High precision electron accelerator experiment with 4 RPs

Tell us how your experiment is running with Red Pitaya
Post Reply
Mutabimini
Posts: 7
Joined: Mon Oct 05, 2015 2:09 pm

High precision electron accelerator experiment with 4 RPs

Post by Mutabimini » Fri Oct 23, 2015 4:47 pm

Hi people,
finally I made it to this forum. In the last months I started to believe, that my (and others) work with the red pitayas may become more efficient when I get in touch with other rp users.
Following is my project I want to share with you:
I use (so far ) four RPs to stabilize a beam for an upcoming electron accelerator. The system is tested at an already existing electron accelerator.
The FPGAs shall provide both a feed forward as well as a closed control loop. For this in the next weeks I will modify the Design. The four RPs need to get the data totally synchronized. For that I need to start the data acquisition with the trigger and not by software arming. I will never use the 125MHz, so that I would always have some jitter of the size of sample distance. ( At the moment it's stopped by the trigger, not started). We could already succesfully stabilize a beam with the in-built PID-controller for demonstration purposes.
I modified an acquire to get data from the adc and save it to a file on a network disk. Red Pitayas are already running as a Run Control server client and are remote controlled. The trigger to synchronize the data is set up as pulses from a Spartan 6 evaluation board connected to all RPs. The power supply is solded to the RPs and has very low ripple (required for high precision).

What I am now on is the new ecosystem which is totally different than the one before. I wanted to load a bitstream but struggle with customization of network configs and profile on the /opt. When I manage this problem, I will squeeze out 2 bit more from the adc averaging. For some reason they are cut in the design and only 14 instead of 16 bit are given to the cpu...

Nils Roos
Posts: 1441
Joined: Sat Jun 07, 2014 12:49 pm
Location: Königswinter

Re: High precision electron accelerator experiment with 4 RP

Post by Nils Roos » Sat Oct 31, 2015 1:54 pm

Hi,
I would be very interested to see how you implement the synchronization of your 4 RPs.
What I am now on is the new ecosystem which is totally different than the one before. I wanted to load a bitstream but struggle with customization of network configs and profile on the /opt.
I believe issues with the network configuration are currently being worked on. Since the new ecosystem is still in beta, some hickups are to be expected.
When I manage this problem, I will squeeze out 2 bit more from the adc averaging. For some reason they are cut in the design and only 14 instead of 16 bit are given to the cpu...
I also found that a worthwhile addition. Take a look at my modified scope code if you like an example implementation.
All in a day's work for Bicycle Repair Man

Mutabimini
Posts: 7
Joined: Mon Oct 05, 2015 2:09 pm

Re: High precision electron accelerator experiment with 4 RP

Post by Mutabimini » Fri Nov 06, 2015 5:23 pm

Hi,
thank you for the answer.
I switched already to older ecosystem, for me it's better not to go through all the issues.

I got already my 2-additional-Bit-solution, they have been reserved in the design. Even in the contraints I just had to uncomment those two bits.

I
would be very interested to see how you implement the synchronization of your 4 RPs.
I connected the RPs all together to a Spartan6 evaluation board, where is running a trigger design, which is controlled by I2C-bus from our run control. DAQ is then triggered by external trigger.

berndj
Posts: 12
Joined: Thu Nov 12, 2015 3:32 pm

Re: High precision electron accelerator experiment with 4 RP

Post by berndj » Tue Apr 26, 2016 9:10 pm

hi,
I have a naive question... what if you sync the clocks on the RPs and say I want to start my measurement at that particular second. What do you thing would be the disadvantage? I would guess you could somehow sync the internal clocks easier and independently than four RPs at the same time. It would probably also be easier to verify if you have an internal control mechanism, no?
Just a curious question, sounds like a very interesting project otherwise...
Best,
Bernd

jerzydziewierz
Posts: 31
Joined: Mon Mar 28, 2016 1:10 pm

Re: High precision electron accelerator experiment with 4 RP

Post by jerzydziewierz » Tue Apr 26, 2016 9:47 pm

if you sync the "wall clocks", they will be synced to, e.g. 0.5 second accuracy,

a way better accuracy is needed i many applications,

more over, the crystals residing on separate boards will have an ever so slightly different frequency and drift. I am not sure about the accuracy of the particular crystal used in RP, but it is common for cheap crystals to drift more than 0.1% from the design frequency in matter of few hours. that might be ok for knowing when your lunch break is, but it may be too much for some measurement applications.

e.g. if you are sampling the signal at 10MHz, you would like the clocks to be synced down to 1e-7 second . . . . and then keep them synced

Post Reply
jadalnie klasyczne ekskluzywne meble wypoczynkowe do salonu ekskluzywne meble tapicerowane ekskluzywne meble do sypialni ekskluzywne meble włoskie

Who is online

Users browsing this forum: No registered users and 0 guests