Accuracy of measurement
The bus load is counted directly by the FPGA in the CAN-Bus Tester 2, in the CANtouch and in the CANobserver. The calculation or counting takes place over the duration of one second including the stuff bits and standardized spaces. The result of this count is highly accurate and reproducible, since it is measured directly on the bus and not after the CAN controller on the logical level. Our measuring instruments are the only ones on the market with this high accuracy.
Depending on the time window with which the bus is looked at, there will be strong differences. Of course, the busload at the moment of transmission is 100%. After the standardized gap to the next telegram, the busload drops to 0% at this moment until another node sends (100%). In order to get a meaningful value, you have to consider a much longer time span.
There are further restrictions when selecting a favorable interval. You don’t want to wait too long for the result, nor do you want to completely suppress temporary fluctuations. However, both would be the case if the counting interval were very large. If the interval is too small, strong fluctuations in short distances of the display are the result, which makes an evaluation or even the reading difficult.
A second as a time base, therefore, seems ideal. Although you don’t have to wait long for an update, the display remains stable or fluctuates only slightly in many real configurations. Theoretically, fluctuations between 0% and 100% are possible, depending on how many bits were counted in one second.
Display fluctuates, what could be the reason ?
In certain bus configurations, rhythmically changing variations in bus load and sampling over one second may cause a fluctuating display, depending on how the tester’s measurement grid (the clock pulse every second) and differences in busload meet.
In sync mode, a master polls some slaves at intervals of 400 ms. The slaves answer in the rhythm of the requests. Once the entire request-response sequence will take place within the measuring second, another time the responses of the slaves will be counted to one or the other second.
The following diagram illustrates this situation. While the same bus load is always counted in the first three seconds, different shares fall on the fourth and fifth seconds. Then the grids temporarily fit on top of the other.
Busload display of the CAN-Bus Tester 2
The CBT2 displays the busload in the last second with an accuracy of one-tenth of a percent. Minimum and maximum values (since the beginning of the measurement) can be recorded. The measurement starts automatically as soon as valid CAN telegrams are detected.
Busload display of the CANtouch
The CANtouch delivers the busload of the last second with an accuracy of one tenth of a percent. In addition to the minimum and maximum values (since the beginning of the measurement), the last minute is also displayed graphically. There is a direct graphical evaluation with our smiley.
Busload display of the CANobserver
After the correct configuration of the baud rate, the CANobserver determines the busload over the last second with an accuracy of one percent. This is displayed in the web interface.
Busload display of the CANalarm
In CANalarm we have no physical measuring device available, the bus load is calculated much less precisely on the firmware side. Here, too, counting takes place over one second. Exceeding a freely definable threshold can be signaled at the switching output.