Ana­ly­se bus­load

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.

Preli­mi­na­ry con­si­de­ra­ti­ons
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.

An examp­le
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 calculation within one second

high-precision CAN busload display in the CBT2 softwareBus­load 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.

high-precision CAN busload display in CANtouchBus­load 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.

Display of busload in the CANobservers webinterfaceBus­load 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.

CAN busload displayBus­load 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.

Sub­scri­be to our news­let­ter now.