new SC mag power supply tests at Danfysik (fwd)

[Elton Smith <elton@jlab.org>]

---------- Forwarded message ----------
Date: Mon, 16 Dec 2002 09:51:10 -0500
From: Paul Brindza <brindza@jlab.org>
To: Rolf Ent <ent@jlab.org>, Elton Smith <elton@jlab.org>,
     Larry Cardman <cardman@jlab.org>, Alex Dzierba <dzierba@indiana.edu>,
     Barbara Swanick <swanick@jlab.org>,
     John Cameron <cameron@iucf.indiana.edu>
Subject: new SC mag power supply tests at Danfysik

Greetings

The two new power supplies that were ordered for the HMS D and the 
Solenoid are superb.

They regulate from 1.5 amps up to 3000 amps !

The 16 hour stability is at 6 ppm.and better for an hour or less.

The power supply has several nested control loops to provide this
excellent stability and regulation. There is a voltage loop that is
outermost then there are five nested current loops. The inside loop is
straight current regulation. Then there is a slew rate loop then a ppm
loop then the DCCT filter and finally a NMR regulation loop.

The DCCT filter is ingeneous. It replaces what would be a very large RC
final filter by using an extra winding in the DCCT of 400 turns to provide
a passive gain of 400 to 1 so that 400 times smaller Caps and resistors
can be used. This filter tricks the tarnsistor bank into regulating its
own noise out.

The new supply uses a new polarity switch and dump switch that are less
than half the size of the present ones. Also a new main control board and
a new interface panel. The Transistor bank is of an all new design as is
the driver module. The T bank has 1300 transistors.


The DCCT is a 5000 amp unit to prevent the stray field saturation that 
we experienced with the HMS D unit.

The four days was spent in testing and debugging and tuning up the
numerous new regulation circuits and filters and controls. The power
supply is mostly a new design. Despite the higher dump power to accomodate
the solenoid the components are actuially smaller except for the new dump
resistor. The supply is thus roomier inside and considerable easier to
work on.

The condition of the supply on Monday was complete but very green.
We spent the entire four days testing and fine tuning.

Monday we ran at reduced input voltage for initial tests. The Voltage
regulation worked great up to ~ the 70 % due to input limit. Then we
closed the I regulation loop. After installing the jumper that selects
Resistive mode regulation all was perfect.

We performed some tests and calibrations and tested the polarity switch.

Tuesday we ran to 3000 amps with a 440 3Phase input. We found some small
problems including an overall 13 % offset between the external DCCT and
the internal. This was traced to a resistor divider in the comparator
between the output and temp stabilized burden resistor. The 3 % was a
second order effect in the driver card caused by the 10 %. We also found a
wiring error in the DCCT (open) and got that fixed. After calibrations we
started a 16 hour stability run overnight.

Wed stability for 16 hours was ~ 6PPM!!!
60 minute stability was ~ 3 ppm!

Absolute accuracy was ~ 100 ppm. full range.

We tested all the safety circuits and verified correct wiring. I found a
logic error in one pair of interlocks and that was fixed overnight. We
also found that the DCCT filter was wired with opposite phase.

Thursday

We verified correct wiring of DCCT filter and the interlock fix. The
system had the water flow and Temp interlocks wired as an "and". I asked
for this to become an "or". checked OK.

We recalibrated the system after some further changes to the comparator
dividers. System is dead on at 0 % and 100 % (better than 10 ppm) at 50 %
we obserced a small descripancy at 50 ppm. Swapped the ADC chip but the
effect was the same. All other calibrations were exact.

We tested the crowbar circuit. The crowbar was turning on prior to the
Transistor bank starting the - 10 volt slow dump. We adjusted some zenor
diodes that sleect the turn on so that the T bank starts dumping when they
are back biased by -10 volts. The crowbar comes in at ~ 10.5 volts and
limits the T bank voltage to protect the T banks. With a large inductor
when the PS shuts down the magnet supplies a large negative voltage. if
this gets too high damage results.

We tested the NMR regulation circuit for correct operation.

We also discussed a analogue quench detector that Danfysik makes. I have
schematics and info. Its a neat package and costs ~ 400 $ for four dual
channel device with latching circuits and power loss protection.


I left out a lot of detail but this was a terrific experience as I
actually got to do some work on the supply and tune it up. Sure beats
sitting around for boring pre-rehearsed tests. A fantastic learning
experience.

The power supply beats all rated parameters by sometimes factors of 10! We
definitely got our moneys worth and then some.

Paul

PS My reputation for having unique travel experiences continues.

SAS/United lost my bags coming and going!
I flew both ways with President Jimmy Carter and got to meet him.
He was on his way to Oslo.