Just increase reliability a little!
This month’s topic is “over engineering: the ugly child of fear”. In this bonus article we will look how small changes to a mission can have drastic effects on cost and schedule.
The DLR TET satellite(s) of the German Aerospace Centre OOV office started out as the quest for affordable IOD/IOV capabilities with a fast cadence to enable and boost new technologies. For this a copy of a flight proven satellite developed by DLR was to be built by industry. Unfortunately due to a series of decisions driven by fear, the planned “carbon copy” was ultimately more twice as expensive and took 6 years to build. As a result the mission was neither affordable, nor did it achieved the desired IOD cadence, nor did it create a German answer to SSTL as the decision makers in had hoped. Let’s have a look!
Introduction
Before we dive into the depth of German space industry from the era 2000-2010 it is important to define a few things for the international audience. For example the fact that there are two DLR’s or the terms OOV and TET which might otherwise be confusing.
What is DLR?
Yes, you have heard it right. There are actually two DLR: one is the traditional space agency (lets call it DLR-Agency) and the other is a group of 27 research institutes (lets call this DLR-Institutes). Until 1997 these two were run separately. The agency part was called Deutsche Agentur für Raumflugangelegenheiten “German Agency for Space Flight Affairs” or short DARA. The institute part Deutsche Forschungsanstalt für Luft- und Raumfahrt “German Research Institute for Aviation and Space Flight” or short DLR (lets call it DFLR for matters of clarity). This marriage has never been free of problems as the DLR-Agency, as part of the new formed DLR, is not allowed to fund projects under their own roof – consequently is not allowed to give project money to the DLR-Institutes. What started out as a political move by DFRL leadership to get control over the pork barrels of DARA backfired. This creates much chargrin for the heads of DLR-Institutes to this day who probably think – so close and yet so far.
What is OOV?
On Orbit Verification (DLR OOV) is the term coined by German Aerospace Centre (DLR-Agency) for in-orbit demonstration (IOD/IOV) activities. Like similar activities by other agencies such as ESA it centers around the idea that due to the fear of mission loss many new technologies are stuck on the ground until proven in space.
What is TET?
The idea for dedicated carrier satellites for IOD was originally envisioned in the early 2000s at DLR-OOV office and since those are supposed to test technologies in Germany they are called TET short for “Technologie Erprobungs Träger” (Technology Test Carrier).
Among the first public notices TET can be found in the DLR annual report 2004/05. It also contains what could be described as programmatic or top level requirements. Lets keep these requirements in mind for later. There it says* (*translation by the author):
“It is planned to use for TET an affordable and already flight proven satellite carrier of the micro / mini satellite class
Similarly like the EC/ESA IOD the focus is on affordability and speed by using a flight proven platform.
“due to high demand it is planned to launch the first TET in 2008 (TET 108) and the second TET in 2010 (TET 210)”
It is planned to conduct phase A in the second half of 2005
This shows that from phase A kick-off to launch there is a 2-3 year period and acknowledging the repetitive nature of the OOV business it is planned to have these satellites with a 2 year cadence with at least one second satellite planned already.
“additional hosted payload opportunities on other satellites shall be used for payloads that cannot be accommodated on TET.
It also acknowledges the fact that additional IOD/IOV opportunities are required and proposes to use hosted payloads for this purpose. These additional opportunities (for example on Russian satellites) are also to be managed by DLR’s OOV department.
There is a choice to be made!
Objectively, considering the requirement for a flight proven carrier in the micro/mini satellite class, there were two potential contenders to be the starting point for TET in 2005. The DLR BIRD satellite built at DLR centre for optical sensor systems (DLR OS) in Berlin and the TUBSAT satellites built at the Chair of Space Technologies from TU Berlin.
DLR BIRD
The DLR BIRD was conceived and built at the DLR institute for optical sensor systems (DLR OS) in Berlin. The BIRD program was lead by a young project manager named Klaus Briess, who later (in 2003) was appointed Professor and became successor of Prof. Renner at TU-Berlin. BIRD was unlike any other: the satellite which was launched in 2001 was the worlds first small satellite with a mission to observe forest fires and thermal hot spot.
The – 94kg, roughly washing machine sized – satellite was undoubtedly a technological marvel and far ahead the satellites that SSTL was building at the same time. The bus included redundant and radiation hard on board computers, deployable solar panels for high power and a three axis attitude control system based on reaction wheels and star trackers and the payload consisted on a two channel actively cooled infrared payload. The BIRD satellite had all the bells and whistles – a truly worthy contender for the future TET satellites.
TUBSAT
TUBSAT in short for TU Berlin satellites is a series of small satellites built at the chair of space technologies of TU Berlin. In 1986 the idea was brought to life by Prof. Renner, also known as the father of the TUBSATs.
TUBSAT pioneered, star trackers, reaction wheels, sub 10m resolution and real time video on 50kg satellites before anybody else.
Under the leadership of Prof. Renner the TUBSAT’s became technology pioneers. The worlds first star tracker (TUBSAT A launched in 1991), the worlds first small satellite reaction wheel, the first use of a fibre optical gyro, the first sub 100kg satellite with 10m GSD (TUBSAT B in 1994), the first German nano satellite with 3 axis stabilisation (TUBSAT N/N1 in 1997) and the worlds first real time video satellite (DLR TUBSAT in 1999) were all made in Berlin! The TUBSAT satellites were so successful that during the 1990s nobody in Germany launched more satellites than TU Berlin.
Interestingly after the initial TUBSATs where each satellite was used only for one mission, from 1999 a new modular bus was build. TUBSAT-C (and evolved TUBSAT-C) was the bus that all TUBSAT satellites were based on (until Prof. Renners retirement). The TUBSAT-C sported a modern separation of Bus and payload section which made it easy to adapt it to new missions. This design approach is similar to that of BIRD which also has a section and a dedicated payload section.
Regarding the cost the TUBSAT team was downright frugal. Since there were never more than 4 people working in the TUBSAT team (Prof. Renner, 1-2 research assistants and 1 student assistant) labour cost which is the main driver for any space mission was very low. In fact the 3 million Marks (about 1.5MEUR) that Prof. Renner received from the German government as a grant for his first satellite were stretched by him over 4 satellites (TUBSAT A, TUBSAT B, TUBSAT N/N1) and the follow on satellites were paid in part by third party.
TUBSAT satellites are always known for their frugality. Even accounting for the free university resources the real cost per mission were lower than 30% of DLR BIRD
Prices for DLR TUBSAT and Maroc TUBSAT are unknown to the author but LAPAN TUBSAT was sold to LAPAN for about 1MEUR (not including launch). This was obviously only possible because the TUBSAT team used the resources already available at the university (such as Prof. Renner’s salary, the university’s mechanical lab and CNC’s, as well as a host of other university facilities). That said the true end to end cost of LAPAN TUBSAT was probably still a factor 3-5 lower than that of BIRD.
And the winner is: DLR BIRD
TUBSAT may have presented a lower cost alternative but in addition to the fear that it might not work (because it had been designed outside the proper environment) it was really dependent on just one man: Prof. Renner. Which in 2005 was in his late 60s. Unfortunately, for all his genius the father of the TUBSAT’s just simply never build a business empire like Sir Martin Sweeting did with SSTL. In the end its a matter of preference as Prof. Renner always liked to work with very small teams of 2-4 people max. Consequently, and maybe not too surprisingly, the DLR OOV program made DLR BIRD the blueprint for TET.
Meet the DLR TET-1 (based on DLR BIRD)
To the defense of DLR, to use the BIRD satellite as blueprint for TET 100% fulfilled the requirements of being flight proven and in addition there was another requirement of the TET program that we need to shed light on.
Preconditions driven by (political) necessity
Remember that initial statement of the two DLR? This will be important here: DLR BIRD was built at an DLR-Institute (DLR OS), therefore the DLR-Agency was unable to fund via OOV building the TET satellite there. As a workaround the plan was hatched to do a technology transfer from DLR-OS to industry. This would mean that the TET project money could go from DLR-Agency to industry and they in turn could give subcontracts to DLR-OS. This was thought to to empower the German space industry in Berlin, namely Astro- und Feinwerktechnik Adlershof (Astrofein) to build BIRD class satellites on their own.
Making a technology transfer from DLR-OS to Astrofein was a logical choice driven by political necessity (originated in the fusion of DARA and DFLR in 1997).
The hope was that similar to the success of SSTL a German contender to the up and coming small satellite market could be created. Being a spin-off of the DLR-OS itself (founded in 1993) Astrofein is very close (literally only a few meters away) to the team that built the DLR BIRD satellite. In addition Astrofein had produced subsystems of BIRD (e.g. RW90 reaction wheels, deployable solar panels) and manufactured the satellite structures.
Choices driven by fear
The following choices however are far less logical and can best be explained by fear of the decision makers…
We need an experienced mission prime!
Astrofein today is a medium sized company with international recognition and close to 100 people. Back then however there were voices that put in question the ability of this much younger and smaller version of Astrofein to handle the mission.
DLR-OS had been the prime of BIRD in all of but the name. To distrust them to be able to guide Astrofein was driven by fear.
Further pressure was levied from the existing industry which wanted a slice of the cake. In the end a half baked compromise was found; Astrofein would only be responsible for the bus and an experienced prime would have mission responsibility.
To make things more acceptable for Astrofein they were given the choice who to take as their overlord. The teams at DLR-OS and Astrofein settled on Kayser Threde (also known as KT). This was the best choice in a bad situation. KT had shown interest to expand to prime role from originally only having developed payloads and so the work split seemed natural. Astrofein would build the bus and KT would act as prime build the payload interface.
The team of DLR-OS and Astrofein was ideal
Adding a extra mission prime not so much.
Unfortunately, not everything went well in the project and so shortly after the honeymoon, with every new setback and delay reservations formed, both in Munich and Berlin, to an extend that was very probable a liability for collaboration in future TET projects.
We need (full) ECSS!
I cannot pinpoint this to one specific decision maker but somebody decided that the TET mission would best be suited by use of (full) ECSS. This was a stark deviation from the more measured approach in the BIRD project. It seems that the decision makers out of fear to be blamed for a mission loss did not tailor the ECSS and instead tried to used it in full.
With that the QA switched to ludicrous speed. The results are as expected and known from similar run projects at ESA: Cost cost explosion and schedules overruns.
“In the TET project the tiniest parts, down the bolts, had two Leitz folders each.”
Source: person involved in TET
In the end more than 100,000 pages were written for the TET project. The folder cabinet filled an entire room at Astrofein. Just imagine that the mass of the paper alone (ignoring the folders) was 5x the mass of the satellite!
The weight of your documentation should not outweigh the weight of your satellite!
We need to increase reliability!
Even though the former DLR BIRD project manager (Prof. Klaus Briess) had famously ignored all requests to calculate the reliability of the BIRD satellite, it is an open secret that the number was around 89% after 12 months. To give you a perspective this is up with the best in industry like the SSTL X-50 or Airbus Astrobus 100 (CNES Myriade).
Unfortunately, someone along the line thought it would be a good idea to increase the reliability to >95% for 14 months. You know, 5% risk of loss is better than 11%…
Increase of reliability does not make mission success a guarantee. What is a given however is the “success” in increasing cost
This person probably did not fully anticipate what will be the effect to move outside optimum of the performance/cost S-Curve. As described in “the ugly child of fear” this seemingly small change (and its consequences) made cost literally double.
Consequences
Each of the decisions had consequences. The political nit-picking (who is prime) already significantly delayed the start of Phase B.
The TET program started with Phase A in January 2005, which ended October 2006. The TET-1 Phase B started in June 2007, […] and was completed […] beginning 2008. Source: Föckersperger et al.
Over-boarding QA/PA (full ECSS) combined with unnecessarily high reliability (for an IOD mission) made the project progress at much slower speed than anticipated. It did not help that there was another gap Phase B and C/D which put financial stress on the companies.
“In July 2008, DLR awarded the prime contract to Kayser-Threde GmbH of Munich (a company of OHB-System, Bremen) for […] TET-1. The microsatellite bus was designed and developed at AstroFein (Astro- und Feinwerktechnik Adlershof GmbH), Berlin in a subcontract to Kayser-Threde.” Source: EOportal
That means instead of having the first satellite launched 2008 was the year that the contract (phase C/D) was given to the prime. It then took another 2.5 years to get to launch readiness.
It was originally planned to launch the TET in March 2011 unfortunately problems with the rocket delayed the launch to July 2012 Source: Wikipedia
End of 2010, when the satellite was finally ready for launch a problem with the launch service provider meant it was stuck on ground for additional 18 month. During this time until the launch in July 2012 the engineering teams continued to do development work and (software) improvements on the satellite. Working teams mean further increase in cost.
TET-1 took 6 years from start of Phase A until launch readiness in 2010
At the end of the day the satellite had consumed more than twice the budget of the BIRD satellite of which it was supposed to be a carbon copy.
The final mission cost of TET until 2012 is 34-37.5MEUR depending how you count.
This drastic cost overrun meant that there was no chance of a TET-2 (TET-1 had basically eaten all budget of both satellites) and so what remains is the realization that while IOD is important the original concept of TET did not have the right formula. Neither was TET able to generate a sustainable IOD service in a regular fashion as per the agencies top level requirements nor did it create a German SSTL.
But life carries on…
Regardless the setbacks of the TET-1 mission the teams in Berlin were able to show that they could build small satellites on a more reasonable budget.
The DLR team
DLR BIROS is a spiritual successor of TET-1 with more reasonable QA/PA levels.
DLR BIROS, spiritual successor of TET was built for 10MEUR at DLR OS until 2016
BIROS was built the way that TET was originally intended to: at DLR-OS with help from Astrofein for around 10MEUR* (* a grant from the German Federal Ministry of Science). BIROS was launch in mid 2016 and together with TET forms the Firebird constellation.
This is a great success even though it personally find it sad, that the industrial team of OHB(KT)/Astrofein was unable to sell the TET satellite or the advanced TET-XL commercially or use it for other German government missions.
The TUBSAT team
The legacy of the TUBSAT team also carries on even after the retirement of Prof. Renner.
Prof. Renner himself continued to act as a consultant to LAPAN (Indonesian Space Agency) which went on to build a successful series of small satellites based on the TUBSAT ideas.
Prof. Renner continued after retirement to consult the Indonesian Space Agency (LAPAN) how to build small satellites.
His successor Prof. Briess built the chair of space technologies into a bustling power house with more than 50 scientific researchers. The fruit of this labour was that until Prof. Briess’ retirement in September 2020 no less than 15 (!) nano satellites (1kg – 20kg) were launched. And this legacy lives on as there are 4 more pico and one nano satellites started under is reign that are to be launched in 2021.
Prof. Brieß managed to grow the chair of space technology from 3 researchers to 50 and to launch 15 (!) nano satellites until his retirement in late 2020.
That means that the chair of space technology of TU Berlin, now under the leadership of Prof. Stoll remains the most active player in Germany when it comes to small satellite research. He also expressed that for the future TUBSAT satellites should move beyond pure engineering missions and collaborate with other Research institutes on scientific missions.
Prof. Stoll is determined to grow the TUBSAT legacy and develop the satellites beyond engineering and towards real science missions.
And while there is no German SSTL – yet, there have been very successful satellite companies grown out of TU Berlin TUBSAT group which are shaping the future of the German industry.
Berlin Space Technologies (BST) was founded in 2010 by former assistants of Prof. Renner and Prof. Briess. BST has contributed to more than 75 missions and built entire small satellite systems for international customers. As a joint venture with its Partner Azista Industries BST has built a factory to mass manufacture satellites in the 50-150kg range.
German Orbital Systems is a spin-off from former scientific assistants of Prof. Briess. They are a global leader in nano satellites and have successfully produced satellites for international customers.
Vectronic Aerospace (Vectronic) was founded in 2003 by former assistants of Prof. Renner and is a leader in small satellite subsystems. Vectronic supplies many international missions including the satellites of the Indonesian Space Agency.
Quō vādis – or what will the future be for IOD made in Germany?
First of all: lets state a regular IOD service is a good idea. and the requirements that DLR had identified were correct and straight forward:
- Regular opportunities (minimum once per 2 years better every year)
- Supported maximum payload size: 50kg, 0.2m³, 200W peak
- Fast turn around (<2 year project cycle, <1 year PL delivery to launch)
- Low cost to allow repetitive missions instead of one every 10 years.
Unfortunately the first try to implement did not work. I think it is time to think about how the original goals can be achieved better. Here are some recommendations:
- Use a simpler platform
Some satellites such as the original TET-1 are like beautiful sports cars. They have its uses on the race track and to impress on the Leopoldstrasse in Munich or Ku’dam in Berlin but probably not the right choice when you want to do the groceries. IOD/IOV needs a platform that is affordable. In the spectrum between TUBSAT and TET-1 future IOD/IOV satellites should be significantly closer to the simple yet effective side.
2. Less politics and less overheads
The example of TET-1 shows how important it is to have a lean management of a project with a low amount of overheads. I think the industry is ready to come up with own concepts and to deliver towards the OOV goal. Why not contracting a service to bring a certain number of IOD payloads into space rather than procuring a satellite?
3. More Reasonable Quality Assurance
Lets spell out bold: BIRD had a reasonable quality assurance. TET did not. For a future IOD program it would be good to keep that in mind. To accept and manage risk rather than trying to completely avoid them will very likely create an affordable and reliable solution. Below shows a concept in which low cost platform can be used reliably. This is based on the fact satellites built in a series are much lower cost that individual pieces of art.
It has been 10 years…
since TET-1 was readied for launch. It is time to work out a smarter and sustainable plan to implement IOD/IOV. Sure, we can “stick with what we know” and continue to waste time and money building super high end satellites once every blue moon but wouldn’t it be better if we use the capabilities and ideas of #newspace instead? Lets use the ability to build affordable satellites at high cadence and and combine that with the new low launch launch opportunities. That will allow to achieve an IOD service with yearly launches that when summed up over a decade will in total cost less than initial TET-1.
Rather than to redo the errors of the past we should strife to achieve do more with less.
Why not set the target to have one IOD mission every year, for 10 years for less than the price of one TET-1?
How can you help:
This text is part of a series of articles in which the author sets the framework to start a discussion about the wrongs of the space industry. If you have experienced similar things, leave a comment. Other views and opinions are very welcome, too, as they may present a way forward. Please be kind to each other.
Disclaimer
The author’s views are his own do not represent the views of Berlin Space Technologies.
Leave a Reply