I have been using the Teensy 3.6 microcontroller board (180 MHz ARM Cortex-M4 processor) to try and implement a driver for a sensor. The sensor is controlled over SPI and when it is commanded to make a measurement, it sends out the data over two lines, DOUT and PCLK. PCLK is a 5 MHz clock signal and the bits are sent over DOUT, measured on the falling edges of the PCLK signal. The data frame itself consists of 1,024 16-bit values.
My first attempt consisted a relatively naïve approach: I attached an interrupt to the PCLK pin looking for falling edges. When it detects a falling edge, it sets a bool that a new bit is available and sets another bool to the value of the DOUT line. The main loop of the program generates a uint_16 value from these bits and collects 1,024 of these values for the full measurement frame.
However, this program locks up the Teensy almost immediately. From my experiments, it seems to lock up as soon as the interrupt is attached. I believe that the microprocessor is being swamped by interrupts.
I think that the correct way of doing this is by using the Teensy's DMA controller. I have been reading Paul Stoffregen's DMAChannel library but I can't understand it. I need to trigger the DMA measurements from the PCLK digital pin and have it read in bits from the DOUT digital pin. Could someone tell me if I am looking at this problem in the correct way? Am I overlooking something, and what resources should I view to better understand DMA on the Teensy?
Thanks!
I put this on the Software Engineering Stack Exchange because I feel that this is primarily a programming problem, but if it is an EE problem, please feel free to move it to the EE SE.
Is DMA the Correct Way to Receive High-Speed Digital Data on a Microprocessor?
There is more than one source of 'high speed digital data'. DMA is not the globally correct solution for all data, but it can be a solution.
it sends out the data over two lines, DOUT and PCLK. PCLK is a 5 MHz clock signal and the bits are sent over DOUT, measured on the falling edges of the PCLK signal.
I attached an interrupt to the PCLK pin looking for falling edges. When it detects a falling edge, it sets a bool that a new bit is available and sets another bool to the value of the DOUT line.
This approach would be call 'bit bashing'. You are using a CPU to physically measure the pins. It is a worst case solution that I see many experienced developers implement. It will work with any hardware connection. Fortunately, the Kinetis K66 has several peripherals that maybe able to assist you.
Specifically, the FTM, CMP, I2C, SPI and UART modules may be useful. These hardware modules are capable of reducing the work load from processing each bit to groups of bits. For instance, the FTM support a capture mode. The idea is to ignore the PCLK signal and just measure the time between edges. These times will be fixed in a bit period/CLK. If the timer captures a two bit period, then you know that two ones or zeros were sent.
Also, your signal seems like SSI which is an 'digital audio' channel. Unfortunately, the K66 doesn't have an SSI module. Typical I2C is open drain and it always has a start bit and fixed word size. It maybe possible to use this if you have some knowledge of the data and/or can attach some circuit to fake some bits (to be removed later).
You could use the UART and time between characters to capture data. The time will be a run of bits that aren't the start bit. However it looks like this UART module requires stop bits (the SIM feature are probably very limited).
Once you do this, the decision between DMA, interrupt and polling can be made. There is nothing faster than polling if the CPU uses the data. DMA and interrupts are needed if you need to multiplex the CPU with the data transfer. DMA is better if the CPU doesn't need to act on most of the data or the work the CPU is doing is not memory intensive (number crunching). Interrupts depend on your context save overhead. This can be minimized depending on the facilities your main line uses.
Some glue circuitry to adapt the signal to one of the K66 modules could go a long way to making a more efficient solution. If you can't change the signal, another (NXP?) SOC with an SSI module would work well. The NXP modules usually support chaining to an eDMA module as well as interrupts.
Related
I am working on a drone project and currently choosing a board to use. Is it possible to use an Arduino Nano for all needs which are:
Gyroscope and Accelerometer
Barometer (as an altimeter)
Digital magnetometer
WiFi (to send telemetry for processing)
GPS module
4 motors (of course)
P.S:
I know nothing about Arduino. However I have a good ASM, C/C++, programming background and I used to design analog circuits.
I would like to avoid using ready-made flight controllers.
Pin count should not be too much of an issue if using I²C sensors, they would simply all share the same two pins (SCL, SDA).
I agree that the RAM could be a limitation, the processing power (30 MIPS for an arduino uno) should be sufficient.
On an arduino mega, the APM project ran for years with great success.
I believe it's possible to do a very simplified drone flight controller with an Arduino nano and several I²C sensors + GPS.
But even with a more advanced microcontroller it's not a trivial task.
*** If you still want to try the experiment, have a look at openlrs project : https://code.google.com/p/openlrs/ . It's quite old (there are several derived projects too), but it runs on a hardware similar to arduino uno (atmega328). It provides RC control, and quad flight controller with i²c gyroscopes, accelerometers (based on wii remote), and barometer.
It also parse data from the GPS, but afaik it doesn't provide autonomous navigation but it should be possible to add it without too much additional work.
edit : about the available RAM.
I understand that at first sight 2kb of RAM seems a very small amount. And a part of it is already used by Arduino, for example the serial library provides two 64 bytes FIFO, using some RAM. Same for the Wire (I²C) library, although a smaller amount. It also uses some RAM for stack and temporary variables, even for simple tasks such as float operations. Let's say in total it will use 500 bytes.
But then what amount of RAM is really required ?
- It will have a few PIDs regulators, let's say that each one will use 10 float parameters to store PID parameters, current value etc. So it gives 40 bytes per regulator, and let's say we need 10 regulators. We should need less, but let's take that example. So that's 400 bytes.
-Then it will need to parse GPS messages. A GPS message is maximum 80 bytes. Let's allow a buffer of 80 bytes for GPS parsing, even if it would be possible to do most of the parsing "on-the-fly" without storing it in a buffer.
-Let's keep some room for the GPS and sensors data, 300 bytes which seems generous, as we don't need to store them in floats. But we can put in it the current GPS coordinates, altitude, number of satellites, pitch, roll etc
-Then some place for application data, such as home GPS coordinates, current mode, stick positions, servo values etc.
The rest is mostly calculations, going from the current GPS coordinates and target coordinates to a target altitude, heading etc. And then feed the PIDs to the calculated pitch and roll. But this doesn't require additional RAM.
So I would say it's possible to do a very simple flight controller using 1280 bytes. And if I was too low or forgot some aspects, there's still more than 700 bytes available.
Certainly not saying it's easy to do, every aspect will have to be optimized, but it doesn't look impossible.
It would be a trick to make all of that work on a Nano. I would suggest you look at http://ardupilot.com/ they have built a lot of cool thinks around the ARM chip (same as an Arduino) and there are some pretty active communities on there as well.
Even if you didn't run out of pins (and you probably would), by the time you wrote the code for the motors and the GPS, you will run out of RAM.
And that's not even getting into the CPU speed, which is nowhere near enough. As mentioned in the other answer, you'll be better off with a Cortex M-x CPU.
Arguably, you could use a few Nanos, one per task, but chaining them together would be a nice mess...
I have a PIC24 based system equipped with a 24 bit, 8 channels ADC (google MCP3914 Evaluation Board for more details...).
I have got the board to sample all of the 8 channels, store the data in a 512x8 buffer and transmit the data to PC using a USB module when the buffer is full (it's is done by different interrupts).
The only problem is that when the MCU is transmitting data (UART transmission interrupt has higher priority than the ADC reading interrupt) the ADC is not sampling data hence there will be data loss (sample rate is around 500 samples/sec).
Is there any way to prevent this data loss? maybe some multitasking?
Simply transmit the information to the UART register without using interrupts but by polling the bit TXIF
while (PIR1.TXIF == 0);
TXREG = "the data you want to send";
The same applies to the ADC conversion : if you were using interruptions to start / stop a conversion, simply poll the required bits (ADON) and thats it.
The TX bits and AD bits may vary depending on your PIC.
That prevents the MCU to enter an interrupt service routine and loose 3-4 samples.
In PIC24 an interrupt can be assigned one of the 8 priorities. Take a look at the corresponding section in the "Family Reference Manual" -> http://ww1.microchip.com/downloads/en/DeviceDoc/70000600d.pdf
Alternatively you can use DMA channels which are very handy. You can configure your ADC to use the DMA, and thus sampling and feeding the buffer won't use any CPU Time, same goes for UART I beleive.
http://ww1.microchip.com/downloads/en/DeviceDoc/39742A.pdf
http://esca.atomki.hu/PIC24/code_examples/docs/manuallyCreated/Appendix_H_ADC_with_DMA.pdf
What is the difference between SPI and serial? In reading an article talking about inter-processor communications, it states that serial interfaces are being replaced with SPI for better/faster comms? What exactly is the difference?
The word "serial" doesn't mean much. But I'll assume that you are talking about traditional serial communication standards. What's fundamentally different about SPI is that it is synchronous. As opposed to, say, RS-232, an asynchronous signaling standard.
An important property of asynchronous signaling is the baudrate, the frequency at which the bits in a byte are sent. The receiver has to do extra work to recover the clock that was used by the transmitter. A typical UART does so by over-sampling the signal at a rate 16 times the baudrate. The start-bit is important, which synchronizes the over-sampling clock. Delays between bytes can be arbitrary, the receiver re-synchronizes for each individual byte. Problems with this scheme are a mismatch between the transmitter and the receiver clock frequencies and clock jitter, effectively limiting the baudrate.
This is not a problem with SPI, it has an extra signal line that carries the clock signal so that both the transmitter and receiver uses the exact same clock. And is therefore immune from mismatches and jitter, allowing higher transfer rates. No stability requirements at all in the clock frequency, the signals can simply be generated in software. Typical four line wiring looks like this:
SCLK is the clock signal. MOSI and MISO carry the data, SS is a chip select signal. Common ground is assumed. More about it in this Wikipedia article. electronics.stackexchange.com is a good site to ask more questions about it.
The previous answer is somewhat misleading.
SPI and UART both transfer binary data as bytes and/or words, depending on the hardware. As explained above, one is synchronous and one is asynchronous. Both require an extra data line to be bidirectional. ASCII is an agreed upon interpretation of the binary data and is not actually a factor in either.
The first answer is almost correct with some small comments:
1) SPI is a subtype of SSI (another example is RS-422)
2) SPI uses the master/slave concept with CS/SS (chips select, slave select) pin ...
Thus a master can have multiple slaves and select between them using the SS pin. Also, on some chips, using the SS the chip can be switched from master to slave.
SPI is a bidirectional data protocol. The difference is that SPI uses an exchange of binary data. And UART uses ASCII, making it much slower data transfer
I have an Arduino Mega connected to a 6 axis robotic arm. All 6 interrupts are attached to encoders (one encoder pin on an interrupt, the other on a vanilla digital input). The interrupts are handled with this code:
void readEncoder1(){
//encoders is a 2d array, where the first d is the axis, and the two pin numbers
//first pin is on an interrupt (CHANGE), and second is a standard digital in
if (digitalRead(encoders[0][0]) == digitalRead(encoders[0][1])) {
positions[0]++;
} else {
positions[0]--;
}
if(servoEnable){
updatePositions(); //// compares positions[] to targets[] and adjusts motor speed accordingly
}
}
This is designed to keep the arm locked at a certain position- if the arduino detects that the position of the motor is off by a certain threshold, it updates the power going to the motor to keep the arm in position.
The problem is this, then -- if two or three (or more) axis are under load (requiring constant updating to stay in position) or they are moving, the Arduino will stop receiving intact commands on Serial input, several characters will be dropped. The interrupts are obviously running quite quickly, and for some reason this is causing commands to become corrupted. Is there any way around this? Architecturally, am I doing this right? My main instinct is to call updatePositions() in the main run loop at, say, 100 ms intervals, will this significantly reduce interrupt overhead? I guess what my question boils down to is how do I get reliable serial commands into the Arduino even if all 6 encoders are pulsing away?
Quadrature encoders were designed to be read by hardware counters. Pulse rates are generally high with the motor running at full speed. One megahertz is not unusual. The higher the number of pulses, the better the servo loop works and the more accurate you can position the motor.
Doing this is in software with a low-power cpu is, well, challenging. It will fall apart when the ISR takes longer than the interval between pulses. You'll lose pulses and thus position. Especially bad because there is no way you can detect this error condition. And that this loss happens when the robot is moving fast, the worst case condition to lose control.
You absolutely cannot afford to update the servo loop in the interrupt handler so get rid of that first. Keep the ISR to the bare minimum, only count the position and nothing else. The servo loop should be separate, driven by a timer interrupt or tick. You cannot properly control a robot with a 100 msec servo update unless it is big an sluggish, this needs to be a handful of milliseconds at most to get smooth acceleration and stable feedback.
There's a limited amount of wisdom in spending forty bucks to control thousands of dollars worth of robot hardware. Not being able to keep up in the servo loop is something you can detect, shut it down when the position error builds up too much. There's nothing you can do about losing pulses, that's a wreck. Get the hardware counters.
First rule of embedded systems:
Do as little as possible in interrupts.
In your case, just update the positions in the interrupt and run your position/speed control loop in the background or at a lower priority.
Aside: I assume you are aware that you are "losing" encoder pulses as you don't have an interrupt on one of the channels?
Also, interrupt-driven encoder-analysis is very noise-prone. If you get a noise pulse, you'll likely only see an interrupt for one of the edges as they'll be too close together to process both.
A more robust way is to use a state machine which watches all 4 transitions, but that requires either interrupts on both edges of both channels, or polling fast enough to not miss anything up the to rate you are expecting to see.
I'm new to microcontroller programming and I have interfaced my microcontroller board to another device that provides a status based on the command send to it but, this status is provided on the same I/O pin that is used to provide data. So basically, I have an 8-bit data line that is used as an output from the microcontroller, but for certain commands I get a status back on one of the data lines if I choose to read it. So I would be required to change the direction of this one line to read the status thus converting this line as an ouput to an input and then back to an output. Is this acceptable programming or will this changing of the I/O pin this frequently cause instability?
Thanks.
There should not be any problem with changing the direction of the I/O line to read the status returned by the peripheral provided that you change the state of the line to an input before the peripheral starts to drive the line and then do not try to drive the line as an output until the peripheral stops driving it. What you must try to avoid is contention between the two driver devices, i.e. having the two ends being driven to opposite states by the processor and peripheral. This would result in, at best a large spike in the power consumption or worse blown pin driver circuitry in the processor, peripheral or both.
You do not say what the processor or peripheral are so I cannot tell whether there are any control bits in the interface that enable the remote device to output the status so that you can know whether the peripheral is driving the line at any time.
I've done this on digital I/O pins without any problems but I'm very far from an expert on this. It probably depends entirely on which microcontroller you are using though.
Yes, it's perfectly fine to change I/O direction on microcontroller repeatedly.
That's the standard method of communicating over open-collector buses such as I2C and the iButton. (see PICList: busses for links to assembly-language code examples).
transmit 0 bit: set output LATx bit to 0, and then set TRISx bit to OUTPUT.
transmit 1 bit: keep output LATx bit at 0, and set TRIS bit to INPUT (let external resistor pull-up line to high)
listen for response from peripheral: keep output LATx bit at 0, and set TRIS bit to INPUT. Let external resistor pull-up line to high when peripheral is transmitting a 1, or let the peripheral pull the line low when peripheral is transmitting a 0. Read the bit from the PORTx pin.
If both ends of the bus correctly follow this protocol (in particular, if neither end actively drives the line to high), then you never have to worry about contention or current spikes.
It`s important to remember that any IO switching in high speed generates EMI.
Depending of switching frequency, board layout and devices sensibilities, this EMI can affect performance and reliability of your application.
If you are having problems in your application use an oscilloscope to check for irradiated EMI in your board lanes.