Timeline for How can I collect data from a hydrophone?
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 30, 2016 at 13:05 | vote | accept | Tyler Durden | ||
| Feb 14, 2016 at 18:52 | history | bounty ended | Tyler Durden | ||
| Feb 12, 2016 at 18:22 | comment | added | Olin Lathrop | @pjc50: Yes, I was being pessimistic with the power requirements. However, fast sampling and the cycles to convolve with a sinc and then dessimate will eat up some, and so will the flash memory writes. I wouldn't want to guess too low without actually doing the design and adding up the worst case. | |
| Feb 12, 2016 at 17:01 | comment | added | pjc50 | My current favourite low power microcontroller: ti.com/lsds/ti/microcontrollers_16-bit_32-bit/msp/… 45microamps per megahertz! Note that going to discrete digital logic will be much worse on power consumption due to increased leakage, greater power required to signal between ICs, etc. Olin's 100ma seems rather high; I'd expect more like 10ma, and potentially much less. | |
| Feb 11, 2016 at 21:16 | comment | added | Olin Lathrop | @Tyler: I just read your updated question. The frequency range is roughly what I was assuming, but you still haven't said anything about signal to noise ratio. The electronic parts shouldn't cost anywhere near $1500. They shouldn't even cost $50. | |
| Feb 11, 2016 at 21:12 | comment | added | Olin Lathrop | @Tyler: Without a micro you won't be easily able to oversample, low pass filter, and decimate. You still need some logic to manage the A/D and the writing to flash memory, which might be over SPI. All the little control logic and sequencing will add up. A micro will definitely be physically smaller, and probably only a little more power, but allow a lot more useful things to be done. | |
| Feb 11, 2016 at 20:01 | comment | added | Robherc KV5ROB | Here's an incredibly low-power MCU that can do everything suggested in this answer & had a 12-bit ADC capable of sampling @ up to 1MHz for anti-aliasing even up to 500KHz noise signals (which your piezos are highly unlikely to be able to pick up that high, anywise): digikey.com/product-detail/en/EFM32LG840F64G-E-QFN64/… | |
| Feb 11, 2016 at 19:15 | comment | added | user1582568 | microcontrollers can be very low power. | |
| Feb 11, 2016 at 19:09 | comment | added | Tyler Durden | Thanks, very valuable analysis. I updated my answer with budget and signal frequency information. I was thinking that by using a direct-to-memory solution and avoiding a computer I could sidestep the power problem. Your suggestion is to use a microcontroller which would have a lower requirement than a computer or miniboard. I wonder if there is a configuration using discrete components that can get the data from hydrophone output to memory chip without using a full blown microcontroller (which has a power-sucking CPU on it). | |
| Feb 10, 2016 at 0:21 | history | edited | Scott Seidman | CC BY-SA 3.0 |
deleted 1 character in body
|
| Feb 9, 2016 at 23:01 | history | edited | Olin Lathrop | CC BY-SA 3.0 |
added 47 characters in body
|
| Feb 9, 2016 at 22:52 | history | answered | Olin Lathrop | CC BY-SA 3.0 |