem?
I'm not sure about the computational complexity (at least not without thinking through the design in more detail), but this shouldn't pose too much of a problem for a decent PC should it?
I think most soundcards use a fixed (relatively low frequency ~20KHz ish) input filter that could make this a problem?
I'm not sure about the computational complexity (at least not without thinking through the design in more detail), but this shouldn't pose too much of a problem for a decent PC should it?
I thought that These products look good. Similar price to your quote, but already done. I think there's a limit to what you can do with a soundcard; I use a soundcard measuring system, and a 100MHz 'scope, but something like those in the link would fill in where these are unsuitable.
I don't mean to discourage you, I have a need for just the thing you are thinking about, but it sounds far too complicated for me.
I don't mean to discourage you, I have a need for just the thing you are thinking about, but it sounds far too complicated for me.
Both of the suggested do indeed have a wider bandwidth, although they only offer 8 and 12 bit resolutions - far below ideal.
My idea behind the digital modulation was to keep the 24bit resolutoin (albeit less in practice) and not have the added expense of a high-resolution, high-bandwidth ADC stage.
I have looked online at a vast number of digital scopes and most tend to be based on Flash ADCs - which are, inherantly, limited in their resolution.
I am curious as to whether or not this digital modulation has been done before in a comercial product (plus I want to create a nice test platform for audio frequency circuits).
My idea behind the digital modulation was to keep the 24bit resolutoin (albeit less in practice) and not have the added expense of a high-resolution, high-bandwidth ADC stage.
I have looked online at a vast number of digital scopes and most tend to be based on Flash ADCs - which are, inherantly, limited in their resolution.
I am curious as to whether or not this digital modulation has been done before in a comercial product (plus I want to create a nice test platform for audio frequency circuits).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.