Real Time Properties and measurements
Introduction
IDCP is a Soft Real-Time application. If the internal timing constraints are not held, no catastrophic failure occurs, just position measurement and velocity profile control is degraded.
Nevertheless the application should perform certain tasks in certain time frames.
IOC UniDrive communication
Note
The following text only applies to the UniDrive motor control devices. The MOCON devices are deprecated, but also use the CAN bus and have a similar concept.
The IOC uses a CAN Bus line to communicate with UniDrive motor control devices. Details of this communication are described in uniserv-CAN.
The communication frequency is determined by the CAN Sync object. This is a CAN Bus message sent by the IOC at a fixed rate. All UniDrive devices receive this broadcast message at the same time. They reply with their current position and status. These unicast messages are collected by the IOC. Since the IOC keeps the precise time stamp when the CAN Sync object was sent, it can assign the positions and status values to this time even when all the messages arrive a bit later.
For positions, IDCP implements an extrapolation correction. It takes into account the last known velocity of each axle and uses this with the position taken at the CAN Sync time to extrapolate the position at the current time. This compensates for the delay caused by the sequential reception of all UniDrive positions and the delay for EPICS position measurement record to process. By this mechanism, the IOC calculates all axle positions for the time the position measurement record is processed.
Process variables which are named in the following text have the following variable parts:
- [DEVICE]
This is the device name of the undulator. At the BESSY storage ring, examples for device names are: “UE112ID7R” or “U41IT3R”.
- [VDRIVE]
This is a single letter that depends on the virtual drive. A virtual drive is a group of axles that is moved to one logical destination with a single command. Each undulator has a gap drive and many, but not all undulators have a shift drive. The VDRIVE part of a PV name is “V” for the gap- and “H” for the shift drive (“V” for “vertical”, “H” for “horizontal”).
Note
The position measurement record is [DEVICE]:NdiPr[VDRIVE]PmsCnv.
CAN Sync Object timing
The CAN Sync Object is sent by a CAN Bus record which is processed at a fixed rate.
Note
Actually record [DEVICE]:AdiAllPmsSynCnt is processed at a fixed rate. It processes via EPICS forward link [DEVICE]:AdiAllPmsSynTrg. This processes [DEVICE]:AdiUnAllGblPmsCnt which processes [DEVICE]:AdiUnAllGblSyn. The last record is the one the sends the CAN Sync object.
The CAN Sync interval is defined in the IDCP parameter database in parameter id-data.ID.config.pos_time_interval. Currently, as of 2026-06-26 it is 0.05s for all undulators except the UE112 and UE56-1 undulator, where it is 0.1s.
The goal here is to send the CAN sync object at a precise time frame with minimal jitter.
How to measure the CAN Sync Interval
The device support used by the IOC for CAN Bus communication, MultiCAN, provides a function to get a time stamp when a CAN object was written. This time stamp is a 64 bit integer with a resolution of one microsecond.
With the Linux implementation of MultiCAN, the time stamp function is implemented in the socan device support.
Note
For the mvme2100 CPU board, the timestamp doesn’t count microseconds but 600/625 microseconds. This must be taken into account when calculating times measured on these systems.
IDCP presents this timestamp for the CAN Sync object in two EPICS PVs:
[DEVICE]:NdiPrVPmsCnv.VALU : The upper 32 bits of the time stamp
[DEVICE]:NdiPrVPmsCnv.VALT : The lower 32 bits of the time stamp
By monitoring these PVs with a simple camonitor command like this:
camonitor [DEVICE]:NdiPrVPmsCnv.VALU [DEVICE]:NdiPrVPmsCnv.VALT > LOGFILE
we can observe the timestamps for longer periods of time.
With Linux we have a second method of measuring time stamps by running candump from package can-utils., assuming that the CAN interface is named “can2”:
candump -t A can2,080:7FF
Measurements
The statistical evaluation was made for the differences between two CAN Sync Timestamps. This difference should be 50ms exactly.
All times that follow have the unit seconds (s).
The measurement method is either:
camonitor: when EPICS PVs with timestamps were used
candump: when the candump program of Linux was used
RTEMS with mvme2100 CPU board
System properties
property |
value |
|---|---|
Device |
UE51 |
operating system |
RTEMS |
kernel version |
4.9 |
CPU |
mvme2100 |
Measurement over 24 hours
property |
value |
|---|---|
measurement method |
camonitor |
start of measurement |
2026-06-22 12:00:00 |
end of measurement |
2026-06-23 12:00:00 |
overall time |
23:59:59.902206 |
number of samples/lines |
1727993 |
absolute minimum |
0.039830 |
absolute maximum |
0.100109 |
max. deviation from mean |
0.050109 |
mean |
0.050000 |
standard deviation |
1.761342e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.039500 to 0.040500 |
1 |
0.00 |
0.040500 to 0.041500 |
0 |
0.00 |
0.041500 to 0.042500 |
0 |
0.00 |
0.042500 to 0.043500 |
0 |
0.00 |
0.043500 to 0.044500 |
0 |
0.00 |
0.044500 to 0.045500 |
0 |
0.00 |
0.045500 to 0.046500 |
0 |
0.00 |
0.046500 to 0.047500 |
0 |
0.00 |
0.047500 to 0.048500 |
0 |
0.00 |
0.048500 to 0.049500 |
465 |
0.03 |
0.049500 to 0.050500 |
1726728 |
99.93 |
0.050500 to 0.051500 |
453 |
0.03 |
0.051500 to 0.052500 |
0 |
0.00 |
0.052500 to 0.053500 |
0 |
0.00 |
0.053500 to 0.054500 |
0 |
0.00 |
0.054500 to 0.055500 |
0 |
0.00 |
0.055500 to 0.056500 |
0 |
0.00 |
0.056500 to 0.057500 |
0 |
0.00 |
0.057500 to 0.058500 |
0 |
0.00 |
0.058500 to 0.059500 |
0 |
0.00 |
0.059500 to 0.060500 |
342 |
0.02 |
0.060500 to 0.061500 |
0 |
0.00 |
0.061500 to 0.062500 |
0 |
0.00 |
0.062500 to 0.063500 |
0 |
0.00 |
0.063500 to 0.064500 |
0 |
0.00 |
0.064500 to 0.065500 |
0 |
0.00 |
0.065500 to 0.066500 |
0 |
0.00 |
0.066500 to 0.067500 |
0 |
0.00 |
0.067500 to 0.068500 |
0 |
0.00 |
0.068500 to 0.069500 |
0 |
0.00 |
0.069500 to 0.070500 |
0 |
0.00 |
0.070500 to 0.071500 |
0 |
0.00 |
0.071500 to 0.072500 |
0 |
0.00 |
0.072500 to 0.073500 |
0 |
0.00 |
0.073500 to 0.074500 |
0 |
0.00 |
0.074500 to 0.075500 |
0 |
0.00 |
0.075500 to 0.076500 |
0 |
0.00 |
0.076500 to 0.077500 |
0 |
0.00 |
0.077500 to 0.078500 |
0 |
0.00 |
0.078500 to 0.079500 |
0 |
0.00 |
0.079500 to 0.080500 |
0 |
0.00 |
0.080500 to 0.081500 |
0 |
0.00 |
0.081500 to 0.082500 |
0 |
0.00 |
0.082500 to 0.083500 |
0 |
0.00 |
0.083500 to 0.084500 |
0 |
0.00 |
0.084500 to 0.085500 |
0 |
0.00 |
0.085500 to 0.086500 |
0 |
0.00 |
0.086500 to 0.087500 |
0 |
0.00 |
0.087500 to 0.088500 |
0 |
0.00 |
0.088500 to 0.089500 |
0 |
0.00 |
0.089500 to 0.090500 |
0 |
0.00 |
0.090500 to 0.091500 |
0 |
0.00 |
0.091500 to 0.092500 |
0 |
0.00 |
0.092500 to 0.093500 |
0 |
0.00 |
0.093500 to 0.094500 |
0 |
0.00 |
0.094500 to 0.095500 |
0 |
0.00 |
0.095500 to 0.096500 |
0 |
0.00 |
0.096500 to 0.097500 |
0 |
0.00 |
0.097500 to 0.098500 |
0 |
0.00 |
0.098500 to 0.099500 |
0 |
0.00 |
0.099500 to 0.100500 |
4 |
0.00 |
Measurement over 18 days
property |
value |
|---|---|
measurement method |
camonitor |
start of measurement |
2026-06-01 00:00:00 |
end of measurement |
2026-06-19 14:00:00 |
overall time |
18 days, 15:28:24.503000 |
number of samples/lines |
30929062 |
absolute minimum |
0.011896 |
absolute maximum |
0.150005 |
max. deviation from mean |
0.100004 |
mean |
0.050000 |
standard deviation |
2.256304e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.011500 to 0.012500 |
1 |
0.00 |
0.012500 to 0.013500 |
1 |
0.00 |
0.013500 to 0.014500 |
1 |
0.00 |
0.014500 to 0.015500 |
0 |
0.00 |
0.015500 to 0.016500 |
1 |
0.00 |
0.016500 to 0.017500 |
0 |
0.00 |
0.017500 to 0.018500 |
2 |
0.00 |
0.018500 to 0.019500 |
0 |
0.00 |
0.019500 to 0.020500 |
3 |
0.00 |
0.020500 to 0.021500 |
0 |
0.00 |
0.021500 to 0.022500 |
0 |
0.00 |
0.022500 to 0.023500 |
0 |
0.00 |
0.023500 to 0.024500 |
0 |
0.00 |
0.024500 to 0.025500 |
0 |
0.00 |
0.025500 to 0.026500 |
0 |
0.00 |
0.026500 to 0.027500 |
0 |
0.00 |
0.027500 to 0.028500 |
0 |
0.00 |
0.028500 to 0.029500 |
0 |
0.00 |
0.029500 to 0.030500 |
0 |
0.00 |
0.030500 to 0.031500 |
0 |
0.00 |
0.031500 to 0.032500 |
0 |
0.00 |
0.032500 to 0.033500 |
0 |
0.00 |
0.033500 to 0.034500 |
0 |
0.00 |
0.034500 to 0.035500 |
0 |
0.00 |
0.035500 to 0.036500 |
0 |
0.00 |
0.036500 to 0.037500 |
0 |
0.00 |
0.037500 to 0.038500 |
0 |
0.00 |
0.038500 to 0.039500 |
3 |
0.00 |
0.039500 to 0.040500 |
5 |
0.00 |
0.040500 to 0.041500 |
0 |
0.00 |
0.041500 to 0.042500 |
0 |
0.00 |
0.042500 to 0.043500 |
0 |
0.00 |
0.043500 to 0.044500 |
0 |
0.00 |
0.044500 to 0.045500 |
0 |
0.00 |
0.045500 to 0.046500 |
0 |
0.00 |
0.046500 to 0.047500 |
1 |
0.00 |
0.047500 to 0.048500 |
24 |
0.00 |
0.048500 to 0.049500 |
12973 |
0.04 |
0.049500 to 0.050500 |
30897546 |
99.90 |
0.050500 to 0.051500 |
12080 |
0.04 |
0.051500 to 0.052500 |
25 |
0.00 |
0.052500 to 0.053500 |
1 |
0.00 |
0.053500 to 0.054500 |
0 |
0.00 |
0.054500 to 0.055500 |
0 |
0.00 |
0.055500 to 0.056500 |
0 |
0.00 |
0.056500 to 0.057500 |
0 |
0.00 |
0.057500 to 0.058500 |
0 |
0.00 |
0.058500 to 0.059500 |
2 |
0.00 |
0.059500 to 0.060500 |
6099 |
0.02 |
0.060500 to 0.061500 |
2 |
0.00 |
0.061500 to 0.062500 |
0 |
0.00 |
0.062500 to 0.063500 |
0 |
0.00 |
0.063500 to 0.064500 |
0 |
0.00 |
0.064500 to 0.065500 |
0 |
0.00 |
0.065500 to 0.066500 |
0 |
0.00 |
0.066500 to 0.067500 |
0 |
0.00 |
0.067500 to 0.068500 |
0 |
0.00 |
0.068500 to 0.069500 |
0 |
0.00 |
0.069500 to 0.070500 |
1 |
0.00 |
0.070500 to 0.071500 |
0 |
0.00 |
0.071500 to 0.072500 |
0 |
0.00 |
0.072500 to 0.073500 |
0 |
0.00 |
0.073500 to 0.074500 |
0 |
0.00 |
0.074500 to 0.075500 |
0 |
0.00 |
0.075500 to 0.076500 |
0 |
0.00 |
0.076500 to 0.077500 |
0 |
0.00 |
0.077500 to 0.078500 |
0 |
0.00 |
0.078500 to 0.079500 |
0 |
0.00 |
0.079500 to 0.080500 |
0 |
0.00 |
0.080500 to 0.081500 |
0 |
0.00 |
0.081500 to 0.082500 |
0 |
0.00 |
0.082500 to 0.083500 |
0 |
0.00 |
0.083500 to 0.084500 |
0 |
0.00 |
0.084500 to 0.085500 |
0 |
0.00 |
0.085500 to 0.086500 |
0 |
0.00 |
0.086500 to 0.087500 |
0 |
0.00 |
0.087500 to 0.088500 |
0 |
0.00 |
0.088500 to 0.089500 |
0 |
0.00 |
0.089500 to 0.090500 |
0 |
0.00 |
0.090500 to 0.091500 |
0 |
0.00 |
0.091500 to 0.092500 |
0 |
0.00 |
0.092500 to 0.093500 |
0 |
0.00 |
0.093500 to 0.094500 |
0 |
0.00 |
0.094500 to 0.095500 |
0 |
0.00 |
0.095500 to 0.096500 |
0 |
0.00 |
0.096500 to 0.097500 |
0 |
0.00 |
0.097500 to 0.098500 |
0 |
0.00 |
0.098500 to 0.099500 |
0 |
0.00 |
0.099500 to 0.100500 |
272 |
0.00 |
0.100500 to 0.101500 |
4 |
0.00 |
0.101500 to 0.102500 |
0 |
0.00 |
0.102500 to 0.103500 |
0 |
0.00 |
0.103500 to 0.104500 |
0 |
0.00 |
0.104500 to 0.105500 |
0 |
0.00 |
0.105500 to 0.106500 |
0 |
0.00 |
0.106500 to 0.107500 |
0 |
0.00 |
0.107500 to 0.108500 |
0 |
0.00 |
0.108500 to 0.109500 |
0 |
0.00 |
0.109500 to 0.110500 |
1 |
0.00 |
0.110500 to 0.111500 |
3 |
0.00 |
0.111500 to 0.112500 |
1 |
0.00 |
0.112500 to 0.113500 |
0 |
0.00 |
0.113500 to 0.114500 |
0 |
0.00 |
0.114500 to 0.115500 |
0 |
0.00 |
0.115500 to 0.116500 |
0 |
0.00 |
0.116500 to 0.117500 |
0 |
0.00 |
0.117500 to 0.118500 |
0 |
0.00 |
0.118500 to 0.119500 |
0 |
0.00 |
0.119500 to 0.120500 |
0 |
0.00 |
0.120500 to 0.121500 |
0 |
0.00 |
0.121500 to 0.122500 |
0 |
0.00 |
0.122500 to 0.123500 |
0 |
0.00 |
0.123500 to 0.124500 |
1 |
0.00 |
0.124500 to 0.125500 |
0 |
0.00 |
0.125500 to 0.126500 |
1 |
0.00 |
0.126500 to 0.127500 |
1 |
0.00 |
0.127500 to 0.128500 |
1 |
0.00 |
0.128500 to 0.129500 |
0 |
0.00 |
0.129500 to 0.130500 |
2 |
0.00 |
0.130500 to 0.131500 |
0 |
0.00 |
0.131500 to 0.132500 |
1 |
0.00 |
0.132500 to 0.133500 |
0 |
0.00 |
0.133500 to 0.134500 |
0 |
0.00 |
0.134500 to 0.135500 |
0 |
0.00 |
0.135500 to 0.136500 |
0 |
0.00 |
0.136500 to 0.137500 |
0 |
0.00 |
0.137500 to 0.138500 |
0 |
0.00 |
0.138500 to 0.139500 |
0 |
0.00 |
0.139500 to 0.140500 |
0 |
0.00 |
0.140500 to 0.141500 |
0 |
0.00 |
0.141500 to 0.142500 |
0 |
0.00 |
0.142500 to 0.143500 |
0 |
0.00 |
0.143500 to 0.144500 |
0 |
0.00 |
0.144500 to 0.145500 |
0 |
0.00 |
0.145500 to 0.146500 |
0 |
0.00 |
0.146500 to 0.147500 |
0 |
0.00 |
0.147500 to 0.148500 |
0 |
0.00 |
0.148500 to 0.149500 |
0 |
0.00 |
0.149500 to 0.150500 |
3 |
0.00 |
Linux with Debian 12 (bookworm) and a desktop PC
System properties
property |
value |
|---|---|
Device |
UE48 |
operating system |
Linux Debian 12 (bookworm) |
kernel version |
6.1.76-1 |
CPU |
Intel(R) Core(TM) i5-3570 CPU @ 3.40GHz |
Measurement
property |
value |
|---|---|
measurement method |
candump |
start of measurement |
2026-05-16 12:00:00 |
end of measurement |
2026-05-17 12:00:00 |
overall time |
23:59:59.950126 |
number of samples/lines |
1727999 |
absolute minimum |
0.049405 |
absolute maximum |
0.050572 |
max. deviation from mean |
0.000595 |
mean |
0.050000 |
standard deviation |
1.151468e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.048500 to 0.049500 |
7 |
0.00 |
0.049500 to 0.050500 |
1727985 |
100.00 |
0.050500 to 0.051500 |
7 |
0.00 |
Linux with Debian 11 (bullseye), Industrial PC regular kernel
System properties
property |
value |
|---|---|
Device |
UE48 |
operating system |
Linux Debian 11 (bullseye) |
kernel version |
6.1.90-1~bpo11+1 |
kernel package |
linux-image-6.1.0-0.deb11.21-amd64 |
CPU |
Intel(R) Core(TM) i3-14100 |
Measurement 1 (candump)
property |
value |
|---|---|
measurement method |
candump |
start of measurement |
2026-06-17 12:00:00 |
end of measurement |
2026-06-18 12:00:00 |
overall time |
23:59:59.949983 |
number of samples/lines |
1727999 |
absolute minimum |
0.048268 |
absolute maximum |
0.051803 |
max. deviation from mean |
0.001803 |
mean |
0.050000 |
standard deviation |
3.478053e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.047500 to 0.048500 |
3310 |
0.19 |
0.048500 to 0.049500 |
82322 |
4.76 |
0.049500 to 0.050500 |
1557417 |
90.13 |
0.050500 to 0.051500 |
82315 |
4.76 |
0.051500 to 0.052500 |
2635 |
0.15 |
Measurement 2 (camonitor)
property |
value |
|---|---|
measurement method |
camonitor |
overall time |
23:59:59.899991 |
number of samples/lines |
1727998 |
absolute minimum |
0.048268 |
absolute maximum |
0.051803 |
max. deviation from mean |
0.001803 |
mean |
0.050000 |
standard deviation |
3.478057e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.047500 to 0.048500 |
3315 |
0.19 |
0.048500 to 0.049500 |
82318 |
4.76 |
0.049500 to 0.050500 |
1557411 |
90.13 |
0.050500 to 0.051500 |
82311 |
4.76 |
0.051500 to 0.052500 |
2643 |
0.15 |
Measurement over 19 days
property |
value |
|---|---|
measurement method |
camonitor |
overall time |
19 days, 0:59:58.899987 |
number of samples/lines |
32903978 |
absolute minimum |
0.047673 |
absolute maximum |
0.052580 |
max. deviation from mean |
0.002580 |
mean |
0.050000 |
standard deviation |
3.484851e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.047500 to 0.048500 |
60153 |
0.18 |
0.048500 to 0.049500 |
1573348 |
4.78 |
0.049500 to 0.050500 |
29647150 |
90.10 |
0.050500 to 0.051500 |
1575115 |
4.79 |
0.051500 to 0.052500 |
48211 |
0.15 |
0.052500 to 0.053500 |
1 |
0.00 |
Linux with Debian 11 (bullseye), Industrial PC realtime kernel
System properties
property |
value |
|---|---|
Device |
UE48 |
operating system |
Linux Debian 11 (bullseye) |
kernel version |
5.10.257-1 |
kernel package |
linux-image-rt-amd64 |
CPU |
Intel(R) Core(TM) i3-14100 |
Measurement
property |
value |
|---|---|
measurement method |
camonitor |
start of measurement |
2026-06-22 12:00:00 |
end of measurement |
2026-06-23 12:00:00 |
overall time |
23:59:59.900000 |
number of samples/lines |
1727998 |
absolute minimum |
0.047862 |
absolute maximum |
0.051794 |
max. deviation from mean |
0.002138 |
mean |
0.050000 |
standard deviation |
3.520560e-04 |
Histogram Table:
Range |
Count |
% |
|---|---|---|
0.047500 to 0.048500 |
3183 |
0.18 |
0.048500 to 0.049500 |
82489 |
4.77 |
0.049500 to 0.050500 |
1557365 |
90.13 |
0.050500 to 0.051500 |
82095 |
4.75 |
0.051500 to 0.052500 |
2866 |
0.17 |
Diskussion
Measurement methods
The two measurements of the Linux bullseye system with the industrial PC and regular kernel show that the ‘candump’ and ‘camonitor’ measurements give almost identical results. It follows that measurements can be compared regardless of the measurement method.
Statistical parameters comparison, 24 hours measurement
property |
RTEMS |
bookworm |
bullseye |
bullseye(rt) |
|---|---|---|---|---|
overall time |
24 hours |
24 hours |
24 hours |
24 hours |
absolute minimum |
0.039830 |
0.049405 |
0.048268 |
0.047862 |
absolute maximum |
0.100109 |
0.050572 |
0.051803 |
0.051794 |
max. deviation from mean |
0.050109 |
0.000595 |
0.001803 |
0.002138 |
mean |
0.050000 |
0.050000 |
0.050000 |
0.050000 |
standard deviation |
1.761342e-04 |
1.151468e-04 |
3.478053e-04 |
3.520560e-04 |
Statistical parameters comparison, 18 days measurement
property |
RTEMS |
bullseye |
|---|---|---|
overall time |
18 days, 15:28:24.503000 |
19 days, 0:59:58.899987 |
absolute minimum |
0.011896 |
0.047673 |
absolute maximum |
0.150005 |
0.052580 |
max. deviation from mean |
0.100004 |
0.002580 |
mean |
0.050000 |
0.050000 |
standard deviation |
2.256304e-04 |
3.484851e-04 |
Histogram comparison, 24 hours measurement
Here we see how the timing histograms of the tested systems compare. We only show the counts, not percentages:
Range |
RTEMS |
bookworm |
bullseye |
bullseye(rt) |
|---|---|---|---|---|
0.039500 to 0.040500 |
1 |
0 |
0 |
0 |
0.040500 to 0.041500 |
0 |
0 |
0 |
0 |
0.041500 to 0.042500 |
0 |
0 |
0 |
0 |
0.042500 to 0.043500 |
0 |
0 |
0 |
0 |
0.043500 to 0.044500 |
0 |
0 |
0 |
0 |
0.044500 to 0.045500 |
0 |
0 |
0 |
0 |
0.045500 to 0.046500 |
0 |
0 |
0 |
0 |
0.046500 to 0.047500 |
0 |
0 |
0 |
0 |
0.047500 to 0.048500 |
0 |
0 |
3310 |
3183 |
0.048500 to 0.049500 |
465 |
7 |
82322 |
82489 |
0.049500 to 0.050500 |
1726728 |
1727985 |
1557417 |
1557365 |
0.050500 to 0.051500 |
453 |
7 |
82315 |
82095 |
0.051500 to 0.052500 |
0 |
0 |
2635 |
2866 |
0.052500 to 0.053500 |
0 |
0 |
0 |
0 |
0.053500 to 0.054500 |
0 |
0 |
0 |
0 |
0.054500 to 0.055500 |
0 |
0 |
0 |
0 |
0.055500 to 0.056500 |
0 |
0 |
0 |
0 |
0.056500 to 0.057500 |
0 |
0 |
0 |
0 |
0.057500 to 0.058500 |
0 |
0 |
0 |
0 |
0.058500 to 0.059500 |
0 |
0 |
0 |
0 |
0.059500 to 0.060500 |
342 |
0 |
0 |
0 |
0.060500 to 0.061500 |
0 |
0 |
0 |
0 |
0.061500 to 0.062500 |
0 |
0 |
0 |
0 |
0.062500 to 0.063500 |
0 |
0 |
0 |
0 |
0.063500 to 0.064500 |
0 |
0 |
0 |
0 |
0.064500 to 0.065500 |
0 |
0 |
0 |
0 |
0.065500 to 0.066500 |
0 |
0 |
0 |
0 |
0.066500 to 0.067500 |
0 |
0 |
0 |
0 |
0.067500 to 0.068500 |
0 |
0 |
0 |
0 |
0.068500 to 0.069500 |
0 |
0 |
0 |
0 |
0.069500 to 0.070500 |
0 |
0 |
0 |
0 |
0.070500 to 0.071500 |
0 |
0 |
0 |
0 |
0.071500 to 0.072500 |
0 |
0 |
0 |
0 |
0.072500 to 0.073500 |
0 |
0 |
0 |
0 |
0.073500 to 0.074500 |
0 |
0 |
0 |
0 |
0.074500 to 0.075500 |
0 |
0 |
0 |
0 |
0.075500 to 0.076500 |
0 |
0 |
0 |
0 |
0.076500 to 0.077500 |
0 |
0 |
0 |
0 |
0.077500 to 0.078500 |
0 |
0 |
0 |
0 |
0.078500 to 0.079500 |
0 |
0 |
0 |
0 |
0.079500 to 0.080500 |
0 |
0 |
0 |
0 |
0.080500 to 0.081500 |
0 |
0 |
0 |
0 |
0.081500 to 0.082500 |
0 |
0 |
0 |
0 |
0.082500 to 0.083500 |
0 |
0 |
0 |
0 |
0.083500 to 0.084500 |
0 |
0 |
0 |
0 |
0.084500 to 0.085500 |
0 |
0 |
0 |
0 |
0.085500 to 0.086500 |
0 |
0 |
0 |
0 |
0.086500 to 0.087500 |
0 |
0 |
0 |
0 |
0.087500 to 0.088500 |
0 |
0 |
0 |
0 |
0.088500 to 0.089500 |
0 |
0 |
0 |
0 |
0.089500 to 0.090500 |
0 |
0 |
0 |
0 |
0.090500 to 0.091500 |
0 |
0 |
0 |
0 |
0.091500 to 0.092500 |
0 |
0 |
0 |
0 |
0.092500 to 0.093500 |
0 |
0 |
0 |
0 |
0.093500 to 0.094500 |
0 |
0 |
0 |
0 |
0.094500 to 0.095500 |
0 |
0 |
0 |
0 |
0.095500 to 0.096500 |
0 |
0 |
0 |
0 |
0.096500 to 0.097500 |
0 |
0 |
0 |
0 |
0.097500 to 0.098500 |
0 |
0 |
0 |
0 |
0.098500 to 0.099500 |
0 |
0 |
0 |
0 |
0.099500 to 0.100500 |
4 |
0 |
0 |
0 |
Histogram comparison, 18 days measurement
Range |
RTEMS |
bullseye |
|---|---|---|
0.011500 to 0.012500 |
1 |
0 |
0.012500 to 0.013500 |
1 |
0 |
0.013500 to 0.014500 |
1 |
0 |
0.014500 to 0.015500 |
0 |
0 |
0.015500 to 0.016500 |
1 |
0 |
0.016500 to 0.017500 |
0 |
0 |
0.017500 to 0.018500 |
2 |
0 |
0.018500 to 0.019500 |
0 |
0 |
0.019500 to 0.020500 |
3 |
0 |
0.020500 to 0.021500 |
0 |
0 |
0.021500 to 0.022500 |
0 |
0 |
0.022500 to 0.023500 |
0 |
0 |
0.023500 to 0.024500 |
0 |
0 |
0.024500 to 0.025500 |
0 |
0 |
0.025500 to 0.026500 |
0 |
0 |
0.026500 to 0.027500 |
0 |
0 |
0.027500 to 0.028500 |
0 |
0 |
0.028500 to 0.029500 |
0 |
0 |
0.029500 to 0.030500 |
0 |
0 |
0.030500 to 0.031500 |
0 |
0 |
0.031500 to 0.032500 |
0 |
0 |
0.032500 to 0.033500 |
0 |
0 |
0.033500 to 0.034500 |
0 |
0 |
0.034500 to 0.035500 |
0 |
0 |
0.035500 to 0.036500 |
0 |
0 |
0.036500 to 0.037500 |
0 |
0 |
0.037500 to 0.038500 |
0 |
0 |
0.038500 to 0.039500 |
3 |
0 |
0.039500 to 0.040500 |
5 |
0 |
0.040500 to 0.041500 |
0 |
0 |
0.041500 to 0.042500 |
0 |
0 |
0.042500 to 0.043500 |
0 |
0 |
0.043500 to 0.044500 |
0 |
0 |
0.044500 to 0.045500 |
0 |
0 |
0.045500 to 0.046500 |
0 |
0 |
0.046500 to 0.047500 |
1 |
0 |
0.047500 to 0.048500 |
24 |
60153 |
0.048500 to 0.049500 |
12973 |
1573348 |
0.049500 to 0.050500 |
30897546 |
29647150 |
0.050500 to 0.051500 |
12080 |
1575115 |
0.051500 to 0.052500 |
25 |
48211 |
0.052500 to 0.053500 |
1 |
1 |
0.053500 to 0.054500 |
0 |
0 |
0.054500 to 0.055500 |
0 |
0 |
0.055500 to 0.056500 |
0 |
0 |
0.056500 to 0.057500 |
0 |
0 |
0.057500 to 0.058500 |
0 |
0 |
0.058500 to 0.059500 |
2 |
0 |
0.059500 to 0.060500 |
6099 |
0 |
0.060500 to 0.061500 |
2 |
0 |
0.061500 to 0.062500 |
0 |
0 |
0.062500 to 0.063500 |
0 |
0 |
0.063500 to 0.064500 |
0 |
0 |
0.064500 to 0.065500 |
0 |
0 |
0.065500 to 0.066500 |
0 |
0 |
0.066500 to 0.067500 |
0 |
0 |
0.067500 to 0.068500 |
0 |
0 |
0.068500 to 0.069500 |
0 |
0 |
0.069500 to 0.070500 |
1 |
0 |
0.070500 to 0.071500 |
0 |
0 |
0.071500 to 0.072500 |
0 |
0 |
0.072500 to 0.073500 |
0 |
0 |
0.073500 to 0.074500 |
0 |
0 |
0.074500 to 0.075500 |
0 |
0 |
0.075500 to 0.076500 |
0 |
0 |
0.076500 to 0.077500 |
0 |
0 |
0.077500 to 0.078500 |
0 |
0 |
0.078500 to 0.079500 |
0 |
0 |
0.079500 to 0.080500 |
0 |
0 |
0.080500 to 0.081500 |
0 |
0 |
0.081500 to 0.082500 |
0 |
0 |
0.082500 to 0.083500 |
0 |
0 |
0.083500 to 0.084500 |
0 |
0 |
0.084500 to 0.085500 |
0 |
0 |
0.085500 to 0.086500 |
0 |
0 |
0.086500 to 0.087500 |
0 |
0 |
0.087500 to 0.088500 |
0 |
0 |
0.088500 to 0.089500 |
0 |
0 |
0.089500 to 0.090500 |
0 |
0 |
0.090500 to 0.091500 |
0 |
0 |
0.091500 to 0.092500 |
0 |
0 |
0.092500 to 0.093500 |
0 |
0 |
0.093500 to 0.094500 |
0 |
0 |
0.094500 to 0.095500 |
0 |
0 |
0.095500 to 0.096500 |
0 |
0 |
0.096500 to 0.097500 |
0 |
0 |
0.097500 to 0.098500 |
0 |
0 |
0.098500 to 0.099500 |
0 |
0 |
0.099500 to 0.100500 |
272 |
0 |
0.100500 to 0.101500 |
4 |
0 |
0.101500 to 0.102500 |
0 |
0 |
0.102500 to 0.103500 |
0 |
0 |
0.103500 to 0.104500 |
0 |
0 |
0.104500 to 0.105500 |
0 |
0 |
0.105500 to 0.106500 |
0 |
0 |
0.106500 to 0.107500 |
0 |
0 |
0.107500 to 0.108500 |
0 |
0 |
0.108500 to 0.109500 |
0 |
0 |
0.109500 to 0.110500 |
1 |
0 |
0.110500 to 0.111500 |
3 |
0 |
0.111500 to 0.112500 |
1 |
0 |
0.112500 to 0.113500 |
0 |
0 |
0.113500 to 0.114500 |
0 |
0 |
0.114500 to 0.115500 |
0 |
0 |
0.115500 to 0.116500 |
0 |
0 |
0.116500 to 0.117500 |
0 |
0 |
0.117500 to 0.118500 |
0 |
0 |
0.118500 to 0.119500 |
0 |
0 |
0.119500 to 0.120500 |
0 |
0 |
0.120500 to 0.121500 |
0 |
0 |
0.121500 to 0.122500 |
0 |
0 |
0.122500 to 0.123500 |
0 |
0 |
0.123500 to 0.124500 |
1 |
0 |
0.124500 to 0.125500 |
0 |
0 |
0.125500 to 0.126500 |
1 |
0 |
0.126500 to 0.127500 |
1 |
0 |
0.127500 to 0.128500 |
1 |
0 |
0.128500 to 0.129500 |
0 |
0 |
0.129500 to 0.130500 |
2 |
0 |
0.130500 to 0.131500 |
0 |
0 |
0.131500 to 0.132500 |
1 |
0 |
0.132500 to 0.133500 |
0 |
0 |
0.133500 to 0.134500 |
0 |
0 |
0.134500 to 0.135500 |
0 |
0 |
0.135500 to 0.136500 |
0 |
0 |
0.136500 to 0.137500 |
0 |
0 |
0.137500 to 0.138500 |
0 |
0 |
0.138500 to 0.139500 |
0 |
0 |
0.139500 to 0.140500 |
0 |
0 |
0.140500 to 0.141500 |
0 |
0 |
0.141500 to 0.142500 |
0 |
0 |
0.142500 to 0.143500 |
0 |
0 |
0.143500 to 0.144500 |
0 |
0 |
0.144500 to 0.145500 |
0 |
0 |
0.145500 to 0.146500 |
0 |
0 |
0.146500 to 0.147500 |
0 |
0 |
0.147500 to 0.148500 |
0 |
0 |
0.148500 to 0.149500 |
0 |
0 |
0.149500 to 0.150500 |
3 |
0 |
Conclusion
RTEMS, not as good as expected
Although RTEMS has a very good standard deviation of 0.23ms, some extreme deviations from the mean value are concerning. Over 24 hours we saw a maximum jitter of 50ms and over 18 days we saw even 100ms. This is longer than the mean difference between two CAN sync objects. So in the first case, one CAN Sync was lost, and in the second case even two were lost.
Linux, very good but questions on details remain
One positive aspect of Linux systems is that the maximum deviation from the mean value even over 19 days was never greater that 2.6ms.
Some questions remain if we look more closely at the performance:
The real time (rt) kernel of version 5 doesn’t perform better than the normal kernel of version 6. Using the real time kernel makes almost no difference
The bookworm system performed measurably better than the bullseye system, even when both run with a kernel of version 6. Therefor more tests should be made with Debian bookworm or even the more recent Debian Trixie.
Linux versus RTEMS
The measurements suggest that a PC with Linux performs better than the VME Bus based RTEMS system with a PowerPC CPU regarding the maximum deviations from the timing schedule.
This indicates that replacing RTEMS VME CPUs with Linux Industrial PCs is the future for this project.