Hardware 

The MEC WAVE8 IOCs run on server ioc-mec-wave8 located in R64B EL40 in the laser area. Connections into the server PGP card are made with a 4 fiber-pair MTP cable which interfaces to a patch panel in the back of R64B. The matching panel panel is on the south side of the beamline in rack PR61. The two EVR signals are taken from PR60.

IPIM Setup

ipmConfigEpics

EVR Setup

MECHome/EVR/SLAC EVR for Users

For IPM2 we connected to EVR0 and EVR4 on the PR60 patch panel.

Note: Here is the previous EVR setup as captured at 16:19 on 2 July 2025.


Detailed Set-up Procedure - from 8 July 2025 Photon MD Time (under edit)

Changing the Peak Index

  • The peak index is specified in ticks. Once it is changed then hit "configure".
  • Changing the peak index alters the location where the IOC picks off the value and sends to BLD.
  • In order to move the curve, you will need to update the WAVE8 Trig 1 delay in EVR, which has a nominal value of 83000.
  • The peak was just off the end of the displayed waveform. In general change the peak index instead of the delay…but either would in principle work.


Bias and Rise Time

  • Adding a bias  would likely make the rise time sharper…but if the peak looks like that without bias, then it might saturate with bias.
  • The peak index needs to come close to where the rise occurs
  • And confirming that ch12 is our diode.

Correlation with the Gas Detector

  • I’ve often had trouble correlating against the gas det…but they should roughly correlate. In this case the waveform shows values up to 250000 but the channel 12 shows 50ish.
  • Possible hardware issues:
    • If there’s something clipping the beam between the gas det and the diode, the correlation will be poor
    • You should also try to look at hx2_ipm1 for correlations - it's less noisy and somewhat closer to MEC (though not a lot)
  •   Software issues:
    • That means the peak index is looking at the wrong bin
    • So try to estimate where the rising point of your waveform start

FIR Filter

Here is how you can see what the IOC is looking at when it tries to grab the peak value. It's a pretty narrow plateau.  You need to set the peak index to within that plateau.

  1. Open ipmConfigEpics
  2. Under waveform FIR, you can find the path to the FIR filter. 
  3. Goto ami, click on the channel corresponding to your diode
  4. Hit "apply FIR filter"
  5. Grab the FIR filter from the path found under ipmConfigEPICS
  6. Eliminate all other channels from AMI

Notes:

  • amplitude or peak finding?
    • There should be two paths, but they should be identical
    • It's likely pointing to somewhere in my home directory under ~nakahara
    • Unless I was smart and copied it to ~mecopr...which is unlikely.
    • right your patOk, you'll need that path when you try to apply it in AM
  • 250000 is saturation
  • it is not clear where to find how to apply a filter in AMI for the channel
    • Click on "ChA" on the waveform screen, for example.
    • I am in Env, then select the source of the channel, 'PEAK_A:CH12'
    • right, this is channel E
    • Ok, so chE
    • Try applying the FIR filter from my home director and uncheck all other channels since they will get in the way
  • Note from Eric at this point: however, the signal jumps from positive to negative



Zooming in

Disable autoscale, and zoom in until you see and resolve the plateau. You need to find the bin where the max appears along the xaxis



Horizontal axis is in seconds
Once you know the X value of the peak, the conversion should be X*4096/819200...


More on Sampling

  • What sampling are we using?
    • After discussion this was 88*4096/819.2 = 440, very different from what I had for all other IPM devices...by a factor of exactly 2
  • Should have channel 12, MEC-XT2-BMMON-02:PEAK_A:CH12 showing the peak as if this was on the waveform.
  • You could move the peak index by 1 bin at a time to see if the value increases or decreases.  It should correspond to the peak value of the waveform as seen via the FIR filter. 
  • We had to set the delay on the evr to 95000
    • Ah, that explains why the peak index is so different. 
    • In that case, you might want to just keep 95000 for this application so you don't have to re-do the peak finding process.
      The curve was really off on the wave8 so I think the evr needs to stay where it currently is






November 2025

Context: Install WAVE8 for MEC 1012351 Descamps ECS-9105 - Getting issue details... STATUS .

Update of engineering_tools/ipmConfigEpics: https://github.com/pcdshub/engineering_tools/pull/324/files


peregrin@mec-daq  R3.0.1]$ more ioc-mec-usr-dio.cfg 
RELEASE=/cds/group/pcds/epics/ioc/common/wave8/R3.0.0
ARCH=rhel7-x86_64
ENGINEER=Peregrine McGehee (peregrin)
LOCATION=MEC User Diode (floating)
IOC_PV=IOC:MEC:USR:DIO

EVR(NAME=MEC:USR:DIO:EVR,TYPE=SLAC)
PGP(NAME=PGP0,MASK=0x40,G3=1)
WAVE8(NAME=MEC:USR:DIO,BOX=PGP0,EVR0,TRIG=0,AUTO=0,BLDID=93)
RUNTRIG(EVR0,TRIG=1)
peregrin@mec-daq  R3.0.1]$ 










  • No labels