industrialNETworX

Hi,

the cycle time is provided by the Managing Node by writing object 0x1006.

The Response Time is related to the time, the Controlled Node needs from the end of the Poll Request frame to start sending of Poll Response frame.

The actual time however is 2us and not as mentioned on Website.

hi,
will hischer release powerlink "managing node" stack firmware? in your schedule?
And,
For a project that contains 8 comx as controled node and a managing node,
MN has 6 bytes to send to each CN one by one, and each CN have 6 bytes to send to MN.
how to estimate the least cyclic time ?
(1.34us+2us+1.34us+2us)*8=53.6us right?

That is to say, Can I use PowerLink to implement a less than 100us cyclic time in this project?

ps:
can RTE Kit analyze the powerlink network?
if do can, is there a application case in the user's manual?

Hi houzenan,

as far as I know is there currently no plan for a "managing node" stack firmware.

Your calculation is wrong, per slave there is at least a 5,7us long frame which requires a interframe gap of 960ns.

The calculation lokks like follows:
6,7us + 2us + 6,7us + 1us at least per slave (16,4us)

that makes up 16,4us * 8 = 131,2us just for the slaves. Not yet accounting the time required for SoC, SoA and the space left for acyclic data exchange.

This makes an additional of at least 30us.

So, you end up with about 160us cycle time as minimum.

Yes, the netANALYZER can also be used for POWERLINK.

Dear AJ
Thank you !!! :D :D :D

I could use comx in the 3 slaves situation, and do not use acyclic data period.
Thus, the minimal cyclic time will be:
16.4*3+10(minimal SoC?)=59.2us

Can it be implemented according to your experiences?

The acyclic data period must exist otherwise the bus will not start up.

the SoC and SoA frame are always there and required. In addition, there must be a specific amount of time behind the SoA frame for being able to Read/Write SDO, Identify Slaves or requesting their status.

The mimimal cyclic time of PowerLink is:
16.4*3+30=79.2us < 80us
It is faster than Sercos III's 250us.
So powerlink is the most fastest stack in Hischer's products ! (After Managing Node FW is developed.)

However, a managing node implementation on the netX would not be capable of handling that entirely. There is a lot more work to do than just copying data.

Just some numbers for comparison on netX100/500 of another protocol:

EtherCAT Slave 30us cycle time with 64 byte each direction EtherCAT Master 1ms

So, you see that on masters the workload is higher than on slaves. Therefore, the same would apply to PowerLink (Managing Node => Controlled Node).
In my opinion, if there is a Managing Node Fw, I would expect something like 250us cycle time due to internal workload required.
Final numbers would have to be examined after such an implementation exists.