Profile cover photo
Profile photo
Monday Geekery
The geekier side of a particle physicist's work.
The geekier side of a particle physicist's work.

Monday Geekery's posts

Post has shared content
Only a day late, here are some details on how I got a bunch of digitisers to work together.
Getting multiple digitisers to work together

This post is a little technical and I doesn't explain my experiment. I plan to give an overview soon, maybe in a hangout on air.

My data acquisition system is made up from a number of digitiser boards. Each one has eight channels and converts analog voltages given out by our light sensors into digital values that are transmitted to a computer. To run our experiment properly I had to synchronise the boards and distribute the triggers from one board to the other boards.

For the synchronisation I define one board to have the master clock. This board outputs its clock signal, which is fed into the next board. All the other boards are slaves. They use the clock input from either the master or another slave as their reference clock. They also output the same clock so that all the boards can be connected in a chain. There are some official cables to do this, but they are expensive and the delivery time is too long, so I searched our electrical workshop and found some suitable connectors and made up the cables myself. You can see that the synchronisation is working in the three plots. These show the time difference of events between the first board and each of the others that are used. With no synchronisation the difference changes a little from event to event. When the boards have their clocks synchronised they have an offset in time, which is due to waiting their turn for a start signal, but importantly this offset doesn't vary with time.

The other issue was to share the trigger signals between boards. Most of the time our light sensors don't see anything, but we check them at 100 MHz. We can't keep all that data, so we only trigger the readout when we see some signals go over a set threshold. Each board can independently trigger when one or more channels exceed the threshold. When any board triggers we want data from all the boards, so we have to share the trigger signal around. We do this by sending out voltage pulses when a board triggers, taking the OR of all these signals and feeding them back into the boards. I searched around the lab and found another board that can perform the simple logic, but it uses an ancient electronic crate standard, so I actually spent more time looking for a crate to power it than I did for the board. Anyway, it works fine, as you can see in the short video. The trigger LEDs of all the channels flash even though only one has a signal to trigger on. WARNING: the video is kinda loud due to the fans in the VME crate. Sorry, I couldn't work out how to mute my camera's mic. Be glad you don't have to sit next to this noisy thing all day...
8 Photos - View album

Post has shared content
I've finished with my thesis so I'm getting back to fun and geeky work...
New experiment, new toys

I've finally started working on the SoLid experiment full time. We're building a small (just over 1 cubic metre) neutrino detector to place right against the wall of a nuclear reactor. This week I've been playing with a crate full of fast digitiser cards. These will convert the analog signals from 32 channels of silicon photomultipliers into digital data that can be read into our data acquisition computer. The boards on the right hand side of the crate are all Caen V1724s. They each contain 8 fast digitiser channels. The board on the very left is an optical readout board that connects to the computer.

So far I've been setting up the computer and writing some basic software to use the Caen libraries to control the boards and read the data over the optical link. The libraris are written in C and my data acquisition program will be written in Go, so I've had to use the cgo package to call the C library functions. The library is fairly complete, so I've only had to resort to directly reading or writing the registers on each board a few times. 

Now that I've got the basic functionality working I have to design the overall acquisition program. This will configure the different input channels and triggers, read out data and save interesting events to disk so that we can analyse them. It will also control the bias voltages for the silicon photomultipliers as well as providing a way for us to monitor and control data taking. This system is going to replace a less powerful one that is currently running as our prototype detector is deployed at the reactor and taking data. Hopefully I'll have this system ready in time for the next reactor run cycle.
2 Photos - View album

Post has attachment
Trigger boards are here

After a couple of days to get from China to Oxford, and then a few more days to get from Oxford to my college in Oxford, my trigger boards are here. They look really good. 10 boards cost me about £30 and took about a week and a half for fabrication and delivery. I did realise one mistake that I had made in the design, so I need to put a jumper between the digital to analog converter's (DAC) latch pin. I soldered up a DAC to one of the boards but didn't quite finish writing the software library to try it out yet.
PCBs - 4 February 2013
3 Photos - View album

Much overdue update

I've got a final year undergrad working on the cosmic ray detector now, so hopefully the development will start speeding ahead again. I designed some trigger PCBs that should be on their way over from China by now. I got them made by, who have a special of 10, 5 cm by 5 cm boards for ~£13. When the boards arrive the plan will be to start searching for coincidences between two SiPM channels, so hopefully we can pick out cosmic events from the sensor's high dark count.

Post has attachment
Sending data from a remote sensor

Over the weekend I was playing around with methods for sending data from an #mbed  to a server over 3G. I'm not sure at the moment whether this type of system would be good for a network of cosmic ray detectors.

Post has attachment
TWEPP Paper Submitted

I didn't get into the lab today, but over the weekend I got my paper for the Topical Workshop on Electronics for Particle Physics submitted. The paper gives some details of the cosmic ray detector I'm building. Now that it is submitted I need to wait for comments from a couple of reviewers, then hopefully it will be published near the end of the year or early next year. TWEPP proceedings are published in the Journal of Instrumentation ( and I think they are freely available to the public.

I also found out last week that I will be supervising a final year undergraduate student next term. He will be working on the cosmic ray detector and hopefully making a high altitude measurement. This means I need to get the analog readout PCBs made up so that we can get two channels running when he starts to try and detect a coincidence signal between the two to show that they really can detect cosmic rays.

Post has attachment
Getting online using 3G

The #mbed  team released their Vodafone 3G library a couple of weeks ago (see ) . I'd bought a dongle some time back when it was in beta testing, but only just  got around to testing it out.

Setting up the circuit for the mbed should be pretty easy since it simply needs a regulated power supply, a USB socket and some pull down resistors for the USB data lines. Unfortunately my dim eyes picked out the wrong resistor at first, so I spent a decent time scratching my head about why it wasn't connecting. With that problem fixed it was simply a case of trying a couple of the different networks until I found the correct one for my data plan (turns out it's £5 for 30 days, up to 250MB).

I plan to have the 3G be a backup connection method for either my cosmic ray detector or the flight computer to send back a decent amount of data after landing in case the detector lands somewhere inaccessible. They could also be used for lab based experiments if a wired connection isn't available but this would probably require the detector to be plugged in since these dongles can draw a decent amount of power (I'll have to measure how much they really use, but they can take the maximum allowed USB power when trying to connect or if they have a bad signal).


Back to work

You may have noticed I haven't posted anything for a while. That was because I was away getting married and then honeymooning over in Hawaii (which is an awesome place, I suggest you visit if you can). However I'm now back in Oxford and starting to work again. The first thing to do is to write up my contribution for the proceedings of the Topical Workshop on Electronics for Particle Physics, which I went to just before I left to get married. It will be a short paper giving some more details of my prototype cosmic ray detector. I have to submit the paper in the next few days and the proceedings should be published in JINST in a few months.

Post has attachment
Today I put together a poster about my cosmic ray detector for the Topical Workshop on Electronics for Particle Physics that we're hosting in Oxford next week. Here are some photos I took for it.
4 Photos - View album

Outreach Project

I've finally found some Monday Geekery time to work on my detector today. I'll be getting into the lab later in the afternoon, but first I am working on an application for an outreach grant.

I'm asking the STFC for some money to deploy my detectors in local schools for GCSE (15 year old) students to do some experiments in their schools. We will also get the students to launch, track and recover detectors launched on high altitude balloons to measure the cosmic ray flux at high altitude. Along with the experimetns the kids will do I'm also planning a series of online lectures related to cosmic ray physics and the experiments they will do.

I need to get the application done pretty quickly (before I go on a break to get married), but then won't hear whether we receive the grant until late this year. I think the school project would be awesome fun for the student but I find these applications where you really have to sell yourself and the project tricky.
Wait while more posts are being loaded