| ADC questions [message #338607] |
Fri, 02 May 2008 15:40  |
 georgewalker
Messages: 1
Registered: May 2008 |
Welcome |
|
|
Bruno-
Its pretty wonderful to have a brilliant mind in audio so accessible to the masses! I have a couple of questions that have come out of my musings re ADC.
1) I gathered from your tech documents on the G.A. site that a) even low levels of jitter are audible and b) jitter noise increases with wavelength. I was wondering, with all the developments in wideband rf and sdr, if doing frequency modulation, digitizing, and then demodulating in the digital realm could improve AD performance by a) shifting low frequency signals to higher freqs, where jitter is a more managable issue, b) by using more of an ADC chip's bandwidth, c) allowing for the use of higher-bandwidth ADCs (ie those intended for RF) and/or d) allowing for jitter detection/correction by slapping an RF modulated signal with a predictable (although, I expect, not constant wrt time) delay on top of the unmodulated signal.
2) It appears to me that there are an awful lot of gain stages between a typical microphone and ADC. For instance, I was looking into building a field recorder with the TI PGA2500 --> PCM4222, and, upon corresponding with TI, learned that a high-dollar, high power consumption buffer is recommended in between the two, and that driving the latter with the former directly is a VBI. This surprised me. Do you think that there exists microphone preamp topolog(y/ies) that can drive s.c. inputs well? I was wondering in particular about your balanced opamp topologies.
(I gathered from the Boyk paper on amplifiers and negative feedback that every time it is applied, prior distortion is shifted to a higher order. Is that correct? I would think that, assuming that is correct, that one global feedback loop (with gobs of feedback) would be ideal for, say, getting a signal from a ribbon mic output to the input of an ADC...but I haven't seen it done, which leads me to believe that I'm missing some things.
The impetus for my second question is that I would like to get a ribbon mic output into the digital realm with extremely low distortion, ideally without using stupendous amounts of (LiFePO4 supplied) power.
Edit: Oops, I read the "about this forum..." sticky after first posting. My b.
My name is George, I'm a student in biochemistry (wishing I had gone EE or CPE...) at VPISU in Blacksburg, VA obsessed with capturing music without the spacial distortions imposed by studio (or even utility power) bound equipment => field recording.
|
|
|
| Re: ADC questions [message #339001 is a reply to message #338607 ] |
Mon, 05 May 2008 04:06  |
Bruno Putzeys
Messages: 631
Registered: November 2006
Location: Rotselaar (Belgium) |
Gold Member |
|
|
1. Jitter on the IF ADC clock would now come in twice. Once demodulated directly, once as phase modulation of the actual signal. Since your suggestion was an attempt to get rid of the latter, I wouldn't bother.
2. One of the original impulses behind my discrete design was to get rid of the input buffer. However, I'll end up having to add something in the next version because there's no way of making a practical gain trim circuit without a buffer. Anyhow I've become wiser since, and it turns out that completely transparent gain stages are quite feasible.
IC ADC's are all switched cap and indeed need a buffer. Not sure what you mean by "high dollar" though. You can knock together a very good ADC input buffer with a few quid's worth of parts*. The best circuit is a 2nd order LPF where one of the poles is the capacitor across the ADC inputs. An OPA1632 balanced opamp (or a discrete equivalent) is a good candidate. If current consumption is a factor, rule out any of my circuits...
Feedback loops have a tendency to create higher order harmonics that weren't there before. This is not a case of "shifting distortion to higher order". There is definitely a net improvement in THD. The only thing is that with low amounts of loop gain, the improvement in low-order harmonics does less good than the new higher order ones do harm.
The best answer is simply to have so much loop gain that the new harmonics are also reduced beyond significance. A global loop is a good tactic especially if you manage to make it a higher order loop.
___________
* 1 quid = 1.97 buck at the time of writing. Looks like the buck is indeed on the verge of stopping somewhere.
Have more faith, Bambi. The only right way of dealing with headlights is staring them down.
|
|
|
Recording Engineering & Production: The nuts and bolts of making records better, not just louder and cheaper. Some of the brightest minds from the creative and scientific world of audio discuss the elements of audio production that interest you most.
Members
Search
FAQ
Register
Login
Home
