[George]

I think i solved the problem by upgrading the firmware to SFW198E060 (Pck051b060).

What is difference between Pck051b060 and PCK054A010?

[George]

I think i solved the problem by upgrading the firmware to SFW198E060 (Pck051b060).

What is difference between Pck051b060 and PCK054A010?

[George]

Hello, if you had read the comments inside “MyProgram” in project i had sent, you could see that i had located the problem in malloc buffer size and the memory allocation mechanism of the linker. The information for malloc memory of SysSetAnOut FB  is very useful.

Another problem that i found is that the measured output values are abnormal. The analog outputs drive the reference inputs of motor inverters. If for example set channel 1 to 1V and all the other channels to 0V, the measured values to channels other than channel 1 are like -3.5V or 2.5V etc.

I checked that between analog inputs of inverters and the earth strip the impedance is about 100KOhm. The inverters specs for reference input give Ri about 40KOhm. If i disconnect the analog outputs from inverters reference inputs the measured values are as expected.

The solution that implemented and solves my problem is to connect opamp buffers between analog outputs of the module and the inverters inputs. If you could suggest a more simple solution, it would be nice.

[George]

I have connected two PCB126C150 on a MPS054A110 CPU (fw:SFW198E000) and i made a program to update the eight analog output channels.

If i run this program from Task background, execution is as expected.

if i run from Task slow, the update is done only on the first module (as i can see at Done output of SysSetAnOut FB). The same even if i play with longer loop time of slow Task.

I may send the program i prepared.

[George]

So, the information (20 kB RAM user data) I read in product characteristics is not valid?

Is there the 20Kb option available?

[George]

I have a MPS056A320 (SFW198E010) and compiling my program i get:

Used code size:       9180h (   36 KByte)
Free code space:      16E70h (   91 KByte)
Total code space:      1FFF0h (  127 KByte)

Used data space:       1E88h (    7 KByte)
Free data space:        178h (    0 KByte)
Total data space:       2000h (    8 KByte)

Used retain data space:        214h (     0 KByte)
Free retain data space:        1D4h (     0 KByte)
Total retain data space:    3E8h (     0 KByte)

In the characteristics of the CPU i see the data memory size is 20Kb RAM which is different from 8KBytes that i get. I have already executed the command “Force target image upload”.

 

[George]

Thank you for your answer.

My problem is not to characterize the device and to find the correction offset (I already found it), but that of the settling time needed (from asymmetric Ton/Toff to symmetric, about 25 .. 30 minutes) from the machine startup every morning to operate. Is there a workaround on this issue?

How affects the load current of the PWM output the operating conditions?

Maybe light load current better for rise time, but not good for temperature drift?

[George]

I want to use the PWM output on a MPS054, but i measured an asymmetry on the Ton / Toff at 50 % duty cycle. At frequency 400 hz i measured Ton: 1.183ms and Toff : 1.318 ms. After adding an offset of 3.0 % to 50 % the asymmetry is compensated.

The problem  (when the device is cold) is that after power up i measure  Ton:1.280 ms, Ton:1.220 ms and only after 20 – 25 minutes the compensation becomes active. Drifts slowly to Ton:1.253ms, Toff:1.247ms. Could be a device hardware drift?

Looks like there is need to warmup for the first power up.

 

 

 

 

[Autore]

Post