Digital IF – Transforming the Satellite and Space Industry
The satellite industry is undergoing its most significant digital transformation yet, poised to reduce costs, enhance resilience, spur innovation, and expand its global market share. However, breaking through the bottleneck of the ground segment is crucial to achieving this goal, as current analog IF systems struggle to scale and meet future capacity demands. Transitioning to digital IF promises a solution.
I recently spoke to Simon Swift, Engineering Director – Digital Technologies at ETL Systems and Director on the Board of the DIFI (Digital Intermediate Frequency Interoperability) Consortium, to find out about the digital transformation of the industry.
Can you provide an overview of the DIFI Consortium?
The consortium was founded in 2021 and for the last 2 years, I have chaired the Specification Working Group. It has been set up as an industry wide consortium to ensure that when Digital IF makes it way into the market, that the products from all of the different manufacturers are interoperable. It is very easy for new products to not work together. If you take 2 pieces of equipment that conform to the VITA 49 specifications, you could almost guarantee that if they are from different manufacturers that they would not work together. The reason for that is that there are many vendor specific extensions, which create opportunities to deviate from a core specification, which means interoperability cannot be guaranteed.
With the DIFI consortium we are tying down all of these holes, specifically for the aim of RF distribution, primarily for SATCOM systems. The consortium contains a wide range of companies including Amazon, Microsoft, ETL, ST Engineering iDirect, Intelsat, Eutelsat and many others from all different segments of the industry. There are equipment providers, cloud providers, end users, who are all working together very magnanimously to provide an interoperable solution to the market place that is truly beneficial.
Why have you got competitors working together?
The SATCOM distribution industry is a large market segment and all involved believe that it will only get bigger if everything is interoperable. A parallel to this would be the 3GPP working groups in the telecommunications sector.
ETL ran a demonstration at Satellite in Washington earlier this year, can you recap this?
ETL demonstrated an end to end system. We had 2 end to end boxes, both which were linked by 100Gbps ethernet ports and over that connection we put a DVBS2X 32ASPK Rate 5/6 signal with 260MHz occupied bandwidth.
With this setup, we ran a live video demonstration through a couple of off the shelf modems, kindly loaned by iDirect, using the mentioned industry representative bandwidth and bitrate. The thing we like about live video as a demo, is that there is no opportunity for smoke and mirrors. Also, when you see zero errors – signal quality, error statistics etc. using another manufacturers modems, it means you are not marking your own homework.
We had a phenomenal set of demonstrations, to over 100 different people from a large range of customer. Everyone was hugely receptive to the demo and were very excited to see the product that ETL had to offer.
ETL differentiates itself in the marketplace from its competitors by providing a very modular approach that allows a wide variety of different modules to be integrated into the solution. It is inline with our Genus standard, which all of our equipment conforms to in terms of form factor and electrical interfaces. This allows different frequency converters to be put onto the front end of it. It can be closely integrated with matrixes and brings a wide range of options.
From the module point of view we currently go into the box at L-Band (850-2450MHz). Because of the architecture we have and the range of frequency converters we have, it would be very easy for us to go in at X, Ku and Ka Band.
One of the challenges facing the end users as digital IF comes in, is that they have already invested a lot in the legacy analog equipment. By allowing the modular integrated and supporting standard interfaces, such as SNMP v3, we are able to ease their transition from the analog to digital world.
We are removing the analog signal over a analog cable from the modulator into the frequency converter and replacing it with digitized signal and putting it out over a 100Gbps optical ethernet.
What is the benefit of moving from analog to digital?
This all sounds quite expensive. One of the things that you need to remember with a RF signal is that whatever you do to it, you cannot make it any better. You can amplify it, but when you amplify the signal you amplify the noise with it. If you turn it into RF over fiber directly, all of the conversion equipment and all of the cables have a loss and a noise figure associated with them. So the further you send it, the more it degrades.
Now with digitalization, what you do is manage all of your signal degradation at the front end. So the DACs / ADCs are deployed there and there are specifications on bit depth and sample rate that are key to the performance of the digital IF system. However once you are in the digital domain, no matter how far you send it, how many times you split it and send it to different areas, every copy of that signal will be identical. So if you are looking to produce a diversity solution, with spatially diverse antennas, maybe 100km apart, you get no further degradation no matter how far apart you move those antennas. We use this for mitigation of rain fade. For example in Indonesia, you can get a tropical storm coming through and you will easily see 35-36dB of rain fade and that is your link margin completely gone. With a diversity site, it is unlikely that you will have that storm over both antenna sites at the same time and that helps providers meet their availability specifications .
One of the other advantages of Digital IF is the opportunity for virtualization. Modems suddenly move from being a hardware, capital expenditure to being a piece of software, potentially running on a x86 compute platform in a data center, and thus becomes OPEX. That is a great advantage, as it suddenly it allows for a whole range of pricing models. You can pay per bit, use, day. You no longer have capital equipment sitting there.
In the scenario for disaster recovery, you can stand up a variety of different modems to be similar to the terminals on the ground to enable communication with them.
If you have a sporting event / large festival, you can stand up more modems on that day without having to invest in a large amount of capital hardware.
Once the signal is in the digital domain, there are all sort of additional processing that can be carried out it.
What are the limiting factors for the Digital IF chain?
The key enabler for digital IF is faster ethernet and faster analog to digital converters and we are seeing both of those only moving up in speed.
For example on the ethernet side, when we first started looking at the digital IF product line and saw 100Gbps ethernet and that could be a challenge for customers. Now customers are talking about redundant 100Gbps down each side of the rack along with 400GHz data concentrators. Looking ahead, the 800Gbps ethernet spec got released earlier this year and the 1.6Tbps specifications is due for publication in 2026. All of this is driven by our need for more bandwidth.
Do you see the digital interface moving further up the RF chain?
At the moment, the digital signal stops at the frequency converters. We can see the potential for a digital interface on BUCs and BDCs as A/D convertors go up in frequency. we can also see the potential for the direct conversion to Ku / Ka band going forwards.