This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Tool/software:
A question about the ADC of the
AM243x:
We would like to use a SinCos resolver.
We assume: MaxRPM= 9000Rpm -> 150 revolutions / s
Number of lines of the
resolver = 120 lines.
This results in
18000 sin-periods / s = 55.5us per sin-period
We would like to use the both channels of the SinCos resolver SIMULTAN.
According to the data sheet of the AM243x, sampling with the ADC is NOT SIMULTANEOUS but only sequential.
How can we sample the two sin and cos values of the resolver into the processor without sampling delay in order to avoid angular displacements?
thanks
Jan
Hello Jan,
I have assigned this thread to an Hw expert to provide inputs on this requirement, and I am also notified in this thread. If there are any software related queries I will provide a reply.
Regards,
Anil.
Hello Jan,
Refer below block in the TRM
Figure 12-6. Functional Block Diagram
According to the data sheet of the AM243x, sampling with the ADC is NOT SIMULTANEOUS but only sequential.
How can we sample the two sin and cos values of the resolver into the processor without sampling delay in order to avoid angular displacements?
As you mentioned the ADC is sequential.
You may have to consider an ADC that samples simultaneously to the Sine
Please check if this helps
https://www.ti.com/tool/TIDA-00178
https://www.ti.com/tool/TIDA-00176
https://www.ti.com/tool/TIDA-01527
Regards,
Sreenivasa
Hello Jan,
Please refer additional input i added based on my discussion with the device expert:
The ADC implemented in AM243x is a single channel SAR ADC with a multi-input analog multiplexer inserted on the input path. Therefore, it is not possible for the ADC to sample two inputs simultaneously.
Regards,
Sreenivasa
Hello Kallikuppa,
yes this has been already clear to the customer. And if we understood correctly also an external ADC like ADS9234R does not work on the QSPI interface (maybe you can confirm that finally). So the only option would be either ADS9234R via PRU (this is not possible here because the PRU will be in use) or the decision that the customer made to use an ADS7253 or ADS7251.
Thank you
Jan
Hello Jan,
Here is the list of ADCs that have been interfaced,
And if we understood correctly also an external ADC like ADS9234R does not work on the QSPI interface (maybe you can confirm that finally). So the only option would be either ADS9234R via PRU (this is not possible here because the PRU will be in use)
QSPI i am not sure. PRU - understand.
ADS7253 or ADS7251
refer section Terminal Descriptions of ADC data sheet. Any thoughts how customer plans to interface to Am243x?
Regards,
Sreenivasa
Hello Kallikuppa,
QSPI i am not sure
The documentation of the AM243 says that QSPI is only for memory interface. ADS9234R has a dual SDO line, so there for normal SPI does not work.
refer section Terminal Descriptions of ADC data sheet. Any thoughts how customer plans to interface to Am243x?
For me it looks like through standard spi interface. Do you have any concerns here?
Thanks
Jan
Hello Jan,
Thank you.
refer section Terminal Descriptions of ADC data sheet. Any thoughts how customer plans to interface to Am243x?For me it looks like through standard spi interface. Do you have any concerns here?
As i understand, these ADCs also seem to have dual SDO output.
Regards,
Sreenivasa
Hello Kallikuppa,
As i understand, these ADCs also seem to have dual SDO output.
The ADS7253 has a Single-SDO mode:
whereas ADS9234R has not such a mode.
Regards
Jan
Hello Kallikuppa,
can you also clarify the contraints of the QSPI interface and why it is just declared as an memory interface?
Thanks
Jan
Hello Jan,
Thank you.
TIs current implementations and software enablement has been around the OSPI/QSPI memory.
Regards,
Sreenivasa
TIs current implementations and software enablement has been around the OSPI/QSPI memory.
So this means that it is a software and not a hardware constraint, correct?
Hello Jan,
Thank you.
TIs current implementations and software enablement has been around the OSPI/QSPI memory.So this means that it is a software and not a hardware constraint, correct?
I do not have a way to confirm the query.
This may require a detailed analysis.
Regards,
Sreenivasa