Measure and log current at least 1000 times per second
- Thread starter hias9
- Start date
DogDipstick
100 kW
You want one log per millisecond? Ok. 1Ks/S. ( kilosamples per second) Thats very slow sampling for some machinery.
DSo...
... or M series DAQ? The M-Series can do a sampling of 1.25 MEGAsamples per second ( 1.25 million per second).... 60 VDC continuous bank isolation, 1,400 Vrms/1,950 VDC channel-to-bus isolation, withstand for 5 sec. 3800$ - 6000$. Used, sell for 1200$ish.
1000+ sampling per second is oscilloscope territory...or a DAQ.
For 100$ you can get a 100kS/sec DAQ datalogger for that voltage. IF you know how to use it.
Example: You would have to design the analog side of the data array: but, this thing is not exactly slow.
https://www.dataq.com/products/di-...IKDVxLWgwyz7xqCAgIE-YvmMVH8KXgTBoCjXIQAvD_BwE
Consumer grade? Well, there is Extech ( a division o FLIR) typical 1 log per second... Or my Flir Bluetooth multimeter, can do 3 sample per second... ... Or the Cycle analysts... My CA3 does 10 samples/sec. This is the fastest current and voltage on a easy spreadsheet that I have... There is some documentation of a use of a faster rate for the system, think I read somewhere? Certainly the clock rate can handle it.. and the CA3 is capable of more I bet... Lol. I am very happy with the 10Hz.
IDK. All the consumer grade thingies like to do 1 log per sec for some reason...
Digital storage oscilloscopes are very useful...
......... Nothing like my Tektronix 422. Nothing like it. That, 15 MHz, top of the line '1965.... ..... Nowadays we have 100 gigaherze.. top of the line oscilloscopes.... Lol. 100 GIGA hertz.... (10^9 Hz, = 1GHz).. (1GHz = 1,000,000,000 Hertz) (100 x 1,000,000,000 (a billion) times per second)... (100 billion times per second, 100 GHz)..... I bet that would leave a little dent in your SD card.
Can I ask why? I, also, am quite intrigued by the idea of capturing and analyzing the data from the output of a motor controller:... and having a recod is quite valuable... ? Have been. You will only average the RMS and not see the peak without an ( faster) oscilloscope? Or something faster than the brush less FOC controller PWM rate ( usually 12-18kHz). ( 12-18000 Hertz)... I have never owned a very fast DSO: But am looking at getting one.. ( Rigol?) Whattya trying to accomplish? A direct measurement of each wave pulse from the controller bridge, before the DC current gets through the diode? I think you would need to have something that can measure faster than the PWM rate f the microchip output (driver) on the (driven) bridge. maybe I don't understand or know enough but this is something that interest me too.
My true RMS clamp meter ( Seaward Solar) can measure harmonics ( very useful... when pwm through an inverter bridge... )(dirty signals and false readings) ( Ghost voltages?) , up to the 25th Order: voltage, current and power, at the push of a button... AC, DC, or AC+DC, power factor; Pretty good, peak and hold too. Does not data log, but my FLIR does.. and that is 3 log / sec. true rms.. on a 1200$ meter. So I think the makers thought there was no reason for a higher sampling rate.Even for a 1000$ + meter? Only 3 sample per second data logged ... The files stay small. Very slow compared to a real scope... 3s/Sec. You ask for 1000 s/Sec.. ( 1Ks/Sec sampling rate, a log every millisecond)... So I think this is DSo or DAQ territory for those that know how to use these systems.
DSo...
... or M series DAQ? The M-Series can do a sampling of 1.25 MEGAsamples per second ( 1.25 million per second).... 60 VDC continuous bank isolation, 1,400 Vrms/1,950 VDC channel-to-bus isolation, withstand for 5 sec. 3800$ - 6000$. Used, sell for 1200$ish.
1000+ sampling per second is oscilloscope territory...or a DAQ.
For 100$ you can get a 100kS/sec DAQ datalogger for that voltage. IF you know how to use it.
Example: You would have to design the analog side of the data array: but, this thing is not exactly slow.
https://www.dataq.com/products/di-...IKDVxLWgwyz7xqCAgIE-YvmMVH8KXgTBoCjXIQAvD_BwE
Consumer grade? Well, there is Extech ( a division o FLIR) typical 1 log per second... Or my Flir Bluetooth multimeter, can do 3 sample per second... ... Or the Cycle analysts... My CA3 does 10 samples/sec. This is the fastest current and voltage on a easy spreadsheet that I have... There is some documentation of a use of a faster rate for the system, think I read somewhere? Certainly the clock rate can handle it.. and the CA3 is capable of more I bet... Lol. I am very happy with the 10Hz.
IDK. All the consumer grade thingies like to do 1 log per sec for some reason...
Digital storage oscilloscopes are very useful...
......... Nothing like my Tektronix 422. Nothing like it. That, 15 MHz, top of the line '1965.... ..... Nowadays we have 100 gigaherze.. top of the line oscilloscopes.... Lol. 100 GIGA hertz.... (10^9 Hz, = 1GHz).. (1GHz = 1,000,000,000 Hertz) (100 x 1,000,000,000 (a billion) times per second)... (100 billion times per second, 100 GHz)..... I bet that would leave a little dent in your SD card.
Can I ask why? I, also, am quite intrigued by the idea of capturing and analyzing the data from the output of a motor controller:... and having a recod is quite valuable... ? Have been. You will only average the RMS and not see the peak without an ( faster) oscilloscope? Or something faster than the brush less FOC controller PWM rate ( usually 12-18kHz). ( 12-18000 Hertz)... I have never owned a very fast DSO: But am looking at getting one.. ( Rigol?) Whattya trying to accomplish? A direct measurement of each wave pulse from the controller bridge, before the DC current gets through the diode? I think you would need to have something that can measure faster than the PWM rate f the microchip output (driver) on the (driven) bridge. maybe I don't understand or know enough but this is something that interest me too.
My true RMS clamp meter ( Seaward Solar) can measure harmonics ( very useful... when pwm through an inverter bridge... )(dirty signals and false readings) ( Ghost voltages?) , up to the 25th Order: voltage, current and power, at the push of a button... AC, DC, or AC+DC, power factor; Pretty good, peak and hold too. Does not data log, but my FLIR does.. and that is 3 log / sec. true rms.. on a 1200$ meter. So I think the makers thought there was no reason for a higher sampling rate.Even for a 1000$ + meter? Only 3 sample per second data logged ... The files stay small. Very slow compared to a real scope... 3s/Sec. You ask for 1000 s/Sec.. ( 1Ks/Sec sampling rate, a log every millisecond)... So I think this is DSo or DAQ territory for those that know how to use these systems.
I also did some research about that yesterday and probably the raspberry pi daq hat with 16bit resolution would be the best choice however I have no experience with raspberry pi.
I want to do some fine-tuning of current regulator and PLL regulator and want to calibrate the current sensors (not really important that they measure the exact value but more important that both current sensors (phase A and C) show equal peak values.
I want to do some fine-tuning of current regulator and PLL regulator and want to calibrate the current sensors (not really important that they measure the exact value but more important that both current sensors (phase A and C) show equal peak values.
hias9 said:
I have no experience with tuning FOS controllers, what does the procedure look like?
You are throwing out little bits of information. Are you interested in peak current? If you have a 3000 PRM motor, and you want to see the peak current at each rotation, you need way more than 1K sample/s. You need 1M sample/s, which would be something like a pocket oscilloscope.
No, I don't need to see peak current of each wave.
100k per second is more than enough. One cycle of the controller is only 13-16kHz.
I want to analyze the waves to fine tune Kp Ki for current regulator and phase-locked loop.
And also make sure that the calibration of the two current sensors inside the controller is fine or if I need to adjust a bit.
There is no real procedere. Riding, analyzing, adjusting.
I dont have a dyno, so I would do very short rides and analyze data later. I guess the raspberry pi can be powered while riding using a DCDC converter.
100k per second is more than enough. One cycle of the controller is only 13-16kHz.
I want to analyze the waves to fine tune Kp Ki for current regulator and phase-locked loop.
And also make sure that the calibration of the two current sensors inside the controller is fine or if I need to adjust a bit.
There is no real procedere. Riding, analyzing, adjusting.
I dont have a dyno, so I would do very short rides and analyze data later. I guess the raspberry pi can be powered while riding using a DCDC converter.
hias9 said:
You also said earlier that only 1 phase needs to be measured. If you need 3 channels, MCC128 can do it. How do you plan to convert current to voltage for the MCC128 to measure it?
Step 1: decide what type of current measurement (min,max,precision,accuracy...)
Step 2: decide on type of current measuring device that satisfies conditions from step 1
Step 3: decide how to convert that to voltage
Step 4: decide what DAC can capture it <<--- you are here (but even step 1 is not defined)
Step 5: decide what device will store data
By using current sensors. The ones I have ordered are open loop and have a delay of <7 microseconds which is okay.
Yes I know, but when I wrote the first post I did not know the MCC products. Only one at a time would have been okay, but measuring all three at the same time makes it easier.
Both the ACC 118 and 128 would be fine, but 128 has much higher resolution, so I will probably go for the 128.
I guess the device to store data could be a microsd card? I have no experience at all with raspberry pi.
Yes I know, but when I wrote the first post I did not know the MCC products. Only one at a time would have been okay, but measuring all three at the same time makes it easier.
Both the ACC 118 and 128 would be fine, but 128 has much higher resolution, so I will probably go for the 128.
I guess the device to store data could be a microsd card? I have no experience at all with raspberry pi.
