Mr. Schildknecht, 5G is currently in the introductory phase - at least if we are to believe the statements of the mobile network providers. How quickly will we get 5G in the factory?

Thomas Schildknecht: In mid-July, I attended the VDI conference 5G in Automation in Baden-Baden. The mood among providers and potential users was very informative. And there was a very important block of topics at the conference, particularly with regard to the availability of 5G in industry. This dealt with the technical details of the parameters of the uRLLC - ultra-Reliable Low-Latency Communication - and mMTC - massive Machine-Type Communication - services, as well as the possibility of setting up separate campus networks. Interested end users were very disappointed to learn that these services will not be defined until Release 16, which is expected in 2020. Only then can the baseband manufacturers begin with the final specification of the associated chips! And it will be another two to three years before the first integration-capable OEM modules come onto the market. As a result, 5G solutions for automation devices will probably not be available until 2023!

Do the mobile network providers and automation specialists at least speak the same language?

Thomas Schildknecht: It would be nice! For example, there are fundamental misunderstandings about latency. This term is used by mobile network providers and network equipment suppliers to describe the transmission time of a data packet via 5G in campus networks with 1 millisecond as the target value for the upcoming Release 16. However, this term has a different meaning for automation engineers - in connection with fieldbus transmissions via Profinet and/or in future via TSN, for example. Incidentally, the latency for WLAN is less than 1 millisecond in undisturbed operation, but this does not mean that Profinet can be realized with an update time of 1 millisecond.

In fieldbus technology, we use latency as a term for the fieldbus update time. In wired real-time applications, the variance of this value - the jitter - may only be a few microseconds for synchronized drives; even with 5G URLLC, however, it will be a few hundred microseconds!

The frequency range that can be used by campus networks should use the TDD principle, according to the Federal Network Agency's current plans. In other words, a kind of half-duplex timeslot procedure. However, a half-duplex data connection is counterproductive for low target latency times. Automation specialists are still familiar with this problem from Profibus, which also worked half-duplex. In contrast, the successor Profinet was introduced, where data can be sent and received simultaneously, i.e. full duplex. TDD is likely to result in latency times that are permanently several orders of magnitude higher than the target values.

"5G solutions for automation devices are not expected before 2023," says Thomas Schildknecht, CEO of Schildknecht AG.

© Schildknecht

What important tasks still need to be solved?

Thomas Schildknecht: To fully achieve the desired 5G parameters, further technologies such as beamforming, MIMO antennas or the use of millimeter waves >30 GHz still need to be developed, to name just a few examples. It is therefore to be expected that the actual and required performance of 5G will only be fully realized and ready for use in a few years' time. Until then, the task of automation specialists will be to identify those use cases for which 5G offers sufficient performance from Release 15 onwards.

Does it even make sense to deal with the topic of private 5G networks?

Thomas Schildknecht: Absolutely! Because the possibility of setting up campus networks will be possible from September 2019 - upon application to the Federal Network Agency. Politicians have made a bold decision for industry here: 100 MHz bandwidth in the 3.7 to 3.8 GHz band is reserved for industrial applications and was not offered to mobile network providers in the license auction! This means that companies can set up and operate their own base station with 4G LTE or 5G technology on their company premises and use it for their industrial processes. For example, for a kind of super WLAN for IIoT networking of mobile and autonomous conveyor vehicles. Around 20 MHz bandwidth is possible here per company, which corresponds to one Wi-Fi channel. If, for example, 500 AGVs are to be connected via Profinet, it should be borne in mind that the resource is likely to be fully utilized and latency times of 2.5 seconds are to be expected. The advantage is certainly that this bandwidth does not have to be shared with other companies and a higher transmission power is probably possible within a factory hall.
transmission power is possible within a factory building. Profinet communication would not currently tolerate this latency. Here you have to work with tricks: A patent implementation in our radio devices, for example, still allows reliable operation.

One technological challenge will be whether safety-critical signals such as emergency stops can also be reliably transmitted via a campus network. Profisafe via Profinet or TSN will be suitable fieldbus technologies here. Initial tests are being carried out on a 4G campus network at RWTH Aachen University as part of the 5GANG research project. We - Schildknecht AG - have already successfully tunneled Profisafe via Profinet at an Ericsson base station, for example.