RE: Kicker grading measurement

["Mike Barnes" <Mike.Barnes@cern.ch>]

Dear Peter,

I suspect that there is a problem with the measurement. The scope set-up
looks different to me; for example the 2005-hv2-dn-01.png had a
sensitivity of 500mV/div, whereas up_01.png has a sensitivity of 1V/div.
However, although 1V/div will result in more noise than 500mV/div, I
don't think the scope setting is the source of the problem. 
I suspect that the ground on the scope probe is NOT adequate. The ground
wire should be as short as reasonable possible and it should be
positioned so as to minimize its inductance and also pickup from the HV
pasts of the circuit. 

What is the HV voltage and frequency for your measurement?

In addition, when I was at PSI during August 2005, we found that the HV
probe you are using at PSI, for the voltage grading measurements, is
very susceptible to capacitive pickup from the stack -- i.e. by changing
the angle of the measurement probe slightly the measured transient
voltage changed. We have found this same problem at CERN with the newer
type HV probes, but not the older style. While I was at PSI Francoise
and I discussed the possibility of making up a conducting piece, to
connect to the HV probe banana plug, so that the probe body could be
several more inches away from the stack: this would reduce the
capacitive pickup from the stack to the probe body. Was this piece used
during the latest measurements?.

Where did you ground the HV probe for your measurements?

Best regards,
Mike 

-----Original Message-----
From: Winter Peter [mailto:peter.winter@psi.ch] 
Sent: Monday, May 01, 2006 1:25 AM
To: Mike Barnes
Cc: Mulhauser Francoise; kiburg@npl.uiuc.edu; kammel@npl.uiuc.edu
Subject: Kicker grading measurement

Dear Mike,

following Francoise proposal and having a long weekend without beam, we
started to perform a kicker grading measurement. However, our findings
are surprising and yet puzzling to us, so we thought it might be a good
time to ask for your opinion and advice.
We setup the measurement as similar as possible like it has been
performed last year before te MuLan run. The oscilloscope settings were
restored, the HV probe is the same and grounded with the same cable and
at the same position inside the kicker. In the attachment you'll find an
example of one card from last year (2005-hv2-dn-01.png) and one from our
recent measurement (up_01.png). The subtracted signal between the drain
and the source is the orange M1.
In general, we observed that the spikes at the transitions are much
larger than last year. While then they had been around 7-15mV with
spikes up to 30-40mV for the worse cards, we now have spikes usually in
the order of 30-60mV with some cards even higher than that. The shown
example is one of the worst.
In addition we observe that the amplitude of the wiggles at the
beginning of each transition are much more pronounced than last year.
They have a frequency in the order of 50MHz. Furthermore, the baseline
of our measurement is in general noisier than last year.
We had a longer phone call with Francoise to ensure that we're in
principal doing the same like last year. There is one major difference,
namely that we ave one pump connected to the kicker tank which is
pumping. We want to repeat the measurements without the pump but the
strong correlation of the wiggles with the transition and the high
frequency might indicate that this is not the source of this.
The major question arising is, if what we see is related to the MOSFET
cards and if that's a bad sign? And why is the amplitude so much
stronger? In other words, are these big spikes really induced on the
card itself or is this something connected with our way of measuring.
Because in case of the latter these large spikes could be just an
artefact of that.
We already see the wiggles with the probe being inside a cabinet but not
touching any card. When we close the cabinet having the probe outside we
do not pickup this signal anymore.
I should also mention that we already reseated all cards and optic
fibres of the cabinet HV1 without any improvement. Francoise idea was to
exchange one optic fibre completely to see if tht makes a huge change.
We hope that you can give us some new ideas. You also might judge if
this current results are consistent with the cards or optic fibres
becoming worse (hopefully not) or if this rather points to some problems
with our measurement.
Looking forward to your answer
greetings from PSI
Peter