The data acqusition of the RedPitayaDAQServer project is based on two data flows to and from the upper 128 MB of the RedPitaya memory. This memory region acts as a ring buffer for the acquired samples and can be queried by clients using SCPI commands.
Signal acquisition within a cluster is based on a shared clock and trigger signal distributed via cables between the RedPitayas. Once triggered, all FPGAs continuously write the samples from their ADC channel to the sample ring-buffer with each clock tick. Both ADC channels on a RedPitaya are written to the buffer at the same time. The 14-bit values of the ADCs are converted to 16-bit signed integer samples and then concatenated into one 32-bit value, which is then written to the buffer. The sampling rate of the system can be adjusted by setting a decimation parameter and the decimation itself is realized with a CIC filter.
Internally, the FPGA keeps track of a 64-bit writepointer register pointing into the ring-buffer and increments this value with each new sample pair. Additionally, the writepointer also counts the number of buffer overflows. As the size of the buffer region is a power of two, these two components of the writepointer can be interpreted as one 64-bit number counting the samples from acquisition start. For the 128 MB buffer, this means that the lower 25 bits of the writepointer are the buffer location and the remaining bits are the overflow counter.
As the writepointer is reset and incremented based on a shared clock and trigger signal, it is synchronized across all FPGA images in a cluster. The logic implemented with the reprogrammable hardware is also the only logic of the RedPitayaDAQServer with predictable timing behaviour. All other components of the system implement their (timing related) logic in reference to the current writepointer values. With a known sampling rate, the writepointer can also be seen as the clock of the server and client components.
To retrieve samples from the server a client can supply a similar pointer, called readpointer, together with the number of samples to retrieve. The server then extracts the buffer position from the readpointer and transmits the requested amount of samples over the data socket. This transmission happens either way, even if the samples are overwritten. However, the server uses the whole readpointer, including the buffer overflows, to check if the requested samples were overwritten.
If the distance between the write- and readpointer is larger than the buffer size the overflow status flag is set. If during the transmission the requested samples are overwritten the corrupted flag is set. These flags can be queried individually or together in a status byte via SCPI commands.
This distance can not only be used to see if samples were overwritten, but also to track how well the client is able to keep up with the FPGA during a series of sample transmissions. If this distance increases over time, the FPGA is creating more samples than the server can transmit to the client. To allow a client to track this value, this distance is stored as a 64-bit value deltaRead for the latest transmission and can be queried. Additionally, the server also tracks the duration of the transmission as writepointer "clock ticks" as a 64-bit value deltaSend, which is just the difference between the writepointer at the start and end of a transmission.
There are several things to consider when attempting to retrieve samples at a high sampling rate, for larger cluster sizes or for longer periods of time. Most of the following points were implemented/considered in the Julia reference implementation, but would become relevant when implementing custom clients.
As the server will always transmit samples just based on the buffer position of a readpointer, if a client wants to only receive certain samples it needs to wait for them to exist in the buffer. This requires querying the writepointer until it is larger than the desired readpointer.
If the number of requested samples is larger than the buffer, the sample should be requested in smaller chunks as the server would otherwise return samples that were not written yet. In a cluster scenario the i-th chunk should be requested from all RedPitayas in the cluster before requesting the next chunk to avoid "starvation" effects.
The status and performance data of a transmission can only be queried after the transmission has finished, which requires additionaly communication overhead.
To help clients with these issues, the server offers a second type of sample transmission in which samples, status and performance data is pipelined. In such a query a client first transmits a readpointer, together with the number of requested samples and the number of samples belonging to a chunk. The server itself then tracks the writepointer and transmits a chunk as soon as it becomes available and immidiatey follows that up with the status and performance data of the transmission. This way additional communication overheard is reduced and after the inital request a client just needs to read data until the transmission finishes.
The samples sent by the server are the 16-bit values of the ADC channel of a RedPitaya. However, one might want to work with voltage values instead or encapsulate samples into a repeating concept like frames. The Julia client library offers functions to convert samples into such a concept or to directly request a number of frames instead of a number of samples.
Here, frames are comprised of periods, which in turn are comprised of samples. During the conversion process the 16-bit binary values can also be converted to floating point numbers representing a voltage if the RedPitaya was calibrated beforehand. In this calibration process, a client can store scale and offset values for each channel in the EEPROM of the RedPitaya. When the client establishes a connection to the server, it reads these values and can use them to translate the 16-bit values into a respective voltage value.
The highest supported sampling rate of the RedPitayaDAQServer is 15.625 MHz or 15.625 MS/s, as this is the sampling rate at which a single RedPitaya can produce and transmit samples continously without data loss given the 1 Gbit/s limit of the ethernet connection from the RedPitaya. This rate is a achieved with a decimation of 8 from the base 125 MHz sampling rate of the RedPitaya hardware.
At this sampling rate a single RedPitaya produces new samples at a data rate of 500 Mbit/s. Furthermore at this rate, once a sample has been written to the buffer it exists for 2.15s before being overwritten again (Time-To-Live, TTL). An overview of these metrics for different decimation factors is shown in the following table:
Decimation
MHz
MByte/s
Mbit/s
TTL
64
1.95
7.81
62.5
17.18s
32
3.91
15.63
125
8.59s
16
7.81
31.25
250
4.29s
8
15.63
62.5
500
2.15s
The table only refers to the data rate of new samples being produced. The data rate of samples being transmitted to a client can differ greatly depending on how the client queries and processes the samples and the available network bandwidth and usage. At the higher sampling rates it is recommended to have client threads that exclusively receive samples and perform any computation on samples in different threads to maximise the transmission speed, as a server can only transmit data at a rate of just above 500 Mbit/s. This exceeds the highest supported sampling rate by only a few Mbit/s and a client with frequency interruptions of its sample reception might not be able to keep up with the sampling rate.
Settings
This document was generated with Documenter.jl version 0.27.24 on Monday 27 March 2023. Using Julia version 1.8.5.
The data acqusition of the RedPitayaDAQServer project is based on two data flows to and from the upper 128 MB of the RedPitaya memory. This memory region acts as a ring buffer for the acquired samples and can be queried by clients using SCPI commands.
Signal acquisition within a cluster is based on a shared clock and trigger signal distributed via cables between the RedPitayas. Once triggered, all FPGAs continuously write the samples from their ADC channel to the sample ring-buffer with each clock tick. Both ADC channels on a RedPitaya are written to the buffer at the same time. The 14-bit values of the ADCs are converted to 16-bit signed integer samples and then concatenated into one 32-bit value, which is then written to the buffer. The sampling rate of the system can be adjusted by setting a decimation parameter and the decimation itself is realized with a CIC filter.
Internally, the FPGA keeps track of a 64-bit writepointer register pointing into the ring-buffer and increments this value with each new sample pair. Additionally, the writepointer also counts the number of buffer overflows. As the size of the buffer region is a power of two, these two components of the writepointer can be interpreted as one 64-bit number counting the samples from acquisition start. For the 128 MB buffer, this means that the lower 25 bits of the writepointer are the buffer location and the remaining bits are the overflow counter.
As the writepointer is reset and incremented based on a shared clock and trigger signal, it is synchronized across all FPGA images in a cluster. The logic implemented with the reprogrammable hardware is also the only logic of the RedPitayaDAQServer with predictable timing behaviour. All other components of the system implement their (timing related) logic in reference to the current writepointer values. With a known sampling rate, the writepointer can also be seen as the clock of the server and client components.
To retrieve samples from the server a client can supply a similar pointer, called readpointer, together with the number of samples to retrieve. The server then extracts the buffer position from the readpointer and transmits the requested amount of samples over the data socket. This transmission happens either way, even if the samples are overwritten. However, the server uses the whole readpointer, including the buffer overflows, to check if the requested samples were overwritten.
If the distance between the write- and readpointer is larger than the buffer size the overflow status flag is set. If during the transmission the requested samples are overwritten the corrupted flag is set. These flags can be queried individually or together in a status byte via SCPI commands.
This distance can not only be used to see if samples were overwritten, but also to track how well the client is able to keep up with the FPGA during a series of sample transmissions. If this distance increases over time, the FPGA is creating more samples than the server can transmit to the client. To allow a client to track this value, this distance is stored as a 64-bit value deltaRead for the latest transmission and can be queried. Additionally, the server also tracks the duration of the transmission as writepointer "clock ticks" as a 64-bit value deltaSend, which is just the difference between the writepointer at the start and end of a transmission.
There are several things to consider when attempting to retrieve samples at a high sampling rate, for larger cluster sizes or for longer periods of time. Most of the following points were implemented/considered in the Julia reference implementation, but would become relevant when implementing custom clients.
As the server will always transmit samples just based on the buffer position of a readpointer, if a client wants to only receive certain samples it needs to wait for them to exist in the buffer. This requires querying the writepointer until it is larger than the desired readpointer.
If the number of requested samples is larger than the buffer, the sample should be requested in smaller chunks as the server would otherwise return samples that were not written yet. In a cluster scenario the i-th chunk should be requested from all RedPitayas in the cluster before requesting the next chunk to avoid "starvation" effects.
The status and performance data of a transmission can only be queried after the transmission has finished, which requires additionaly communication overhead.
To help clients with these issues, the server offers a second type of sample transmission in which samples, status and performance data is pipelined. In such a query a client first transmits a readpointer, together with the number of requested samples and the number of samples belonging to a chunk. The server itself then tracks the writepointer and transmits a chunk as soon as it becomes available and immidiatey follows that up with the status and performance data of the transmission. This way additional communication overheard is reduced and after the inital request a client just needs to read data until the transmission finishes.
The samples sent by the server are the 16-bit values of the ADC channel of a RedPitaya. However, one might want to work with voltage values instead or encapsulate samples into a repeating concept like frames. The Julia client library offers functions to convert samples into such a concept or to directly request a number of frames instead of a number of samples.
Here, frames are comprised of periods, which in turn are comprised of samples. During the conversion process the 16-bit binary values can also be converted to floating point numbers representing a voltage if the RedPitaya was calibrated beforehand. In this calibration process, a client can store scale and offset values for each channel in the EEPROM of the RedPitaya. When the client establishes a connection to the server, it reads these values and can use them to translate the 16-bit values into a respective voltage value.
The highest supported sampling rate of the RedPitayaDAQServer is 15.625 MHz or 15.625 MS/s, as this is the sampling rate at which a single RedPitaya can produce and transmit samples continously without data loss given the 1 Gbit/s limit of the ethernet connection from the RedPitaya. This rate is a achieved with a decimation of 8 from the base 125 MHz sampling rate of the RedPitaya hardware.
At this sampling rate a single RedPitaya produces new samples at a data rate of 500 Mbit/s. Furthermore at this rate, once a sample has been written to the buffer it exists for 2.15s before being overwritten again (Time-To-Live, TTL). An overview of these metrics for different decimation factors is shown in the following table:
Decimation
MHz
MByte/s
Mbit/s
TTL
64
1.95
7.81
62.5
17.18s
32
3.91
15.63
125
8.59s
16
7.81
31.25
250
4.29s
8
15.63
62.5
500
2.15s
The table only refers to the data rate of new samples being produced. The data rate of samples being transmitted to a client can differ greatly depending on how the client queries and processes the samples and the available network bandwidth and usage. At the higher sampling rates it is recommended to have client threads that exclusively receive samples and perform any computation on samples in different threads to maximise the transmission speed, as a server can only transmit data at a rate of just above 500 Mbit/s. This exceeds the highest supported sampling rate by only a few Mbit/s and a client with frequency interruptions of its sample reception might not be able to keep up with the sampling rate.
Settings
This document was generated with Documenter.jl version 0.27.24 on Friday 28 April 2023. Using Julia version 1.8.5.
diff --git a/dev/architecture.html b/dev/architecture.html
index 00a26c26..1ccdcd0d 100644
--- a/dev/architecture.html
+++ b/dev/architecture.html
@@ -1,2 +1,2 @@
-Architecture · RP DAQ Server
The RedPitayaDAQServer project is implemented as a distributed system in which one client connects to a cluster of RedPitaya boards. The project has four software components:
The FPGA image is responsible for generating and acquiring synchronized out- and input signals. The server acts as an intermediary to the FPGA over a TCP/IP connection, which allows remote clients to configure the FPGA image and retrieve samples. Furthermore, the server also maintains a thread that takes part in signal generation.
The Julia client library can be used to implement a data acquisition client application, which controls a (cluster of) RedPitaya(s). This Julia library acts as a reference, but in principle it is possible to write clients in any programming language, as the communication is language agnostic. In the example directory we provide a rudimentary Python client that allows to perform a simple data acquisition experiment.
The various components of the distributed system communicate over different interfaces. Communication within a RedPitaya is based on memory-mapped I/O, while communication between the server and a client is based on SCPI commands over a TCP/IP connection, usually over Ethernet. Lastly communication between RedPitayas is limited to signals distributed over cables as described in Cluster.
The FPGA image is directly connected to certain memory regions that can be memory mapped on the CPU side of the RedPitaya. Both the CPU and the FPGA image access the reserved main memory region as a sample buffer. The C library rp-daq-lib, which is located under src/lib/ in the project repository, encapsulates these memory accesses into a convenient C library. It is possible to use this C library directly on the RedPitaya when no communication with the host system is required, i.e. if one wants to write the acquired data into a file. When making changes to the FPGA image one may need to adapt the rp-daq-lib C library.
The server itself uses the rp-daq-lib library to interface with the FPGA image.
The server on each RedPitaya has two TCP sockets to which a client needs to connect. The first is the command socket on port 5025 and the second is the data socket on port 5026. Over the former, a client can send SCPI commands to the server and receive replies, while the latter is used for sending binary data such as the samples acquired by the ADCs.
SCPI commands are ASCII strings, such as RP:ADC:DECimation, which the server translates into C function calls. As an example these calls could invoke a function of the rp-daq-lib library to set the decimation of the sampling rate or instruct the server to transmit data over the data socket. A list of the available SCPI commands can be found here.
At any point a server is only connected to one client and establishing a new connection stops any current signal generation and acquisition.
Settings
This document was generated with Documenter.jl version 0.27.24 on Monday 27 March 2023. Using Julia version 1.8.5.
The RedPitayaDAQServer project is implemented as a distributed system in which one client connects to a cluster of RedPitaya boards. The project has four software components:
The FPGA image is responsible for generating and acquiring synchronized out- and input signals. The server acts as an intermediary to the FPGA over a TCP/IP connection, which allows remote clients to configure the FPGA image and retrieve samples. Furthermore, the server also maintains a thread that takes part in signal generation.
The Julia client library can be used to implement a data acquisition client application, which controls a (cluster of) RedPitaya(s). This Julia library acts as a reference, but in principle it is possible to write clients in any programming language, as the communication is language agnostic. In the example directory we provide a rudimentary Python client that allows to perform a simple data acquisition experiment.
The various components of the distributed system communicate over different interfaces. Communication within a RedPitaya is based on memory-mapped I/O, while communication between the server and a client is based on SCPI commands over a TCP/IP connection, usually over Ethernet. Lastly communication between RedPitayas is limited to signals distributed over cables as described in Cluster.
The FPGA image is directly connected to certain memory regions that can be memory mapped on the CPU side of the RedPitaya. Both the CPU and the FPGA image access the reserved main memory region as a sample buffer. The C library rp-daq-lib, which is located under src/lib/ in the project repository, encapsulates these memory accesses into a convenient C library. It is possible to use this C library directly on the RedPitaya when no communication with the host system is required, i.e. if one wants to write the acquired data into a file. When making changes to the FPGA image one may need to adapt the rp-daq-lib C library.
The server itself uses the rp-daq-lib library to interface with the FPGA image.
The server on each RedPitaya has two TCP sockets to which a client needs to connect. The first is the command socket on port 5025 and the second is the data socket on port 5026. Over the former, a client can send SCPI commands to the server and receive replies, while the latter is used for sending binary data such as the samples acquired by the ADCs.
SCPI commands are ASCII strings, such as RP:ADC:DECimation, which the server translates into C function calls. As an example these calls could invoke a function of the rp-daq-lib library to set the decimation of the sampling rate or instruct the server to transmit data over the data socket. A list of the available SCPI commands can be found here.
At any point a server is only connected to one client and establishing a new connection stops any current signal generation and acquisition.
Settings
This document was generated with Documenter.jl version 0.27.24 on Friday 28 April 2023. Using Julia version 1.8.5.
diff --git a/dev/client.html b/dev/client.html
index e1c28d3c..2db1a31d 100644
--- a/dev/client.html
+++ b/dev/client.html
@@ -1,105 +1,105 @@
-Client Library · RP DAQ Server
This page contains documentation of the public API of the Julia client. In the Julia REPL one can access this documentation by entering the help mode with ? and then writing the function for which the documentation should be shown.
Struct representing a connection to a RedPitayaDAQServer.
Contains the sockets used for communication and connection related metadata. Also contains fields for client specific concepts such as periods, frames and calibration values.
RedPitaya(ip [, port = 5025, dataPort=5026, isMaster = false])
Construct a RedPitaya.
During the construction the connection is established and the calibration values are loaded from the RedPitayas EEPROM. Throws an error if a timeout occurs while attempting to connect.
Examples
julia> rp = RedPitaya("192.168.1.100");
+Client Library · RP DAQ Server
This page contains documentation of the public API of the Julia client. In the Julia REPL one can access this documentation by entering the help mode with ? and then writing the function for which the documentation should be shown.
Struct representing a connection to a RedPitayaDAQServer.
Contains the sockets used for communication and connection related metadata. Also contains fields for client specific concepts such as periods, frames and calibration values.
RedPitaya(ip [, port = 5025, dataPort=5026, isMaster = false])
Construct a RedPitaya.
During the construction the connection is established and the calibration values are loaded from the RedPitayas EEPROM. Throws an error if a timeout occurs while attempting to connect.
Receive a string from the RedPitaya command socket. Reads until a whole line is received or timeout seconds passed. In the latter case an error is thrown.
Receive a string from the RedPitaya command socket. Reads until a whole line is received or timeout seconds passed. In the latter case an error is thrown.
Struct representing a batch of SCPI commands for a RedPitaya. Only commands that interact exclusively with the command socket should be used in a batch.
Struct representing a batch of SCPI commands for a RedPitaya. Only commands that interact exclusively with the command socket should be used in a batch.
Executes all commands of the given batch. Returns an array of the results in the order of the commands. An element is nothing if the command has no return value.
Executes all commands of the given batch. Returns an array of the results in the order of the commands.
Each element of the result array is again an array containing the return values of the RedPitayas. An element of an inner array is nothing if the command has no return value.
Executes all commands of the given batch. Returns an array of the results in the order of the commands. An element is nothing if the command has no return value.
Executes all commands of the given batch. Returns an array of the results in the order of the commands.
Each element of the result array is again an array containing the return values of the RedPitayas. An element of an inner array is nothing if the command has no return value.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
Transmit the client-side representation seq to the server and append it to the current list of sequences. Return true if the required commands were successful.
Store calibration DAC offset val for given channel into the RedPitayas EEPROM. This value is used by the server to offset the output voltage. Absolute value has to be smaller than 1.0 V.
As with single RedPitaya. The chan index refers to the total channel available in a cluster, two per RedPitaya. For example channel 4 would refer to the second channel of the second RedPitaya.
Transmit the client-side representation seq to the server and append it to the current list of sequences. Return true if the required commands were successful.
Store calibration DAC offset val for given channel into the RedPitayas EEPROM. This value is used by the server to offset the output voltage. Absolute value has to be smaller than 1.0 V.
For val equals to true this is the same as calling the function on the RedPitaya returned by master(rpc). If val is false then the keepAliveReset is set to true for all RedPitayas in the cluster before the master trigger is disabled. Afterwards the keepAliveReset is set to false again.
Request and receive numOfRequestedSamples samples from wpStart on in a pipelined fashion. The samples and associated PerformanceData are pushed into channel as a SampleChunk.
For val equals to true this is the same as calling the function on the RedPitaya returned by master(rpc). If val is false then the keepAliveReset is set to true for all RedPitayas in the cluster before the master trigger is disabled. Afterwards the keepAliveReset is set to false again.
Request and receive numOfRequestedSamples samples from wpStart on in a pipelined fashion. The samples and associated PerformanceData are pushed into channel as a SampleChunk.
The RedPitayaDAQServer allows to use multiple RedPitayas in a fully synchronized fashion. One of the RedPitayas will act as the master and distribute its clock to all other RedPitayas acting as slaves.
Unfortunately, the STEMlab 125-4 does not allow cluster synchronization without hardware modifications. It is therefore necessary to resolder all slaves according to this documentation. The required mode for this project is 'Directly from FPGA' which requires resistors on R27 and R28. The heatsink has to be removed temporarily in order to unsolder the two resistor below it. In the following image you can see the new position of the 0 Ohm 0402 resistors. Since they get lost easily, make sure you have some in stock.
To run a cluster of RedPitayas one needs to connect the devices using different cables. An exemplary cluster with 3 devices is shown in the following image.
The clock is distributed from the master to the first slave via an SATA cable (green). Additional slaves can be used by connecting the next slave to the previous one. Additionally all slaves have connection from +3.3 Volt to DIO0_N.
In order to send a mutual trigger signal for starting the acquisition and the signal generation, you also have to connect the master's DIO5_P pin with the DIO0_P pin of all devices including the master.
Settings
This document was generated with Documenter.jl version 0.27.24 on Monday 27 March 2023. Using Julia version 1.8.5.
The RedPitayaDAQServer allows to use multiple RedPitayas in a fully synchronized fashion. One of the RedPitayas will act as the master and distribute its clock to all other RedPitayas acting as slaves.
Unfortunately, the STEMlab 125-4 does not allow cluster synchronization without hardware modifications. It is therefore necessary to resolder all slaves according to this documentation. The required mode for this project is 'Directly from FPGA' which requires resistors on R27 and R28. The heatsink has to be removed temporarily in order to unsolder the two resistor below it. In the following image you can see the new position of the 0 Ohm 0402 resistors. Since they get lost easily, make sure you have some in stock.
To run a cluster of RedPitayas one needs to connect the devices using different cables. An exemplary cluster with 3 devices is shown in the following image.
The clock is distributed from the master to the first slave via an SATA cable (green). Additional slaves can be used by connecting the next slave to the previous one. Additionally all slaves have connection from +3.3 Volt to DIO0_N.
In order to send a mutual trigger signal for starting the acquisition and the signal generation, you also have to connect the master's DIO5_P pin with the DIO0_P pin of all devices including the master.
Settings
This document was generated with Documenter.jl version 0.27.24 on Friday 28 April 2023. Using Julia version 1.8.5.
diff --git a/dev/connections.html b/dev/connections.html
index b5d6f004..4a7955f1 100644
--- a/dev/connections.html
+++ b/dev/connections.html
@@ -1,2 +1,2 @@
-Connections · RP DAQ Server
An overview of the extension connectors (see also here) is given in the following image
The project uses most but not all connections that are used in the original RedPitaya image. From the connector E2 only the analog inputs and outputs are used. From the connector E1 several pins are reserved for the following purposes:
DIO0_P for the ADC and DAC trigger. Connect it with the master's DIO5_P to distribute the trigger signal to all RedPitayas in a cluster. As long as the input is high, the DACs and ADCs are running.
DIO1_P is the input for the watchdog (see configuration register section for further details)
DIO2_P is used to acknowledge a received watchdog signal.
DIO3_P can be set to high, to stop all DACs instantly.
DIO4_P outputs a high for 10 ms after a 100 ms pause on low to provide an alive signal.
DIO5_P can be set to high via the configuration register to provide the mutual trigger signal.
DIO7_P, DIO7_N, DIO6_P, DIO6_N, DIO5_N, DIO4_N, DIO3_N, DIO2_N can be used as arbitrary outputs set via the server.
DIO0_N and DIO1_N are used for the clock selection in a cluster.
Settings
This document was generated with Documenter.jl version 0.27.24 on Monday 27 March 2023. Using Julia version 1.8.5.
An overview of the extension connectors (see also here) is given in the following image
The project uses most but not all connections that are used in the original RedPitaya image. From the connector E2 only the analog inputs and outputs are used. From the connector E1 several pins are reserved for the following purposes:
DIO0_P for the ADC and DAC trigger. Connect it with the master's DIO5_P to distribute the trigger signal to all RedPitayas in a cluster. As long as the input is high, the DACs and ADCs are running.
DIO1_P is the input for the watchdog (see configuration register section for further details)
DIO2_P is used to acknowledge a received watchdog signal.
DIO3_P can be set to high, to stop all DACs instantly.
DIO4_P outputs a high for 10 ms after a 100 ms pause on low to provide an alive signal.
DIO5_P can be set to high via the configuration register to provide the mutual trigger signal.
DIO7_P, DIO7_N, DIO6_P, DIO6_N, DIO5_N, DIO4_N, DIO3_N, DIO2_N can be used as arbitrary outputs set via the server.
DIO0_N and DIO1_N are used for the clock selection in a cluster.
Settings
This document was generated with Documenter.jl version 0.27.24 on Friday 28 April 2023. Using Julia version 1.8.5.
diff --git a/dev/devtips.html b/dev/devtips.html
index cdcef40c..a78eba08 100644
--- a/dev/devtips.html
+++ b/dev/devtips.html
@@ -1,2 +1,2 @@
-Development Tips · RP DAQ Server