So far, for all sensor studies, I have been using either a single channel active board (Santa Cruz or Chubut) or a multi channel passive readout board. Though this works, it is not ideal. Firstly, the passive readout board is not the best because usually the amplifier is installed as close as possible to the readout board. Moreover, if the board is changed for different studies then it is an additional variable to take into account when looking at the results. The ideal case would be a multi channel readout board with amplifiers integrated on the board itself. So I decided to solve this and implemented the “Chubut 2 prototype”.
Below there are some pictures of this prototype:
This prototype has two active channels with integrated amplifiers close to the DUT. It also features a Sensirion humidity and temperature sensor close to the sample to precisely monitor the conditions and, if needed, implement a temperature control (for example in the TCT setup). Additionally, I went for a design that combines a “main board” with a “carrier board”, which can be seen in the pictures. This I made because to ease the life of the guy that is in the lab testing the devices (i.e. me). The size of this prototype is almost identical to the size of the current Chubut board and I switched to using MMCX connectors because SMA just don’t fit.
After assembling everything together I verified that not only the board works but it seems to be a bit better than our current boards. Since the design has a carrier board, I mounted a few detectors and tested that it is easy to swap between one another. Then I went for a “full test” in the beta setup, for which I installed a TI-LGAD with two out of the four pads connected to the outputs and installed it in the Robocold setup.
Below there is an example of a signal from a beta particle:
The only detail I don’t like is the long tail of the signal, but anyway it does not affect our tests because we only look at the rising edge, and for the charge it only changes the proportionally factor to convert to Coulomb. I still don’t know where is it coming from.
So far I only measured rigorously one of the two channels, but I verified that both of them work similarly.
If we look at the time resolution measured with our Robocold setup, we get this:
which is pretty good.
The collected charge (in arbitrary units) also looks quite good, despite the long tails in the signals:
The Signal to Noise Ratio (SNR) is also amazingly good. For example, at the highest voltage for the device I tested it became even better than for the MCP-PMT in our setup, which is usually not the case. See the plot below.
To conclude, just to mention that this design should be easily scalable to an arbitrary number of channels, as long as there is space in the board to put the components.