I suppose this topic could go under many of the categories already underway - Bridge reliability, Modeling, Prototype Failures, but it probably deserves its own topic.

It probably goes down as one of the most expensive failures ever, and you can bet the calulated MTBF was about one million hours. There were probably more years of modeling that can be imagined.

Does anyone know what happened? Probably a well-guarded secret, but maybe a bridge? Or radiation?

In all seriousness, you can complete as much WCCA as you like and lifetime calculations but unless you can model ALL of the expected conditions your models can never be accurate.

As in the other thread about what to do where is no model then possibly the same applies here. I am sure that there are effects on electronic systems that are not modelled as they have not been modelled/calculated/found by engineers down on planet earth.

(It may be the aliens though, I mean we have invaded Mars..... :) )

One challenge with designing for satellites is that they are going in an out of sun light, which usually creates significant temperature fluctuation that is repeated thousands of times. For most parts and systems this is very hard on them not to mention years of accumulated radiation damage to many materials. A design that is good for even the harsh environment of automobiles may not stand up to years in space.

I agree with the comment that modeling and Worst Case Analysis (WCA) can miss things. Often the concentration on the WCA is like very carefully studying the tree next to you and missing the forest. Many space designs do not allow plastic parts but in space where there is no moisture, I don't know why plastic is not allowed; maybe because they are used in commercial equipment so they assume plastic to be inferior. It is true some vendors do make low grade parts designed for lowest price but not all.

There are offgassing requirements for certain plastic/polymer materials to consider, which could interact with other materials and cause performance issues.

I would hazard a guess with radiation: even though the screening process for space-level components is stringent (and is another costly procedure in itself), the fact that the parts are certified to be "rad hard" does not rule out a failure due to transient radiation levels that are greater than the peak limits for which the components are rated.

Sorry to get the discussion side tracked about plastic. I do have to say I have successfully flown high quality plastic on space shuttles with out any problems. Admittedly space shuttles is unique in that it does not stay up for years and the accumulated years of radiation.

How does the current flowing in the MOSFET's diodes of the totem pole switches cause the MOSFETs to fail? If operating correctly in the phase shifted bridge the current in the diode has virtually stopped before the voltage reverses. This helps reduce the stored charge - Is the diode conduction really causing problems or something else that turns on both FETs in the totem pole structure. I believe in the non-phase-shifted bridge when the load is not being driven there are two series MOSFETs withstanding the total supply voltage and any unusual voltage spikes but the phase-shifted bridge has only has one FET having to hold off the supply voltage. --- Just an observation

The telescope was only expected to have a service life of 20-25 years; to achieve this level of continual service from a power supply without any maintenance is commendable. Parts do degrade over time (namely filter capacitors) which can cause erratic behavior over time. Even if it was a PSFB (phase shifted full bridge) topology, the output filter capacitors could degrade, causing disturbances in the feedback loop, corrupting the load line characteristics. When the power supply falls out of ZVS range and theoretically becomes discontinuous, the secondary and leakage current reflected to the primary for the series switches diminishes, which can cause momentary dissipation in the switches. Insufficient reverse recovery in the body diodes can cause shoot-through which can easily destroy them. I am assuming the capacitors are of the electrolytic type.

Electrolytics in space are completely forbidden, it would have been completely done with MLC and other non-degrading types.

Electrolytics in space are completely forbidden, it would have been completely done with MLC and other non-degrading types.

Ray if you mean Aluminum electrolytic you are correct but Tantalum wet and solid electrolytics are used in space. I believe the wet are always hermetically sealed. I prefer the ceramics but for dampening the large value in a small space makes the Tantalum very attractive. As you well know they design the box and everything else in it and then give us the space and poor heat sinking that is left. They are paranoid about surge current in the Tantalums though.

Ernest, I bet you have spent many days like me, performing analysis to prove to Lockheed or Boeing that the surge current levels through Tantalums are acceptable. It was easier after AVX provided more supporting test data for their TAZ and TAJ series.

Back to the Hubble failure though: Nice to read some intelligent speculation on radiation effects but it wasn't electrical. The cause of failure was the air pressure inside of a power brick which was not evacuated before sealing. The module stayed together for many years until finally it gave way. The internal pressure caused the case to bulge and, with potting compound attached, it pulled components right off of the PC board.