Re: Effect of the 1mF Electrolytic Cap on the Fanout
Doug Michael (michael)
Mon, 24 Mar 1997 10:39:48 -0800 (PST)
Ioannis wrote:
>
> Following Doug and Ed's e-mails, expressing concerns as to what the
> 1mF electrolytic cap (instead of the 2.2uF) on the fanout can do to our
> PMT signals, Erik, Chris and myself went in the tunnel again to perform
> a series of tests.
> First we obtained some more of these 1mF electrolytic capacitors from
> the electronics shop here at G.S. These were manufactured by Daewoo and
> were rated 85V. They were different from the very first one I tested and
> also from the 4 that Chris brought from the U.S. and are now installed
> on the detector on the WFD-outputs of the fanout for channels 3B01-0, 3B02-0,
> 3B03-0 and 3B04-0.
> Following a series of testing on the bench, we decided to take SM3&4 off
> in order to have a look on the ones that are on the detector right now.
> I believe that the following are the conclussions of all 3 of us (Erik, Chris
> and myself).
> The 1mF electrolytic capacitor affects negatively our signal in the
> following 2 aspects:
> (1) Amplifies the noise that comes out of the fanout. The noise I am referring
> to was observed when we plugged on the backplane of the fanout crate the
> BNC output of the portable HP pulser, still keeping the pulser OFF.
> So, you may call it "0-input noise". This noise, was approximately
> -0.5 to +0.5 mV for the unmodified channels while it is -1.0 to 1.0 mV
> for the modified. The figures I've just quoted are only indicative
> because the noise varies with the channel used and the capacitor used.
> But it is a clearly observable phenomenon. The most "annoying" effect
> of this noise is observed when you simulate 4mV, 13nsec, 20Khz spes,
> feed them to the fanout and look at them with the *analog* scope,
> triggering with the trigger output of the pulser. We looked at both the
> modified and the unmodified outputs of the same channel.
> The spes from the modified output look "fuzzy", i.e. the spe has a band
> of almost 1mV, and is no longer a thin line, as it is from the unmodified
> one.
> The frequency component of this noise could not be isolated, because
> it presented a different component each time we changed the time-scale.
> The bottom line is that it is there.
What happens if the usual capacitor is kept in parallel?
1 mV level of noise is acceptable. The question is whether it can
get much worse.
There must be *some* idea of the frequency of the noise! I don't
understand that it never changes on any scope setting.
Here you quote the noise with the HP pulsser plugged in... what
about with no pulser? That is what really matters of course. I don't
obviously care about the noise with the pulser plugged in.
> (2) It attenuates the input signal by an additional 2-4%, with respect to
> the unmodified output. To be more exact, we created a signal with amplitude
> roughly 70mV and measured it to be 70.4mV at the modified output, while
> it was 72.8 on the unmodified one (~3%).
>
What was the rise and fall times? This needs to be checked for signals
as fast as SPE's.
Without seeing it on a number of channels, I might consider 2-4% to be
negligible attentuation. Are you sure that isn't just due to slight
differences in gain in two different channels? Or put another way, are
you sure that the attentuaion might not really big bigger than quoted?
> Now, as to what effect these two phenomena may have to the digitized
> waveforms, Chris has been looking at the real data from the 4 modified
> channels and is going to mail you his findings.
>
Doug