July Switch Update

Posted by Rezwan on Aug 09, 2010 at 08:06 AM
Tools: Print | | (8) Comments

We completed testing the new spark plugs that we had designed.  The first good news is that, with the Lexan insulators and tungsten pins, the spark plugs have lasted through over 200 shots without breaking.  While there is some melting of the tungsten electrodes, it is small and manageable.  On Dr. Subramanian’s suggestion, we have modified the copper electrodes to eliminate a small sharp edge that could contribute to insulator break-down.

Second, the amount of current we are producing per firing capacitor has improved.  We have achieved 1 MA at 27 kV with only 8 capacitors firing, something that required all 12 capacitors with the old spark plugs.  We believe that the large size of the tungsten pins and the better distribution of current has reduced the inductance of the switches and led to the increase in current, which makes us more confident that we can reach the design current for FF-1 of 2.8 MA.

However, despite careful adjustment of the spark gaps, the simultaneity of firing with the new spark plugs is worse than with the old ones, and on average, only five switches are firing on the trigger. We believe we know the cause and cure of this problem.  Our HV probe tells us that the trigger voltage only rises to about 12 kV.  We think that is because the Lexan insulator gives off a small amount of vapor that shorts the trigger electrode too quickly to the adjacent electrode.  Amgad Mohammed pointed out that, since almost all capacitors that do not fire on the trigger then do fire at the time of the pinch, when the voltage spikes to about 50 kV, a sufficiently high trigger pulse will indeed trigger them.  He asked how the trigger voltage could be increased and the answer was by replacing the DC power supply that powers the trigger.

We have ordered this new power supply, which will increase the charging voltage on the trigger from 20 to 40 kV and will arrive near the end of August.  We think that by doubling the rate of rise of the trigger pulse, we will be able to get the trigger voltage up to at least 20 kV before it shorts.  That, together with the capacitor voltage of 30 kV, should get us to the 50 kV needed to fire all of the switches together.

While waiting for the new power supply, all the switches have been removed for maintenance.  We observed that the copper inserts that were used were not completely aligned with the aluminum plates underneath them.  This misalignment was at microscopic levels, but even this small misalignment seemed to generate coronal discharge in the sharp edges.  One such corona discharge caused a surface breakdown inside switch #2 and resulted in a short.  We have taken all the plates to the workshop and have polished them.  We are also inserting Mylar sheets at the interface to prevent any surface breakdowns from occurring.  This task will be completed by the end of this month.

Note that the ongoing trouble with switches is related to attempts at cost cutting.  The large DPF in Vegas is said to work at higher voltage with greater reliability.  Per Eric:

Vegas has rail gaps, which we rejected on grounds of cost. Also, because they are difficult to adjust—they must be highly parallel.

Sing Lee, who has the widest knowledge of various machines, basically thought there was a pretty clear correlation of quality, complexity and cost.

The best switches cost a lot and are complicated to maintain. On the other hand, spark gap switches (our kind) with much higher trigger voltages—120 kV vs our current 20 kV—have been made to function reliably recently.  So it is not unreasonable to hope that at 40 kV we can get ours to work as well.

It is admirable to see the crew working so hard within their budgetary constraints to make the less expensive components work.  However, the switch savings could be offset by all the time and effort that has been spent bringing these switches up to par (equalizing the complexity).  Although Eric is confident that the crew can get the switches consistently to par, I would hate to see the project come to a halt because of a continued switch accommodation issue and lack of adequate, reasonable and necessary financial support.  It would be a classic case of “for want of a nail, the battle was lost”.  The researchers, too, all want to move beyond the switch issues and get on with the advanced fusion research that relies on the switches. 

While the LPPX crew takes pride in being able to operate on a shoestring, and while they continue to make creative, resourceful improvements on the switch system and extract fascinating research results, I would be a whole lot happier if they had better (above and beyond) financial support.  An infusion of funds would give them the ability to spare the time and cost to upgrade to better components.  A committed funding source would ensure that they can take care of other contingencies as they emerge and that the project will be completed in style.  It would enable them to take the time they need to produce the highest quality experiment.