[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Test pulser for FDC calibration and other FDC issues



Hi Simon,
	I agree with all the below... And also note that if the F1TDC is needed 
on the anodes, and so a discriminator is needed, it will either be a 
different preamp/discriminator board design or else there will be a 
discriminator board somewhere. In either case, this means there really 
is no motivation on that account to use the standard cables for the 
preamp board to ADC board connection.
	I'll send some more details on the pulser control scheme later, but 
basically I conceive of a single LVDS output from the ADC board to the 
front end (preamp) board, which sets the pulser amplitude _and_ times 
its firing. The actual pulser would consist of an LVDS receiver, some 
decoding logic probably, a DAC, and a switching transistor, and of 
course a capacitor for every channel. Some of this could possibly be 
incorporated in the ASIC, though I think I have identified off-the-shelf 
parts that can do the job.
	We need some specifications for the pulser. Previously I had thought of 
it just as a hook for testing electronics. That is certainly needed. 
However, if it is elevated to the status of a calibration device we need 
more specifications:
	- Do channels need to be pulsed individually? Or all channels on a 
board together?
	- What rate? Low rates (~100Hz max) certainly is easiest, but maybe not 
good enough for calibration uses?
	- Amplitude accuracy? Both absolute accuracy and balance from channel 
to channel and from board to board need to be specified. Capacitors are 
available in 1% tolerance, but only for >10pF which may be a little 
high. If the capacitors are in the ASIC, they will match very well 
within the ASIC but not so well between different ones, in general; the 
abosolute tolerance is poor, maybe 15 or 20%. Low value capacitors might 
be made with good tolerance (few %) on the printed circuit board itself.
	Well, we need to discuss this for sure...

	Gerard

Simon Taylor wrote:
> Dear all:
> 
> Here are some thoughts about the FDC electronics that I hope will generate
> some discussion.
> 
> My studies of the prototype indicate that a channel-to-channel gain
> calibration scheme is necessary to get the optimum position resolution.
> This ment injecting a known amount of charge into each strip channel.  In
> practice I used several different charge amounts (via attenuation) to
> check the linearity of the preamp response.  In order to implement such a
> scheme Gerard proposed using one pair of wires from the twisted pair cable
> connecting the front-end boards to the readout electronics to control a pulser
> on the preamp boards we would use for the final FDC design.
> This sounds like a good idea to me.  I am mentioning this because it has
> implications for the design of the drift-chamber preamp boards (and the
> FADCs for the FDC) and want to make sure the feature does not get left
> off.
> 
> To simplify the design of the front end (preamp boards) I think it makes
> sense to have the boards for the cathode and anode signals be the same
> with a minor modification to account for the opposite polarity of the
> signals.  Gerard proposed having 24 channels (3 ASICs with 8 channels
> each) on each board so that we would use 25-pair cables to connect between
> the preamp boards and the flash-ADCs we would use to read out (at least)
> the cathode channels.  I think this is also a good suggestion because it cuts
> down on the cable count but it has further ramifications for the design of the
> FDC readout system.  I have not been able to demonstrate that we can obtain
> the desired time resolution from the flash-ADCs and the old nominal design
> called for reading out the wires with F1 TDCs anyway.  If we use the F1 TDCs
> then there is potentially a cable mismatch (the current F1 design uses 34-pin
> connectors) and we would have to add discriminators somewhere in the
> chain.  We are working on shaping the pulses so that we can refine the
> algorithm for picking off the time from the FADC data, so the last point
> may be moot, but until we can demonstrate that this works, the F1 TDC
> option is still on the table for the anode wires.
> 
> Simon
> 
>