THE ECAL MGPA

Overview
The pre-amplifier chip is the first stage in the ECAL read out electronics after the photodetectors. Its purpose is to boost the weak signals from the photodetectors before they are passed to the ADC for digitization. Imperial played a major role in the revision of the ECAL electronics system in 2002. A modified architecture reduced substantially the number of costly optical links, and made use of optical and electronic developments from the Tracker. Imperial developed the new pre-amplifier chip, the multiple gain pre-amplifier (MGPA). The MGPA has three gain channels and the channel which is in range is decided by logic in the ADC chip. The MGPA can be used by both the APDs in the barrel and the VPTs in the endcap thanks to a choice of external feedback components to the first gain channel.
	performance
	The chip was designed within a very short timescale; the programme was started in mid 2002 and the first design was submitted in early 2003. So successful was this first version in meeting the requirements of noise, linearity and pulse shape matching that very little was changed for the final version in late 2003. A pre-amplifier must boost signals whilst contributing as little noise as possible to the read out system. Simulations of the expected noise contribution from the MGPA suggested that it would be within specifications and actual measurements (shown in the plot on the left) have confirmed this. The chip non-linearity was specified to be within ±0.1% of the full scale signal for each gain channel. A sample of ten chips from both iterations were tested and it was demonstrated that the non-linearity can be achieved.
radiation Hardness
The MGPA is made using 0.25μm technology which is known to be a radiation hard process. Nevertheless, the chip's performance after irradiation was checked to ensure that it would still function after receiving a large integrated dose. A chip was irradiated at Imperial College using 10 keV X-Rays to a total dose of 50 kGy, twice the expected total dose at CMS. The only difference noticed was a 3% reduction in gain.
	Links
	Mark Raymond's Documents