interfacing an ADC (LTC1408-12) with PC

You said RPi can only SPI 20K bytes per sec. 6 chains would only be able to transfer 2 to 3 Ksps each. That's below audio quality.

I made a huge mistake, i totally forgot that, sorry for that. I guess i thought i can do every thing with it and face no problems and some how im still driven by that thought.

OK, i saved you $8.27 plus shipping.

Try this:

You can sample 6 channels at 1Msps and still do 2 channels of PWM for your servo motors. Only $4.09.

This one is 128K Flash/32K ram. See if you can get the basic working before upgrading to 2M Flash/512K ram version.

I was looking at the 14 bit version which is more like $15. To me that is a pricey chip.

You may be better off with multiple chips rather than one, it depends on whether the interface can get all the data from all six channels in one transfer or not. But it sounds like the rPi is not a good choice for this reason. I'm also not sure about your ADC. I just feel you are putting the cart before the horse. Do you even know your requirements for number of bits and SN, etc.?

Stepper motors usually need a driver circuit. How much current does this motor use? I seriously doubt you will find a MCU that can drive it directly.

I'm not sure why that is not an option and 1 ADC with the rPi is.

I would not order anything until I knew what my algorithm required of the hardware. There are pieces of hardware missing from your design, notably the stepper driving circuit. There are parts you don't know how to integrate, the ADC. You can't pick an MCU until you have some idea of the computation rates you will need.

I think you need to figure out a way to get some raw data and spend some time on a PC massaging an algorithm. Maybe synthesize some data by writing a program. That will let you size the problem. Then you will have an idea of how fast a processor you need and can try to find one that has support for your I/Os. Someone else suggested the Beagle Bone Black which may be a good choice, but you won't know until you understand your problem better.

--

Rick

He is using a servo motor, not a stepper. A servo motor usually have build-in driver, you just need to send a PWM signal to enable the driver. All you need is a digital PWM pin to signal it.

Do you want to expend your time and expertise in reinventing a wheel, or do you want to go buy it?

If you can't do what you want with a sound card (or cards), you need a data acquisition card with sufficient channels and sampling rate. They're out there. Start with DAQ, then look at National Instruments, then use those search terms to Google for it.

Be warned that National Instruments is THE NAME in DAQ hardware and software, so they demand a premium that's really only worth it if you're doing some big LabView application in an environment where time is big money. You can get equivalent stuff for a lot less money elsewhere.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

e i have used it before.

s the power to do what you want.

microphones

es to be used with those specifications

hard to me since its my first and im not experienced with hardware at all. I want to do some processing (Generalized cross-correlation with phase tran sform) on those signals gathered by the mics in order to obtain the time de lay between each pair of mics and find the direction of sound arrival in 3D and steer a camera to that direction.

do you want to go buy it?

ta acquisition card with sufficient channels and sampling rate. They're ou t there. Start with DAQ, then look at National Instruments, then use those search terms to Google for it.

But he also need to control two servo motors and close the loop. Putting P C in the loop is a big big problem.

are, so they demand a premium that's really only worth it if you're doing s ome big LabView application in an environment where time is big money. You can get equivalent stuff for a lot less money elsewhere.

He just need a $4 to $6 micro. Do a proof of concept with a $4 128K/32K (f lash/sram) and test it off-line on the PC. Then do the real-time version w ith 2M/512K chip, which is more involved in PCB constructions.

For the 128K/32K PIC32MX1, all you need is this to hook it up to the progra mmer:

http://173.224.223.62/motor/mx1.jpg

Yes, the more the better. I'll try 12 mikes.

The MAX5566 chip they are using is very difficult to get (min 2500 perhaps). I just brought 50 DIP-8 LM386 (a very popular pre-amp) for $6. So, 12 of them cost me $1.30.

So, total project cost so far: $4 (micro) + $1.3 (pre-amp) + $5? (dumb mikes).

Opps, my bad. Posting late again.

What voltage is the motor? How much current. Just saying PWM is not much of a spec.

--

Rick

That is a silly way to put it. An MCU needs to be on a board with support circuitry including whatever is needed to interface to the various I/Os. Then there is the need for whatever is required to program and debug the MCU.

Ok

--

Rick

his motor use? I seriously doubt you will find a MCU that can drive it dir ectly.

uild-in driver, you just need to send a PWM signal to enable the driver. A ll you need is a digital PWM pin to signal it.

ch of a spec.

I have seen 6V 500mA servo, with logic PWM control (3.3V, a few mA) of 1.5m s neutral pulse. Turn to the left if pulse is less than 1.5ms. Turn to th e right if pulse is more than 1.5ms. Range is 180 degree. Will try to loo k up actual current and PWM frequency cycle. But it depends on specific mo del anyway.

ftware, so they demand a premium that's really only worth it if you're doin g some big LabView application in an environment where time is big money. You can get equivalent stuff for a lot less money elsewhere.

K (flash/sram) and test it off-line on the PC. Then do the real-time versi on with 2M/512K chip, which is more involved in PCB constructions.

t circuitry including whatever is needed to interface to the various I/Os. Then there is the need for whatever is required to program and debug the M CU.

Yes, but the programmer can be reused. And cost nothing if you already hav e it. I just need to find mine somewhere in storage.

ogrammer:

I knew someone would bring that up.

I can hand-wire (no PCB) a MX1 ($4.09), 12 LM386 ($1.30) and 12 Electret Co ndenser PCB Inserts MIC Unit Microphone 6050 $($5.38).

Actual cost of less than $20.

I can run a 12 dimensions convolution at 83Ks/s. Might not be able to do r eal time with 128K/32K. So, extract the data and test it on PC first.

software, so they demand a premium that's really only worth it if you're do ing some big LabView application in an environment where time is big money. You can get equivalent stuff for a lot less money elsewhere.

32K (flash/sram) and test it off-line on the PC. Then do the real-time ver sion with 2M/512K chip, which is more involved in PCB constructions.

ort circuitry including whatever is needed to interface to the various I/Os . Then there is the need for whatever is required to program and debug the MCU.

ave it. I just need to find mine somewhere in storage.

programmer:

Condenser PCB Inserts MIC Unit Microphone 6050 $($5.38).

plus $4.40 for two servo motors:

I asked the seller for detail spec and will post the response here if any.

with 128K/32K. So, extract the data and test it on PC first.

I will use this to detect intruder and fire the AK-47 if necessary. Please stay away from my property.

Well, good for you. If the OP could do that he already would have and he wouldn't be asking the questions he's asking.

I wasn't making suggestions that would work for YOU, I was making suggestions that might work for HIM.

And, if he wants to spend a huge amount of time learning how to design circuits, write code, prototype, etc., all the while pushing his real goal back -- sure, he can take your suggestion.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

/32K (flash/sram) and test it off-line on the PC. Then do the real-time ve rsion with 2M/512K chip, which is more involved in PCB constructions.

port circuitry including whatever is needed to interface to the various I/O s. Then there is the need for whatever is required to program and debug th e MCU.

have it. I just need to find mine somewhere in storage.

programmer:

t Condenser PCB Inserts MIC Unit Microphone 6050 $($5.38).

do real time with 128K/32K. So, extract the data and test it on PC first.

wouldn't be asking the questions he's asking.

ions that might work for HIM.

rcuits, write code, prototype, etc., all the while pushing his real goal ba ck -- sure, he can take your suggestion.

This is doable for the OP as well. I'll reduce it to eight, because there are eight points in a typical umbrella frame. He needs to put the mike apa rt but fixed known location; for example: 8 tip of an umbrella. He also ne ed to put the pre-amp close to the mike. So, two wires between the mike an d LM386, and 10 wires (power, ground and 8 signals) back to the controller board. Might need a few passive (R & C) to configure the LM386. He can att ach them directly to the DIP-8, no PCB necessary. He can do all the wiring s in a day or two.

Regarding the MX1, he can write a very simple setup program (10 to 20 lines ) to setup the ADC array and hunt for signals. Just check the memory buffe r and stop the ADC if any large signals are detected. For initial testing, he can just use the programmer to download the sram content to the PC and test his theory.

lM386 in a power amp, not pre-amp. (not much power, ballpark 1W depending on which variety you have) you might be able to use it as a pre-amp if it has enough gain.

--
umop apisdn 


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

Perhaps using two stages, just enough to bring the output voltage to around 1V for the ADC, and better S/N for the long wires. The microphones will probably be several feet away from the controller.