Re: Is it possible to delay one input to the WFD?
MICHAEL@CITHE1.CITHEP.CALTECH.EDU
Thu, 08 Jul 1993 16:23:32 -0800 (PST)
I know of only two approaches to producing on order of 100 ns delay for
these signals. The first way is simply to make 100 ns long cables. These
cables would best be made from a low attenuation, low dispersion cable.
If built from RG58c/u then the length of the cables would be 20 meters.
RG58 has an attentuation at 100 MHz of about 5 decibels per 100 ft. Hence,
we could expect that single photoelectron pulses would be attenuated by
this much cable by about 50%. I don't think that is acceptable and I don't
think that we want to run phototube gains differently in order to account
for this. At higher cost, one can use a lower attenuation cable. Say that
we set a critereon of no more than a 5% difference for the short vs long
cables. This would require use of RG218/U, a large diameter low attenuation
flexible cable (at least its not hard-line) which I think probably costs about
5 times more than RG58 (one would have to check prices). In addition to the
cost for the cable, this kind of cable is much trickier for building connectors
and I think those are probably more expensive. Some kind of an adapter panel
would very likely be required which adds an additional cost of 4 connectors
(I think) per channel. If I assume that the cable is $2.00 per foot and the
connectors are $8 apiece then this would work out to about $164 per channel
additional cost compared to using short cables.
A cheaper alternative might be to use bulk delay circuits (like the autocoils
in the ERP). Unfortunately though, I think these generally have attenuation
characteristics similar to RG58 at best. That much delay could probably be
purchased for only $10-$30 per channel I guess (it probably depends on the
attentuation properties). I am not optimistic that this could be a solution
though.
My conclusion is that you need a very strong reason to pursue this option.
My numbers here are or course very "back of the envelope". However, I think
they show that making this delay is at least not trivial so that it needs
some real research if it is to be pursued. I know that they are planning to
start building cables at Texas very soon. However, I think that since this
kind of cable probably needs adapter cables anyways, the folks at Texas can
just proceed making all of the regular cables and then this additional delay
could be added at a later time.
Doug