More Progress in Solving Switching Problems
(11) Comments[Report from LPP] We continue to make progress in solving the switching problems, but have not yet cured them. Eight switches have been fitted with the new copper plates and we have tested them with the new Argon-SF6 mixture. Now we can get all eight switches to fire during the pulse, which we were not able to do before.
The best rise time for the current is now 1.8 microseconds and the maximum current obtained is 0.85 MA at 30 kV. Based on this data, we have no doubt that, with all 12 switches firing, we can achieve our second technical goal of getting 1 MA at 25 kV.
However, we have not prevented a lot of pre-firing, generally just a kV or two before full charge. In addition, the best simultaneity we have achieved (time from first to last switch firing) is 150 ns, which is barely acceptable, and most of the time it is 300 ns. There is no difference in the simultaneity when the switches pre-fire and when they are triggered to fire. Most seriously, sometimes the trigger fires, but there is no breakdown within the chamber, and firing of the bank is delayed by microseconds. Finally, the spark plugs that fire the switches have been breaking.
The manufacturer, R.E. Beverley, believes he can solve the pre-firing problem by changing the profile of the electrodes within the switches, and we have sent the electrodes to him for this purpose. He will supply us with custom spark plugs as opposed to the current automotive spark plugs. The other problems seem to indicate that the trigger is not delivering as big or as sharp a pulse as it should, failing to trigger the switches rapidly. We have been unable to test this directly as the HV probe we purchased was non-functional and had to be returned. It is now fixed, according to Team Technology, the supplier, and we will soon be able to diagnosis, and hopefully fix the problem with the trigger. We will also be consulting the literature and our DPF colleagues for the best way to solve the switching problems.
Far Time-of-Flight calibrated, noise problem eliminated
Fusion Food Campaign
Comments
For a more in depth discussion, start a thread in the forums.Regarding rise times, this can be modeled as an RLC series circuit.
The capacitive load is the DPF device, in series with the capacitor bank. The resistive and inductive loads are due to the wires feeding it.
R = V/I = 30 kV / 106.25 kA = 0.2823 ohm; so
this is saying that R requires 1.8 µs to fall to 0.2823 ohm.
1.8 µs rise time corresponds to a 1/2 wave, (which makes f = 277.7 kHz.)
period = 3.6 µs = 2 pi sqrt( LC ); so
sqrt(LC) = 0.573e-6.
It’s doubtful that you could reduce C, but L maybe?
Could you reduce L by replacing the feed wires with a pair of pipes, arranged coaxially?
Page 2 of 2 pages of comments < 1 2
Post a Comment