Topics created by knebb

@cleve
Thanks for letting me know.

To be honest- there is no urgent need for upgrading.

I just prefer to be on the up-to-date version knowing the old stable gets outdated faser than expected ;)

Do you have an estimation when it might be available for Trixie?

Thanks!

/KNEBB

@cleve

Neuron M103 (the old one, not the current) which is still running with Raspbian 10/ buster.

How do I get the information about the firmware?

(BTW: I tried to upgrade to latest Raspbian, but failed)

Hello @knebb,
we do not provide support for Buster based systems anymore. However, on the latest firmware with BaseOS 12, the value is the actual value (the datatype is real - float32), there is no further calculation needed.

Replying to mysqlf as I found a nice guy who did the cabling for me.
As I want to control two vents the wiring has been extended accordingly:
a4ab227e-bcae-4181-a2bc-968b0c0611af-grafik.png

So for each vent I need a 47k resistor and a 12V diode.

That's it! Easy for someone who knowns. Impossible to do it for me.

/KNEBB

@cleve
Thanks for replying.

I will try the update again and create a new topic in case it fails.

/KNEBB

Hi everybody,

It has been fixed with a new version of the unipi-kernel-modules package.
At the same time, we increased the output testing of packages.

Best regards,
Antonin

@gabrielsoudry Hello Gabriel,

We plan to release the new Bookworm-compatible EVOK, along with new OS images, by the end of February. However, if we encounter bugs during testing, this may extend until the end of March.

ZR

tag @martin_triska

Hi @knebb,
you have verified, that the HW is working and returns correct value via standard tool. Checking and fixing your code is unfortunately beyond the scope of the free support.

Thank you for understanding and have a nice day,
Martin

Hi @knebb,
the firmware has been updated on our Knowledge Base as well: https://kb.unipi.technology/en:files:products:05-sensors:iaq

Martin

Hi @knebb,

Thanks for this post, it's very helpful.

Best regards,
Antonin, Unipi

@knebb Try running it on lower voltage, or put another resistor in series and do the recalculation in the software.

Martin

@Martin-Kudláček :

Of course. Here you are. I set the registers and am measuring the AO port during the sleep() cycle.

#include "header.h" #include "vars.h" void main(int argc, char* argv[]) { modbus_t *modbusglob; union { uint16_t word[2]; uint32_t dword; float fl; } counter,direct; int n[9],mode; /********************************* Function **********************************/ void fieldbus_free(modbus_t *modbus) { if (modbus != NULL) { modbus_close(modbus); modbus_free(modbus); } } /********************************* Function **********************************/ modbus_t * fieldbus_new(void) { int err; modbus_t *modbus; char host[15]; int port; strcpy(host,"127.0.0.1"); port = 502; modbus = modbus_new_tcp(host, port); if (modbus == NULL ) { fieldbus_free(modbus); exit (55); } err = modbus_connect(modbus); if (err == -1) { fieldbus_free(modbus); modbus = NULL; exit (55); }; err=modbus_set_slave(modbus, 0); if( err == -1) { fieldbus_free(modbus); modbus = NULL; exit (60); } else { return modbus; } } /********************************* Function **********************************/ int setvoltage(float wish) { int errlvl; union { uint16_t word[2]; uint32_t dword; float fl; } wert; uint16_t ao; wert.fl=wish; errlvl=modbus_write_registers(modbusglob, 3000, 2, (&wert.word[0])); if (errlvl == -1) {printf("ERR1 modbus_write_register, %s\n",modbus_strerror(errno));} else { printf("Register set to %f\n",wert.fl);} return errlvl; } modbusglob=fieldbus_new(); mode=1; setvoltage(1.3); sleep(60); setvoltage(3.5); sleep(60); setvoltage(5.7); sleep(60); setvoltage(7.9); sleep(60); setvoltage(10.0); sleep(60); setvoltage(0); sleep(60); setvoltage(5.0); fieldbus_free(modbusglob); exit (0); }

Hi @knebb, there are two approaches and you figured out both of them:

Do the calculation and write to register 2 Write a value to the register 3000. The value has to be 32bit signed integer.

I checked the documentation and I think there is a mistake with the registers 3000, 3001 and 3002. Those should be 3000, 3002 and 3004 since the 32bit values need two consecutive registers. I will check what is in the firmware and let you know!

Best regards,
Martin

Hi @knebb,
definitely not. The DOs are NPN type (open collector). You can read more about how to read the NPN outputs with digital inputs here: https://forum.unipi.technology/topic/712/neuron-12v-digital-input/2

Best regards,
Martin

Just as an update.

I was struggling with multiple registers which are used i.e. for the AO1.1 port of my Neuron M103.

So for the AO registers 3000 and 3001 are given as "real".

I extended the above union as follows:

union { uint16_t word[2]; uint32_t dword; float fl; } counter;

And I am reading the register like this:

modbus_read_registers(modbusglob, 3000, 2, &counter.word[0]);

It reads 2 registers starting at 3000 and writes the result in the word array where it is automatically converted to "float" by the union structure.

Works fine!

/KNEBB

@knebb said in Need Some sort of Jumpstart:

it writes the state as 0 or 1 into each of the uint8_t .

Or, using semantic constants, as FALSE and TRUE.

Directly using bits would be really slow and error prone (and a nightmare for language bindings).

@tomas_hora It really depends on the actual application. I needed to measure 50mm difference in water level in tight space (approx. 30mm x 30mm). So for me, mechanical solution was a no-go, since there are no such small sensors.

But for @knebb's purpose, some set of float switches (at the bottom, 1/4, 1/2, 3/4 and at the top) could be easy and affordable solution. I would connect them to a resistor ladder and 10V power supply thus converting the 5 states into evenly spreaded 0-10V which can be directly measured by UniPi's AI input. This could be better than occupying 5 DI inputs, if the number of DI's is also an issue.

Ah, well. No I am absolutely fine with this.

My Cacti monitors the system every 5 minutes. So no need for faster- I am fine with a minute.

Thanks again!

Oh, and it is working stable. Now running for nearly 2days without a reboot.