What's a good, really hgih speed 'soundcard'

Are there any Soundcard-like Systems out there that will do 16, 18, or even 20 bits at 100, 200, or even 500MSs ??

Know all the pieces exist, just need an AC coupled system already running and can be quickly implemented for a one off test.

On a sunny day (Sun, 30 Mar 2014 06:53:19 -0700) it happened RobertMacy wrote in :

100Ms/s and 16 bits wide stereo would be 100000000/(16*2) = 3,125,000 sample rate. Think I have seen 96,000, but that would no longer be audible 'sound' now would it? Niquist, it would go to about 1.5 MHz. Try a an audiophile group :-)

Bats for Humans?

Math right?

I phrased that as "Soundcard like" more to reference operation/performance rather than listening pleasure.

By the way, I once wrote a 'proof' of why one NEEDS at least 1MHz bandwidth in a sound system to adequately recreate what one would expect in an orchestral performance, else you can get into bandwidth/slew rate limiting that 'squashes' statistically/potentially occurring waveforms. But, that's another story, right now I NEED this "souncdard like" system.

On a sunny day (Sun, 30 Mar 2014 07:12:34 -0700) it happened RobertMacy wrote in :

OK, now... you are aware of thw wavelength of 1 MHz in air? Now about that position of you next to your 100% perfect 1MHz tweeters.

You seem to need a digitizer.

I knew I shouldn't have brought it up, but I'm not talking about accurate speakers, nor excellent hearing. I'm talking about subtle effects an astute ear will discern. In a sound reproduction system ALL the pressure wvaeforms from ALL the voices are combined into a single time varying waveform. [that in itself causes some information loss] So in the electronics we now have v(t) and its related v(f) [which everybody thinks only needs to be out to 20kHz, or so], In a way, the ear has 'learned' to respond to v(f) and response does NOT go out very far in frequency. The problem arises in the electronics [which is trying to characterize that myriad of signals from all the 'voices' as a SINGLE v(t) waveform] when suddenly the system cannot preserve 'accurate' v(f) because of its non-linear response to excessive dv/dt in the v(t). The ear WILL hear that!

Do a statistical analysis on the 'expected' time waveforms from multiple voices and you'll drop your jaw at some of the dv/dt (which 'needs' 1MHz bandwidth to accurately recreate] you can get.

Yes, but do NOT need DC.

Need Manufacturer, Model, etc.

On a sunny day (Sun, 30 Mar 2014 07:58:25 -0700) it happened RobertMacy wrote in :

I have heard that sort of argument before. You need a reality check, at that wavelength even the blood pulsing in your head will move it far enough to change the phase and destroy your waveform, everything else being perfect. And now in stereo, the distance between your ears, the distance from each ear to the speaker, the speaker's tweeter, the speaker's midrange, the position of the tweeter relative to the midrange, 'phase'. it will (phase) vary when you chew your chewing gum! You really got to see these things in consort with the rest of the system, you, your ears, included. Else it is just theoretical drivel, although part of it maybe be true, it has no significance. Room reflections, even eliminating that with in-ear plugs, the ear canal at that frequency is not to be ciphered away either.

1 MHz at about 330 m/s 330/100000 = 0.003300 or 3.3 mm!!! Here in this youtube video I do an experiment with reflection at a simple 44 kHz or so: Get a life!

Sorry try that in alt.audiophiles.

Well google is your friend.

No. The amplitudes of component signals have been reduced, during mixing, to prevent clipping. dV/dT of the sum will be no more than dV/dT of any one signal channel before it was attenuated and mixed.

20 KHz at full p-p amplitude is all a sound card needs to handle, because that's all that will be in the source signal.

Can you 'learn' to hear 100 KHz?

Problem is, whatever signal source you have (CD, MP3, vinyl) has been brutally mixed, filtered, equalized, distorted, compressed, echoed, Dolbied, and especially bandlimited by whoever produced it. A decent sound card will faithfully reproduce what's left.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

We have a customer who wants us to build a box that will record and play back signals from a jet engine, stuff like load cells, accelerometers, VR speed pickups. We're thinking, why not a good many-channel sound card and a little industrial PC box? All those DRAM interfaces and hard drives and Ethernet stuff would be done, cheap.

Thing is, he wants DC coupling, and sound cards seem to be AC. We had three ideas:

1. Use one channel to carry the DC (sub 20 Hz, I guess) data for, say, 7 others. FM subcarriers or DSB AM or something.
2. Get a super wideband multichannel sound card and FM the baseband data.
3. On each channel, add a narrowband subcarrier, near the upper bandlimit, that carries the DC component of the signal.

All three ideas would need an external signal processing box, but that wouldn't be too awful.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

I think they are AC only because they have an input coupling capacitor. If you jumper the capacitor out, they will go to DC. At least some will. There are USB soundcards for about $1 on ebay. Grab a few and try it.

tm

Something like this?

jamie

On a sunny day (Sun, 30 Mar 2014 09:28:44 -0700) it happened John Larkin wrote in :

Stay clear of cheap soundcards if you do not use FM and want decent signal to noise, The cheap on board soundcard I am just now using for video editing sounds a rumbling noise coming from the PC supply when playing back. I am sure recording uses the same reference. Some soundcards resample to 48 kHz (Creative for example). FM seem a logical choice but what is the highest frequency you need?

If you design a PCI card with a decent ADC you can make a good product, store digital. Better than to make boxes that depends on somebody else's soudncard design that may change any time. Also how many bits do you need? Cheap video digitizer ADCs (flash) that go to above 30 MHz exists, 8 bits though. FPGA, PCI, could not be that hard, most IP is availabe from the FPGA vendors. Need more specs...

Buy a Rigol scope, reverse engineer the hardware and write your own DAQ interface. Find MCU IOs (possibly hijack the LCD interface pins, you won't be needing it?) and install a USB3.0 interface or whatever.

Or wave some money in front of anyone who has an existing product like this. A lot of money.

Being that there aren't many things that need that much bandwidth......do you care to explain why you [think you] need that much? If it were that important, other people would be doing it, after all.

Example: 30-2000MHz SDRs are cheap and plentiful these days, and there is sophisticated free software to support it. Can't find the modulation you need? Add a script, or write your own driver from scratch. The downside (relative to the present case) is, although the frequency range is huge, the sample rate and therefore bandwidth is relatively low, usually 20MHz or so.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com 

"RobertMacy"  wrote in message  
news:op.xdjfi6oo2cx0wh@ajm... 
> Are there any Soundcard-like Systems out there that will do 16, 18, or  
> even 20 bits at 100, 200, or even 500MSs ?? 
> 
> Know all the pieces exist, just need an AC coupled system already  
> running  and can be quickly implemented for a one off test.

How often are you prepared to do any DC calibration ? Once a second ?

I don't know why he just don't bite the bullet and get a PC based capable scope? A little software engineering will give you exactly what you need to monitor the scope, or use their tools.

In the end, you'll have a tool that can be used for other things.

Most of these scopes have dev tool packages or information needed, available.

Jamie

On a sunny day (Sun, 30 Mar 2014 14:18:09 -0500) it happened "Maynard A. Philbrook Jr." wrote in :

Yes that would work and save a lot of time. I could do it with a Raspberry, 2 FIFO chips, and an ADC. That would fit in the Raspberry box I think, and give you an ethernet interface. The software is simple, basically my DVB-S transmitter in reverse, to do 1 MByte per second at 16 bits wide should be no problem.

16 bit ADC should be just fine. But he also has that Xilinx new FPGA board (cannot remember the name), and could add an ADC to that if it not already has one. Ethernet interface will be around for some time I think.

Yes:

Hermes - A DUC/DDC Transceiver:

Quote: "... Supports Real-Time display of entire spectrum from 0-55MHz (with suitable PC software) ..."

Uses ADC AD9744ARU:

-

Almost ready to use (needs PSU, firmware, box, PC):

/Glenn

Datasheet:

AD9744 (Rev. C) - Analog Devices:

Quote: "...

..."

/Glenn

The playback function will be used more often than record. Sometimes we'll acquire data, but more often it will be acquired by some other gear.

Sound card seems like one way to do it.

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

If you dig hard enough in the rtl sdr world, someone managed to get this thing working in a baseband mode, i.e. bypass the mixer. I think it does a few MHz at 8 bits. I never tried this myself.