The readout cards described here were designed by Stan Sherman of the Rutgers University Department of Physics & Astronomy electronics shop. Cards were constructed with outside vendors, and tested by the electronics shop.
Readout cards are designed to detect when a charged particle has passed through the straw chamber. The cards are designed to output an ``ECL''-like logical pulse whenever an input signal is seen.
Input signals from the chamber at operating voltages will typically be ~10 mV negative pulses, with about 10 ns rise and 30 ns fall times. The 10 mV peak corresponds to a peak current of about 32 microamps. This signal is coupled to ground through a 1500 pF capacitor, then fed into the input of an NEC1663 amplifier. The output of the amplifier is a ~300 mV positive signal, leading to a gain for the analog electronics of about 10 mV per microamp.
The amplifier output is fed into a LeCroy MVL407 quad comparator. This is a leading edge discriminator that gives a logical true when the signal exceeds a supplied positive threshold. The threshold is set by a voltage divider, a 10 ohm plus a 1500 ohm resistor on the board. We also limit the current with a 1500 ohm resistor in the circuit before the board. (Front chambers are fused, rather than having an external resistor.) For a 10 V power supply output, the threshhold level is 10*10/3010 = 33 mV. One can see that this is an order of magnitude less than the expected signal size. A lower threshold should not improve efficiency much, but it will lead to better timing and more noise.
Checks of the lowest operating threshold show that oscillations occur once threshhold voltage is dropped below ~2 V (6 mV at the comparator, corresponding to ~ 6 mV/25 = 0.25 mV at the input to the readout board). The oscillations self-sustain, until the threshold is increased to typically 3 - 5 V (9 mV to 15 mV at the comparator, corrsponding to 0.4 mV to 0.6 mV at the input to the board). Thus, the readout cards should always be operated with a threshold of at least 5 V. The noise onsite at CEBAF may be greater than the noise in testing conditions at Rutgers, resulting in a modified minimal required threshold.
The output of the comparators is then fed to shaping circuitry. The readout board is divided into two halves, each of eight channels. The shaping circuitry for the eight channels gives a different width logic pulse for each of the channels. This allows all eight to be multiplexed together with OR chips into a single output channel, reducing the number of cables and channels of TDC needed. The output pulse widths, generally adjusted to +/- 1 ns, are:
| Channel | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|
| Width (ns) | 25 | 45 | 35 | 55 | 85-90 | 65 | 100-105 | 75 |
| Width (ns) | 25 | 35 | 45 | 55 | 65 | 75 | 85-90 | 100-105 |
|---|---|---|---|---|---|---|---|---|
| Channel | 1 | 3 | 2 | 4 | 6 | 8 | 5 | 7 |
The multiplexing does reduce the rate capability of the chamber somewhat, but we do not expect this to be a problem. At 1 MHz singles rate, the maximum expected, there is only a ~10 % chance of a second hit within a chamber plane. The number of multiplexed channels per plane ranges from 11, for the front chambers, to 18 for chamber 4. Thus, the chances of a random hit in the same chamber section are only 0.5 to 1.0%. If other detectors in the system work, it should be generally possible to extract the track using the VDCs to project to the front chambers, and requiring that the distance of closest approach be small.
To reduce the noise generated by the logical pulses, and possibly picked up by the input wires between the straws and the readout boards, the output signals are sent from the boards with an attenuated signal size. The size is reduced from the 1 V swing that is ECL standard to a 0.1 V swing. Level shifter boards sited near the TDCs are used to increase the signal size back to ECL levels, before input to the TDCs.
WARNING:
The readout cards have to be handled with care, since they are static sensitive. Ground yourself, and preferably wear wrist straps, when handling the cards. Wrap cards in conductive plastic before transporting. A typical problem is ``frying'' the ECL output chips. This results in loss of one or both polarities of an output differential signal.
Created: May 6, 1996 Norma Lucero Last revised: August 23, 1996 Ronald Gilman