die Lösung für Lost-Telegramme-Problem mit FT12 Firmware?

[vannachen]

gibt es hier die Lösungen dafür?
wenn der Bus beschäftigt ist(z.b. mehre switch on/off Aktionen mit Status zurück nehmen), kann die Senden-Aktion nicht richtig durchgeführt wird.
Ich habe auch mit neuest FT1.2 Firmware aus git.selfbus.org die Testen gemacht, kein-Erfolg...

hoffe etwas helfen...

[StefanT]

Hello vannachen!

Ich verstehe die Frage leider nicht.
You can also ask your question in English.

LG
Stefan

[vannachen]

Hello vannachen!

Ich verstehe die Frage leider nicht.
You can also ask your question in English.

LG
Stefan

sorry for my poor deutsch! aber ich kann lesen

that would be fine and kindly!~

yes, i met the problem:

i'm using ft12(lpc+firmware) with eibd under the linux(Ubuntu) of my PC.
and now i can use the command "groupswrite" to open an output of the out8.
But when i use "groupswrite" to open the outputs one by one, and some outputs can not be opened. The problem is sure if we connect with more outputs.
Finally i found that , if i deactive the status of the outputs, all outs can be excuted!

So that means, the actions are not excuted in queue if the bus is so busy!
(one open-telegramm is sent and after that followed the status-telegramm, and the ack-telegramm!)

[StefanT]

I think this is not a problem of the FT1.2 but rather a problem of how it is used. The bus is slow, compared to modern computers. You should wait a moment between sending two telegrams. At least 100ms, maybe more.

[vannachen]

I think this is not a problem of the FT1.2 but rather a problem of how it is used. The bus is slow, compared to modern computers. You should wait a moment between sending two telegrams. At least 100ms, maybe more.

hello stefan,

thanks for your reply!
at first i thought just as you said! but:

1. i change the interface to KNX-USB, and no telegramm lost
2. i change the time-interval between 2 Sends, the result seems better, but when the bus is busy from others such as busy sensors-polling, the result is also bad.

mfg

[StefanT]

Ok, thanks, that's interesting.

I think I can see where in the FT1.2 code this could happen. We will have a look at it, thanks for reporting!

[vannachen]

hello stefan,

I have also checked the code and eibd-ft12 backend, and below are my effects(i'm not sure all, just for the reference)

1. if the eibd-ft12-backend doesn't receive the ACK(0xE5), and the backend repeat the send-frame (frame with variable length) infinitely.

2. If the Send-frame is received OK by LPC-Ft12, and the LPC-Ft12 sends ACK immediately(but the Send-action may not be excuted until followed using send_obj_value(0), but the eibd-backend will not send the frame any more).

3. when the send-action arrived build_tel(), and we want to build the array telegramm[23], but if now the INT1 interrupt active is(that means the status-telegramms for example), the priority is high, so the telegramm-array changed with new value before the send!

mfg

[StefanT]

Ah, that are interesting findings, thanks again for investigating!
Oldcoolman and I already have a look at the code.

[StefanT]

A short status report....

I am still working on the FT1.2 firmware. There are several little problems that could trigger the behaviour. So the topic is still in active bug fixing.

[vannachen]

A short status report....

I am still working on the FT1.2 firmware. There are several little problems that could trigger the behaviour. So the topic is still in active bug fixing.

Hello StefanT,

how is the status now? I have noticed that there is a new function in the lib. How can we use it?