Replies to This Discussion

Permalink Reply by Andrus Kangro on November 10, 2009 at 11:25pm

Hi, MPXV5004DP zero pressure voltage is close to zero, so you can use it as is, assuming that you are not reaching speeds where sensor output voltage exceeds 3.3V. MPX4115 voltage, however, is close to VCC (5V), so I use resistor bridge to bring it down (I use 10k and 20k). I think that simpler is solution is better.

• ▶ Reply

Permalink Reply by Faisal Shah on November 10, 2009 at 11:41pm

But, what if I were to use a simple RC filter on the output to clean up the signal? Can I use a simple voltage divider after the filter, or do I need to buffer it with an op amp first (as illustrated in the application note)?

Attachments:

• pres_sensor_outputfilter_opampbuffer.PNG, 12 KB

• ▶ Reply

Permalink Reply by Andrus Kangro on November 11, 2009 at 12:01am

Hi, all my ADC inputs have low-pass RC filter in front of them. I used to use 10k resistor and 0.1uf capacitor, but will probably use 20k and 1uf for next board, as I find no reason for being aware of movements that are faster than servos can move. This 10k-20k bridge has capacitor as well, in parallel to 20k resistor. I'm not against using opamps, but any component tends to add noise and drifts with temperature, therefore I'd avoid them if its possibe to do without.

Why to run CPU on 3.3V
1. thermal resons. 5V dsPIC running close to half of its spec frequency gets noticeably warm.
2. interfaces with gps - EM-406 needs buffers when talking to PIC TTL serial port, its "high" voltage is not enough. With 3.3V PIC you can wire it directly
3. interface with RX - some modern RX'es output voltage high level is close to 2.7V, this will barely function with PIC CN inputs and will not funtion with IC inputs. But, it may be a good idea to have a buffer before RX anyway.
4. connection with SD-card - SD-cards run at 3.3V. If pic is also running on 3.3V then you can connect directly, otherwise you need to use buffers or resistor bridge.
5. Some gyros run at 3V, letting them talk to 5V adc will decrease prescision.

• ▶ Reply

ModeratorPermalink Reply by Howard Gordon on November 11, 2009 at 1:10pm

Only issue with interfacing 5v devices to 3.3v processors is where the 3.3v processor has to send data to the 5v device (e.g. with a UART interface). Otherwise, a simple resistor bridge as a voltage divider will scale your incoming voltages to the right range. Note that you can directly drive servos with 3.3v logic, even when the servo is powered at 5V. Also, most micro servos run fine from a 3.3v supply.

• ▶ Reply

ModeratorPermalink Reply by Krzysztof Bosak on November 11, 2009 at 12:45pm

'Hi, MPXV5004DP zero pressure voltage is close to zero'
1V is close?

• ▶ Reply

Permalink Reply by Andrus Kangro on November 11, 2009 at 1:53pm

Hi, if I remember correctly, then airspeed is well over 100m/s before MPXV5004DP output reaches 3.3V. 100m/s is about 360km/h, so we should be safe until going over about 400km/h. F5 Open Aeroplane Electrical Motor Rechargeable Sources speed record is 420 km/h (or at least, was quite recently), we should be safe.

• ▶ Reply

ModeratorPermalink Reply by Krzysztof Bosak on November 11, 2009 at 2:16pm

my both MPXV are 1V at stady state (0km/h), as confirmed in manual.

• ▶ Reply

Permalink Reply by Andrus Kangro on November 12, 2009 at 12:38am

you are correct, Krzysztof, mpxv5004dp offset is 1v. I did mean by saying that its close to zero that its, well, closer to zero than to vcc:-)

• ▶ Reply

Permalink Reply by Faisal Shah on November 11, 2009 at 2:31pm

I think I've decided to use an instrumentation amplifier op-amp to interface the 5V pressure and differential pressure sensors to my 3.3V uC. I've been reading up on ADCs, precision measurements, etc ... and it seems that is the best way to go. Present the sensor with high impedance on one side, and a low impedance driver for the ADC on the other side.

* Additionally, I can adjust the gain of the op-amps as I please. For example, I would want to adjust the gain of the differential pressure sensor such that 0-100mph is full scale on the ADC. I can use a 3 or 3.3V zener diode to limit the voltage just in case.

* For the pressure sensor (altitude), since it's nominal reading is at the top end of the output range, I can use a difference amplifier to remove the DC-offset which is equivalent to the smallest output from it, and then amplify the left-over to some percentage of the ADC's full-scale input.

I believe taking the above measures would increase the quality and resolution of measurements.

Thoughts? Over-kill?

• ▶ Reply

Permalink Reply by Andrus Kangro on November 12, 2009 at 12:35am

Hi, using opamps to zoom into interesting range of sensor output makes good use of them, I guess. You will get actual benefit from added complexity, so no overkill here imo. If I wouldn't be so lazy, then I would probably also shift&amplify interesting range of my mpx4115 output (currently it gives resolution of 2m per adc bit, I would not mind to get it down to 1m or even to 0.5m), but I would not bother with mpxv5004dp, its good as is with my 12-bit adc. I would not do anything with added bits.

• ▶ Reply

ModeratorPermalink Reply by Chris Anderson on November 10, 2009 at 11:42pm

Given that the RC world (servos and Rx) work at 5V, you might want to think about switching to a 5V microcontroller and running the whole board at 5V, like we do.

• ▶ Reply

Permalink Reply by jstoezel on November 25, 2009 at 8:59am

Well there's your approach, and the approach that thousands of other people take.

To start with 5V MCUs are not a viable solution. 5V MCUs also do consume more power at the same frequency than 3.3V MCUs. And this should be considered when developing an autonomous application. Not to mention that most 3.3V MCU usually do integrate more ROM/RAM/peripherals than 5V MCUs.

As shown in the following paper, interfacing 3.3V MCUs to 5V is simple, and it is commonly used:

http://focus.ti.com.cn/cn/lit/an/spra550/spra550.pdf

• ▶ Reply