Using STM32F413 NUCLEO-144 development board and having issues with DAC on PA4 CH1.

[MCrew.1]

Hello everyone,

I am using the STM32F413 Nucleo-144 development board to generate 2 sinusoidal wave forms at 60Hz. To do this, I have DAC_OUT1 (PA4) and DAC_OUT2 (PA5) conversions running through DMA. The signal from PA5 appears as expected when I measure on the oscilloscope (approximately 3mV ~ 3.3V), however this is not the case with the signal from PA4. The top and bottom of PA4 signal gets clipped as seen in the following image:

The top signal is generated by PA4, the bottom signal is generated by PA5.

I cannot see any reason for this being the case as the same values are being converted for both signals, and I do not see any datasheet difference between the 2 DACs. if anyone can suggest any reason why this might be the case, it would be greatly appreciated!

Thanks

[waclawek.jan]

Different settings of DACs (buffer on/off)? Show registers.

Different loading on the board? Review board schematics.

JW

[MCrew.1]

The timer registers are set the same (buffer = on).

Can you explain what you mean by different loading on the board? I reviewed the schematic and they appear to be the same.

MC

[gregstm]

Weird... the signal from PA4 is not only clipped, but appears to be more linear than a sine wave - more like a triangle wave being clipped. Is triangle wave output somehow being enabled?

[waclawek.jan]

> The timer registers are set the same (buffer = on).

Read out and post DAC registers content.

Also, try to switch off the buffers and post result.

JW

[MCrew.1]

By switching the output buffers to disabled, I was able to generate the following waveforms. This is the results require, thank you! Can you explain why this works? I am not sure I understand the output buffers correctly

[waclawek.jan]

> Can you explain why this works?

Not quite.

Search this forum for "DAC inverted" output. The reason is physical damage to the buffer. Your symptoms are somewhat different, but similar in that the suspicious output is related to buffer being switched on, and I suspect damage is the underlying reason here, too.

JW

[Paul1]

NucleoL476-RG: 

* DAC1 Output Buffer = Disabled OK for accurate Output (BufferDriver requires Calibration, else may offset/clipping)

* DAC2 Output Buffer = Enabled for LED drive else heavy load drags down output (BufferDriver requires Calibration, else may offset/clipping)

I haven't done DAC calibration yet

The output buffers provide a stronger drive out of the DAC pins, at the cost of needing calibration to compensate for offset/gain effects (see AN4566). (maybe the DAC buffers are opamp circuits)

[waclawek.jan]

Is there a question?

You may want to start a new thread, perhaps linking here if it's a similar problem.

JW

[Paul1]

Just a note to others using this demo and the DACs.