May 2018


not literally but our balancing services at GomSpace, is in transferred meaning taking us into orbit. We did more than 2 years ago, have the first meeting with founder and partner at GomSpace, Morten Bisgaard, developer Jonas Morsbøl and had initial discussions about balancing of gyroscope reaction wheels.

GomSpace is among other things, developing and manufacturing Nano satellites and inside each Nano satellite, gyroscopes controlling the satellites position in space (Gyroscope is a device used for measuring – maintaining the satellites rotational degrees of freedom, orientation and angular velocity. In this case, it is a spinning wheel in which the axis of rotation is free to assume any orientation by itself. When rotating, the orientation of this axis is unaffected by tilting or rotation of the mounting, according to the conservation of angular momentum. The numbers of gyroscope reactions wheels varies depending of satellite type.

Over the first couple of meetings, technical aspects were discussed as well the design of the reaction wheel, in order to perform the balancing. Søren Uhre, TCO at Uhre & Nybæk, had not doubts about whether was possible to perform the balancing with VibXpert II. The reaction wheels just had to have, the physicals possibilities to make the adjustments and the balancing grade, had to be determined.

Determining the permissible residual balancing was given by ISO standards Uper = (Eper x m) / r where

Uper is unbalance permissible

Eper Rotor permissible specific unbalance

m is the weight of the rotor and

r the balancing radius.

With a rotor weight of approximately 105 grams and a balancing radius of 15 mm, the balancing grade was calculated to 0,8 g/mm in total. Hence the balancing should be in 2 plans, the exact figure per plane is 0,4 g/mm.

Talks about design ended with 2 rows of 2 mm threaded screw holes, 12 in each end of the 30 mm long reaction wheel.

Using the VibXpert II instrument from Prüftechnik for these balancing jobs, was of course an easy choice hence we use this instrument every day and for all our balancing jobs. The accelerometers normally used for the balancing, would though be an issue due to the weight.

Our standard accelerometer VIB 6.142 R weight of 36 grams and the weight would have to high an influence overseeing the weight of the reaction wheel. Even the reflective tape, normally used for the laser trigger was considered to have an influence. Søren Uhre found alternative solutions and with a new accelerometers, with a weight of 4 grams and a silver marker pen for the trigger mark, we were ready to our first trial balancing.

The first test was performed in September 2016. Morten Bisgård and Jonas Morsbøl had prepared two reaction wheels, number 10 and 11. The initial test on wheel 11 included a run-up test to determine if there were any resonant areas. The test showed a good stabile run-up, with a minor resonant point at 1500 rpm.


The first balancing run, showed an unbalance of 1,4mm/s RMS at the running speed. The unbalance was balanced according to below polar plot. The unbalance was reduced to < 0,1mm/s


After the balancing runs, there were still some vibrations in the unit.

The  spectrum shows how the vibration level at the running speed (110hz) is low, but instead we see “hay stacks” in the spectrum, with harmonics, that indicates looseness of a bearing fit. The spectrum RMS is now 0,76mm/s RMS. The loose bearing fit was a known fact, and would of course be improved.




Below is a final run-up test to 6000 rpm.


Since wheel 11 had the problem with the bearing fit, we changed to wheel 10.





Wheel 10 was balance from 0,55mm/s to 0,1mm/s RMS. Compared to wheel 11, the spectrum now shows a dominant peak at 10x ( 1.100 Hz). This peak has been discussed and is related to electrical interference.


To prove that the 10X peak is electrical related, a coast down test was made, where the power was cut. This means that no external power disturbs the vibration picture

Below waterfall plot shows how an 8X peak now instead is visible in the spectrum. This peak was discussed and found to be related to the magnets in the motor.





Besides above vibration readings, a resonant area was detected, when the unit was put in upright position. Another run-up test was performed to determine the frequency area and level. The waterfall plot now clearly shows a resonant area between 2700 – 3050 rpm. The resonance was clearly induced by the motor end covers.







The conclusion from the initial test was, that balancing could be performed without any issues, despite the ‘low tech test setup’ as seen in picture.

First demo balancing. To the left, laser trigger reading the speed signal from the mark made with pen, upon the reaction wheel inside housing. The tweezers is for operating the screws used to correction.

Conclusion was also changing the setup for adjustments. The 2 mm threaded screw holes were to ‘coarse’ and to adjust any wobbling phenomenon, another row of screw holes around the centre of the reaction wheel, would be helpful. The screw holes is now changed to 1,6 mm and tiny unbrako screws in different length and therefor weight, is used for adjustment.

GomSpace has worked with the design of housing and ongoing, tested different bearings for the gyroscope reaction wheels.

Anders B Nielsen and Mathias M Nielsen at balancing setup


Anders Bo Nielsen to the left, is electrical engineer at Gomspace and responsible for test and balancing of the reaction wheels.







Mathias Møller Nielsen, is one of Uhre & Nybæk technicians performing the balancing at GomSpace and is here, marking the flywheel with a pen, for the trigger reading.

Deeply concentrated with focus on accuracy and cleanliness


Uhre & Nybæk invested in another 4 grams accelerometer and when the balancing is performed today, both axes are measured simultaneously, which improves the time used for the balancing job.

Result? Not to be seen from surface of our Earth but somewhere up there, the Nano satellites have been in orbit for more than one year.


Want to know more? Contact us or visit Gomspace