Hi,
I need an analog to digital converter that runs at 100mhz sample rate, and connects to the PC via usb or something similar. Can anyone suggest a product?
Thanks,
James
100mhz analog to digital converter USB
-
- KVRist
- 48 posts since 8 Aug, 2009 from Shenzhen, China
I assume you mean megahertz and not millihertz. Basically what you are describing is a digital oscilloscope and usually only has 8 bits of resolution. Here is one:
http://www.google.com/products/catalog? ... CEkQ8wIwAQ#
There are more models with higher resolution in a PCI card format:
http://www.signatec.com/products/daq/hi ... pda14.html
Do a Google search for 100 MS/s data acquisition.
http://www.google.com/products/catalog? ... CEkQ8wIwAQ#
There are more models with higher resolution in a PCI card format:
http://www.signatec.com/products/daq/hi ... pda14.html
Do a Google search for 100 MS/s data acquisition.
-
- KVRist
- 48 posts since 8 Aug, 2009 from Shenzhen, China
I have been selling test and measurement equipment like digital oscilloscopes for many years and also a musician with a home recording studio. Now I am the Asia sales manager for Audio Precision, the recognized standard in audio test.
http://ap.com/
Good luck with your search.
http://ap.com/
Good luck with your search.
-
- KVRist
- Topic Starter
- 89 posts since 25 Jan, 2010
dvbssman wrote:I have been selling test and measurement equipment like digital oscilloscopes for many years and also a musician with a home recording studio. Now I am the Asia sales manager for Audio Precision, the recognized standard in audio test.
http://ap.com/
Good luck with your search.
hey atm im waiting on my digitizer to be quoted so that I can buy one, but I was thinking...
What would audio sound like if recorded at 100mhz or higher and played back at 44khz?
I am really interested to hear the results, would you be able to record 30 seconds of audio on the highest sample rate you have in a silent room and send me the raw wav or pcm? I wonder if you would be able to hear the particles in the air vibrate? Would it just sound like noise?
-
- KVRian
- 610 posts since 6 Mar, 2005 from USA
If you resampled sound recorded at 100MHz to 44.1kHz you would just hear "normal" sound since the last step of the resampling process will low pass filter signals higher than 22.05kHz. It would sound lousy because the DAQ is only 8 bits.
If you resampled sound recorded at 100mHz to 44.1kHz you wouldn't hear anything, since the highest sound frequency you could record is 0.05Hz, well below hearing and blocked by a DC coupling capacitor in your mic premap anyway.
If you didn't resample at all, but saved raw data coming from a a microphone connected to a 100MHz DAQ, and then played it back at a 44.1kHz rate, you might hear some very low rumbling, or more likely hear nothing. Sounds at 50MHz would be remapped to 22.5kHz...you'd probably hear none of them though because even if your microphone preamp doesn't have a built-in low pass filter (and it does), the mass inertia of the microphone's diaphragm itself is a low pass filter. But maybe you'd hear something at the very low end as a deep rumble; a 100kHz sound which perhaps on a few high-end mic/preamp combos might get passed would get re-mapped to 100k*44.1k/100M = 44Hz. But no, nothing within orders of magnitudes of thermal noise.
If you resampled sound recorded at 100mHz to 44.1kHz you wouldn't hear anything, since the highest sound frequency you could record is 0.05Hz, well below hearing and blocked by a DC coupling capacitor in your mic premap anyway.
If you didn't resample at all, but saved raw data coming from a a microphone connected to a 100MHz DAQ, and then played it back at a 44.1kHz rate, you might hear some very low rumbling, or more likely hear nothing. Sounds at 50MHz would be remapped to 22.5kHz...you'd probably hear none of them though because even if your microphone preamp doesn't have a built-in low pass filter (and it does), the mass inertia of the microphone's diaphragm itself is a low pass filter. But maybe you'd hear something at the very low end as a deep rumble; a 100kHz sound which perhaps on a few high-end mic/preamp combos might get passed would get re-mapped to 100k*44.1k/100M = 44Hz. But no, nothing within orders of magnitudes of thermal noise.
-
- KVRAF
- 6323 posts since 30 Dec, 2004 from London uk
You would be better off trying 192khz which is pretty much the highest available for audio cards these days.
-
- KVRist
- Topic Starter
- 89 posts since 25 Jan, 2010
Oh my god, just got a quote from TTI (uk):
Aquitek PM66-14 HSA 14 3470.39 3470.39
Gage Octopus CS8240 Compuscope 2894.74 2894.74
Gage Octopus CS8242 Compuscope 3348.68 3348.68
Gage Octopus CS8244 Compuscope 3802.63 3802.63
Gage Octopus CS8245 Compuscope 4256.58 4256.58
Gage Octopus CS8247 Compuscope 4710.53 4710.53
Gage Octopus CS8249 Compuscope 5164.47 5164.47
Basically I need 4 analog inputs that run at 100mhz each.
I can't afford this!! I'm looking at more of a price range around £200 - £300 max!!
Why are they so expencive?! Can anyone reccomend a product that can do 4 analog inputs at 100mhz each around the price range im after?
You know I don't understand these prices because Analog Devices sell ADC chips that run around 100mhz for a few pounds each! How hard would it be for me to buy a few chips and attach them to a breadboard?! But how would I get them to send data up a USB 3.0 connection? I would have to buy some kind of module? I would then need to write a USB 3.0 driver or something?? How does all that work? Where would I even start?
I am an experienced C++ programmer, so if I found a starting point I could make this happen - or atleast do my best. I don't take failure easily.
Aquitek PM66-14 HSA 14 3470.39 3470.39
Gage Octopus CS8240 Compuscope 2894.74 2894.74
Gage Octopus CS8242 Compuscope 3348.68 3348.68
Gage Octopus CS8244 Compuscope 3802.63 3802.63
Gage Octopus CS8245 Compuscope 4256.58 4256.58
Gage Octopus CS8247 Compuscope 4710.53 4710.53
Gage Octopus CS8249 Compuscope 5164.47 5164.47
Basically I need 4 analog inputs that run at 100mhz each.
I can't afford this!! I'm looking at more of a price range around £200 - £300 max!!
Why are they so expencive?! Can anyone reccomend a product that can do 4 analog inputs at 100mhz each around the price range im after?
You know I don't understand these prices because Analog Devices sell ADC chips that run around 100mhz for a few pounds each! How hard would it be for me to buy a few chips and attach them to a breadboard?! But how would I get them to send data up a USB 3.0 connection? I would have to buy some kind of module? I would then need to write a USB 3.0 driver or something?? How does all that work? Where would I even start?
I am an experienced C++ programmer, so if I found a starting point I could make this happen - or atleast do my best. I don't take failure easily.
-
- KVRian
- 610 posts since 6 Mar, 2005 from USA
It's frustrating that's its so high, but the reasons include (and why you shouldn't waste time trying to roll your own):
* It's difficult to do the analog design at high freq's (for instance, you can't use a solderless breadboard to prototype above about 10MHz because the capacitance between adjacent insulated strips shorts out the circuit)
* It's difficult to do the low-level programming (being a C++ guru isn't the same as being able to write efficient kernal-mode drivers).
* It's difficult to do the firmware programming (USB handshaking protocols aren't simple like RS-232, and framing is more complex than Ethernet).
If you really need 4 channels at 100MHz (100mHz means 0.1Hz) then I recommend you purchase PCI plug-in boards for a desktop computer to eliminate the USB stage entirely. NI has several, although four channels will still cost over $5k. I've used PXI versions instead of PCI with success, although those setups are more expensive.
Usually, though, engineering will give you a better solution that doesn't require 4 channels at 100MHz. For instance, if you're looking for 20ns glitches, write the logic in VHDL to put in an FPGA. You can easily do that in one chip, and then generate time stamps for the glitches. Program the FPGA to do its glitch timestamping in ASCII, and feed that into a low speed RS232-to-USB shield, pre-programmed (e.g. on Sparkfun.com). Or, if you need to catch a 10,000 samples of activity on four channels at 100MHz, again use a front end FPGA to start streaming to local RAM after your trigger. Then once the event is over, stream it over a low-speed USB connection (again, like the Sparkfun shields).
If the 100mHz isn't a typo, though, I've used Measurement Computing (http://www.mccdaq.com/) USB-1208 witih success. For 190USD you'll get 8 channels at 12 bits, and it can go as low as 100mHz with no problem.
* It's difficult to do the analog design at high freq's (for instance, you can't use a solderless breadboard to prototype above about 10MHz because the capacitance between adjacent insulated strips shorts out the circuit)
* It's difficult to do the low-level programming (being a C++ guru isn't the same as being able to write efficient kernal-mode drivers).
* It's difficult to do the firmware programming (USB handshaking protocols aren't simple like RS-232, and framing is more complex than Ethernet).
If you really need 4 channels at 100MHz (100mHz means 0.1Hz) then I recommend you purchase PCI plug-in boards for a desktop computer to eliminate the USB stage entirely. NI has several, although four channels will still cost over $5k. I've used PXI versions instead of PCI with success, although those setups are more expensive.
Usually, though, engineering will give you a better solution that doesn't require 4 channels at 100MHz. For instance, if you're looking for 20ns glitches, write the logic in VHDL to put in an FPGA. You can easily do that in one chip, and then generate time stamps for the glitches. Program the FPGA to do its glitch timestamping in ASCII, and feed that into a low speed RS232-to-USB shield, pre-programmed (e.g. on Sparkfun.com). Or, if you need to catch a 10,000 samples of activity on four channels at 100MHz, again use a front end FPGA to start streaming to local RAM after your trigger. Then once the event is over, stream it over a low-speed USB connection (again, like the Sparkfun shields).
If the 100mHz isn't a typo, though, I've used Measurement Computing (http://www.mccdaq.com/) USB-1208 witih success. For 190USD you'll get 8 channels at 12 bits, and it can go as low as 100mHz with no problem.
-
- KVRian
- 610 posts since 6 Mar, 2005 from USA
Everyone's been assuming that by 100mHz you've really meant 100MHz (mHz is 0.001Hz and MHz is 1 000 000 Hz).
But a sampling period of 100ns implies not 100MHz but 10MHz. That's much easier to accomplish!
1/sampling period = sampling frequency.
n = nano = 10^-9
u = micro = 10^-6
m = milli = 10^-3
k = kilo = 10^3
M = Mega = 10^6
You also need to state how many bits of resolution you need. 8 bits for audio, for instance, isn't nearly enough. 16 works for my ears.
You also need to say more about your application. Do you need just to datalog 1000 samples for offline analysis? Then I can recommend very cheap ways if you're willing to break out the soldering iron. Or do you need to do real-time analysis with it?
And is KVR the right place to be asking about high-speed data acquisition? You'll probably do better talking to DAQ specialist, like National Instruments (the other NI).
But a sampling period of 100ns implies not 100MHz but 10MHz. That's much easier to accomplish!
1/sampling period = sampling frequency.
n = nano = 10^-9
u = micro = 10^-6
m = milli = 10^-3
k = kilo = 10^3
M = Mega = 10^6
You also need to state how many bits of resolution you need. 8 bits for audio, for instance, isn't nearly enough. 16 works for my ears.
You also need to say more about your application. Do you need just to datalog 1000 samples for offline analysis? Then I can recommend very cheap ways if you're willing to break out the soldering iron. Or do you need to do real-time analysis with it?
And is KVR the right place to be asking about high-speed data acquisition? You'll probably do better talking to DAQ specialist, like National Instruments (the other NI).
-
- KVRist
- Topic Starter
- 89 posts since 25 Jan, 2010
Ah, thank you I will heck NI out.
Sorry yeah I mean every 10 nanoseconds, 100MHZ, can that rate of data acquisition using these contact microphones:
http://cgi.ebay.co.uk/4-x-Piezo-Contact ... 1c17e545b4
Be achieved cheaply? Would 10MHZ be more realistic? Depending on the price difference I may be happy to settle with less than 100MHZ.
8bits is fine for me, all I want to do is sample contact vibrations for analysis not really playback. The amplitude resolution is not a big deal.
oh yes and real time
This looks ok but still out of my price range tbh:
http://www.mccdaq.com/pci-data-acquisit ... 20-12.aspx
Sorry yeah I mean every 10 nanoseconds, 100MHZ, can that rate of data acquisition using these contact microphones:
http://cgi.ebay.co.uk/4-x-Piezo-Contact ... 1c17e545b4
Be achieved cheaply? Would 10MHZ be more realistic? Depending on the price difference I may be happy to settle with less than 100MHZ.
8bits is fine for me, all I want to do is sample contact vibrations for analysis not really playback. The amplitude resolution is not a big deal.
oh yes and real time
This looks ok but still out of my price range tbh:
http://www.mccdaq.com/pci-data-acquisit ... 20-12.aspx