New MCA
Posted: Sat Feb 14, 2009 3:18 am
I have just massively upgraded my pulse-height analysis equipment, thanks to some throwaway NIMs and some wonderful open-source software.
The new system comprises a Canberra 556 AIM module and a Canberra 9635 ADC (both single-width NIMs). These modules and their relatives from Canberra's proprietary-standard "ICB" NIM family are often on eBay or in other surplus quite inexpensively, and I'm guessing it is because people do not know how to use them or assume that the Canberra Genie(R) software is needed to operate them. Indeed this was my initial assessment of the package, but I'm not one to turn down good loot. It turns out that (A) the cabling needed to attach the ADC to the AIM is straightforward; (B) all you need to get the AIM connected to a PC via ethernet is a cheap IEEE 802.3 to 10-BASE-T transceiver; and the real kicker is that (C) there is FREE, OPEN-SOURCE SOFTWARE available to collect real-time data from Canberra AIM modules.
Let me just describe in briefest terms how the 9635 ADC stacks up against some other ADCs I have experienced in hobby settings. The conversion time is calculated for the photopeak of the spectrum shown, in hypothetical channel 1076 (provided the unit has enough memory). Conversion time is inversely proportional to the maximum count-rate the ADC can process.
ADC / MCA...........................Memory.........................Circuit...............Conversion time
Tracor-Northern 1705.............1023.....................50 MHz Wilkinson...........(20 us)
Canberra 35 Plus....................4096....................100 MHz Wilkinson..........10 us
SpecTech UCS30...................4096....................(Fixed-time ADC).............10 us
Canberra 9635........................8192....................(Fixed-time ADC).............0.9 us
Another way of looking at this is that the 35 Plus and Spectrum Techniques UCS30 would be able to beat the 9635 in count-rate capability only if the peak in question were placed below Channel 96 in those units.
The software I needed to put on my Windows computer was the following, all of it free:
-Cygwin/X (Linux emulator)
-IDL Virtual Machine
-Dr. Mark Rivers' "mca" binaries, here:
http://cars.uchicago.edu/software/pub/mcaPrebuilt.tgz
-EPICS binaries for Windows, here:
http://cars.uchicago.edu/software/pub/E ... naries.zip
-Dr. Mark Rivers' MCA GUI for IDL, here:
http://cars.uchicago.edu/software/pub/idl_mca.tar
I'll give a fuller set of instructions on my blog sometime soon. But for now I will say that I'm incredibly indebted to the hard work of Dr. Rivers and all the other contributors to this fine open-source software, especially the EPICS system. Setup could be a little hairy for some, but the end result is a real-time graphical interface to a top-quality MCA, for ultracheap. The pic just shows this software running, with a Cs-137 spectrum on a 2" NaI(Tl) detector with about 5% resolution.
-Carl
The new system comprises a Canberra 556 AIM module and a Canberra 9635 ADC (both single-width NIMs). These modules and their relatives from Canberra's proprietary-standard "ICB" NIM family are often on eBay or in other surplus quite inexpensively, and I'm guessing it is because people do not know how to use them or assume that the Canberra Genie(R) software is needed to operate them. Indeed this was my initial assessment of the package, but I'm not one to turn down good loot. It turns out that (A) the cabling needed to attach the ADC to the AIM is straightforward; (B) all you need to get the AIM connected to a PC via ethernet is a cheap IEEE 802.3 to 10-BASE-T transceiver; and the real kicker is that (C) there is FREE, OPEN-SOURCE SOFTWARE available to collect real-time data from Canberra AIM modules.
Let me just describe in briefest terms how the 9635 ADC stacks up against some other ADCs I have experienced in hobby settings. The conversion time is calculated for the photopeak of the spectrum shown, in hypothetical channel 1076 (provided the unit has enough memory). Conversion time is inversely proportional to the maximum count-rate the ADC can process.
ADC / MCA...........................Memory.........................Circuit...............Conversion time
Tracor-Northern 1705.............1023.....................50 MHz Wilkinson...........(20 us)
Canberra 35 Plus....................4096....................100 MHz Wilkinson..........10 us
SpecTech UCS30...................4096....................(Fixed-time ADC).............10 us
Canberra 9635........................8192....................(Fixed-time ADC).............0.9 us
Another way of looking at this is that the 35 Plus and Spectrum Techniques UCS30 would be able to beat the 9635 in count-rate capability only if the peak in question were placed below Channel 96 in those units.
The software I needed to put on my Windows computer was the following, all of it free:
-Cygwin/X (Linux emulator)
-IDL Virtual Machine
-Dr. Mark Rivers' "mca" binaries, here:
http://cars.uchicago.edu/software/pub/mcaPrebuilt.tgz
-EPICS binaries for Windows, here:
http://cars.uchicago.edu/software/pub/E ... naries.zip
-Dr. Mark Rivers' MCA GUI for IDL, here:
http://cars.uchicago.edu/software/pub/idl_mca.tar
I'll give a fuller set of instructions on my blog sometime soon. But for now I will say that I'm incredibly indebted to the hard work of Dr. Rivers and all the other contributors to this fine open-source software, especially the EPICS system. Setup could be a little hairy for some, but the end result is a real-time graphical interface to a top-quality MCA, for ultracheap. The pic just shows this software running, with a Cs-137 spectrum on a 2" NaI(Tl) detector with about 5% resolution.
-Carl