I've shared the idea over at #sigrok:libera.chat :-)
After some modification, the oscilloscope works.
I'm reading the python3.9/site-packages/pslab-2.4.0-py3.9.egg/pslab/instrument/oscilloscope.py and power_supply.py
I'm wondering why the abstraction of the two instruments is totally different, why the oscilloscope command is mainly decoded on the PIC24, but power source code sends commands to control I2C bus directly?
Since I'm using two 2-channel DACs on SPI bus, shall I change the abstraction level of the code?
The best practice of abstraction is not overlay between different layers. For example: layer 1, PSLab, layer 2, instrument(oscilloscope), layer 3, (hardware devices MCP4728, ADC, etc).
I'm reading the python3.9/site-packages/pslab-2.4.0-py3.9.egg/pslab/instrument/oscilloscope.py and power_supply.py
I'm wondering why the abstraction of the two instruments is totally different, why the oscilloscope command is mainly decoded on the PIC24, but power source code sends commands to control I2C bus directly?
Since I'm using two 2-channel DACs on SPI bus, shall I change the abstraction level of the code?
The best practice of abstraction is not overlay between different layers. For example: layer 1, PSLab, layer 2, instrument(oscilloscope), layer 3, (hardware devices MCP4728, ADC, etc).
@RafaelLeeImg: see here https://github.com/fossasia/pslab-desktop/blob/development/public/electron.js#L102 to enable debugging with the background processes that are used to talk to the Python library, just set
debug: true
These are 2 screenshots with analog front end error. The problem is fixed today.
2 replies
These are 2 screenshots with gain=1 and gain=4, amplitude is wrong since the circuit is changed. The wave form is correct.
That limit is due to there are protection diode on the 3 resistor pins, the diode are added by me on the image.
Is there any good way to use a different DAC in this project?
The DAC I'm using is MCP4822 which is a 2-channel DAC with SPI bus.
The DAC I'm using is MCP4822 which is a 2-channel DAC with SPI bus.
# pslab-python/pslab/instrument/power_supply.py
self._mcp4728 = I2CSlave(self._ADDRESS, self._device)
self._pv1 = VoltageSource(self._mcp4728, "PV1")
self._pv2 = VoltageSource(self._mcp4728, "PV2")
self._pv3 = VoltageSource(self._mcp4728, "PV3")
self._pcs = CurrentSource(self._mcp4728)
That makes the oscilloscope noisy with high gain. The image is showed with the gain of 16.
+5V is significantly affected by the switching cap voltage inverter. Since the operating frequency of 7660 is too low. There is no good way to filter this ripple with small capacitors.
I want to use some other ways to generate negative voltage. I want try buck-boost.
I want to use some other ways to generate negative voltage. I want try buck-boost.
Since the semiconductor industry is evolving rapidly, the era that use low frequency power converter and big capacitors are already faded away.
The calculation value of voltage ripple is quite near the measured result. If I increase the cap to 100uF, the noise will be several mV, that is achievable but the problem is not solved elegantly.
The current of +10V rail is 15mA, measured with DMM.
Voltage ripple = 1510^-3 mA(1/20000 Hz)/(1010^-6 F)=0.075V.
The current of +10V rail is 15mA, measured with DMM.
Voltage ripple = 1510^-3 mA(1/20000 Hz)/(1010^-6 F)=0.075V.
I'm going to modify the ±6V to ±7V by adding a diode to 6.8V zener diode. The good side is to add a reverse PN node which has a negative temperature coefficient to compensate the
I'm considering to change the U10 OPAMP which used for wave generator to ±7V. Thus, I can remove ±10 rail for OPAMPs voltage supply. The output voltage of ±3V can be covered with this configuration.
I'm considering to change the U10 OPAMP which used for wave generator to ±7V. Thus, I can remove ±10 rail for OPAMPs voltage supply. The output voltage of ±3V can be covered with this configuration.
I didn't find other replacement of MCP4822. I tried to find some MCU with 2 12bit DAC, for example, STM32L431RCT6 the price is OK (around ￥25), but the stock is only like 100pcs. Or, sample-and-hold amplifier is another option, with that, we can reduce the DAC channel to 2 or 1. That leads to some work on firmware.
There are exceptions, for example, use 2 audio DAC connected as bipolar (BTL) to drive speakers. For 2 channels, it needs 4 DAC. That's only used in very high-end audio amplifier systems.
Sorry for not having joined the last meetings. I've been a little busy again, about to present something fancy again, today noon at CLT: https://chemnitzer.linux-tage.de/2022/en/programm/beitrag/226
During FOSSASIA Summit, I might be in Berlin, not 100% sure yet. Let me know if you'd like to meet. I'd arrive at the Wednesday, and stay until Sunday evening. Will let you know when it's for sure. :-)
1 reply
-6V rail, the noise is really low.
The problem is here, the boost converter is working under discontinuous mode, that causes 5V tail near 100mV ripple. This leads to the noise of 3.3V rail and VDDA.
The 10V rail consumes 15mA of current, with that condition, I shall use an inductor with larger value. I'm using a 4.7uH now, and I'm waiting for the components to arrive.
I remove all the big capacitors, after I solve the discontinuous oscillation problem. I think the overall performance will be good.
One issue I ran into with this I2C sensor:
https://github.com/tasmota/docs/pull/954/files
The docs led me to believe the boards had the same pin order.
I.e., I had GND and SDA swapped.
Anyhow, there's still nothing detected. I will see if it works with an ESP board here when I have the time.
Could be because of:
Only a single AHT10 can be connected to the I^2^C bus and no other I^2^C devices can be connected.
The pslab's I2C bus also has the mcp4728 on it, as well as the rtc if you're using the v6 prototype.