The QA procedure

1. Carefully put the assembly on the detector mount on the platform of the laser station and connect the MCM output cable to the J6 of the power/comm board. Be sure to wear a ground strap and use a vacuum pen to pick up the Si-strip part.

2. Slide in the conductive rubber under the kapton cable where the small pad is ,which the bias tab will be connected to. Squeeze down very gently the kapton cable and the conductive rubber against the copper surface of the detector mount with small clamps and screw down the Si-strip detector on the detector mount. Finally, a Si-bias connection is set through the conductive rubber and the surface of detector mount.

3. Shield the whole things with the steel box.

4. Turn on everything.

• motherboard : from the low voltage distribution module #3(+-5V)
• MCM : from two separate stand alone supply(LAMDA LPD 421, HP 6205C)
• Si bias : ~35.0 V from a stand alone supply(HP E3612A)
• laser bias : ~33.0 V from a stan alone supply(KEITHLEY 487)
• water cooling for the laser : outside of the clean tent
• MCM test board : ~6.0V from a supply under the laser station
• CAMAC crate : used to handle the LVL-1 delay
• stepping motor controller, a cooling fan for the MCM, etc

5. Start MCM control softwares(pcb_mcm , mcm_read, mvd_main ) on mvdonl5(a Pentium-PC for the assembly QA). The mcm_read user interface has a button to pop up a panel for stepping motor controlling, which is named as 'MM3000'. Its usage is self-evident. Using this panel, you can move the laser anywhere you want and monitor the current position.
First of all, download the xilinx and the serial control. Check the readback. Bring the MCM to run state. Now issure LVL1 and see the data packet on 'mcm_read'. If everything is O.K up to now, go on. Detail procedure can be found here.

6. Find the first ch of the Si-strip by manualy moving the laser.

7. Move the laser at the position where the first channel shows the biggest ADC value. Reset the stepping motor controller. From this point, the scan starts.

8. Start the automatic scan moving 100 steps at a time. (Approximately 100steps correspond to the strip pitch. Look here.) It takes about 2 hours to scan whole 256 strip with 10 triggers on each strip. Automatic scan can be done by running a script in MVD MCM software. I have added new commands to control stepping motor in MCM software.

Type the control file name(run11.txt;256 channels at a time, or run11_64.txt; 64 channels at a time)in the small text window of mvd_main . You need to change the hist filename before running the control file. Use the 'Edit Ctrl File' menu to do this. Finally click 'Exec Ctrl File' to launch mcm_read (if there is already a mcm_read running, close it first.) and start the scan.

9. As soon as the automatic scan is finished, take the pedestal data of 500 event with keeping everything on and moving the laser away from Si-strip.

10.Analysis data files using ROOT. Simple ROOT macro files c:\mvd\histos\root\pedestal.c and c:\mvd\histos\root\analyze1.c(analyze2.c if you scan 64 channels at a time) on mvdonl5.lanl.gov, the Pentium PC dedicated to the laser station, determine bad channels. Before running these macro files, you should edit 'c:\mvd\histos\assembly_num.txt' which should contains the proper assembly number. These macors save pedestal data and bad channel infomation on the disk with the file name '[assembly_num]_ped.dat' and '[assembly_num]_bad.dat'. Beware these macros are so stupid that it can make wrong dicisions. Bad channels should be rechecked by carefully scanning the laser around bad channels.

11. If you see a bunch of dead channels and a Si short is there, then maybe most of them are not dead. They are just affected by the shorted channel. We can revive them by disabling the shorted channel. We can do this by cutting the trace or by disabling the preamp channel correspoding to that shorted channel. A preamp channel can be disabled by turning a bit of PRE_DIS of TGV serial string.

12. Sometimes unexpected bad things can happen during the scan. For example, the laser spot can go into the readout strip during the scan, which reflect the laser. In that case, you may have to go back to the first channel and repeat the procedure again.

NOTE: All dead channel numbers are in MCM convention, 0 - 255.

Assemblies with evaluation lot MCM's:

First batch(?):

Second batch:

Third batch:

Statistics

Fig.1 Summary chart.

Total 35 assemblies have been received.
4 out of 35 have one block failure or lots of dead channels.(not included in the bar chart.)
1(IT-14) failed serial download.
1(IB-06) unstable.