From r.bainbridge@ic.ac.uk Thu Jan 2 17:53:50 2003 Date: Thu, 2 Jan 2003 11:41:13 -0000 From: Rob Bainbridge To: Anthony Allen Affolder Cc: mark Subject: Re: APV25 testing for final production [ The following text is in the "iso-8859-1" character set. ] [ Your display is set for the "US-ASCII" character set. ] [ Some characters may be displayed incorrectly. ] Hi Tony, the documentation that Mark mentions was never finished, but I have summarised the selection criteria you require below. 1) The baseline position is adjusted to a level corresponding to 1/4 of the available signal range (defined by the digital header) by varying the VPSP parameter. The APV then is (randomly) triggered 1000 times and the channel pedestals calculated. All channel pedestals must fall within a range of 50 - 90 ADC units (8-bit ADC). At the same time, the channel noise is calculated (and recorded). There is no cut on channel noise. 2) The APV is repeatedly triggered so that all 192 pipeline columns are read from. The pedestals of the 128 x 192 cells are then calculated. The rms of the 192 pedestal values for each pipeline must be less than 2 ADC counts. (There is an additional cut on the pedestals: no single pedestal must deviate more than 5 ADC counts from the mean of the 192 pedestals of a pipeline.) 3) Pulses are reconstructed using the calibration circuitry and the maximum gain is recorded for each of the 128 channels. A minimum gain of 20 ADC counts is required of all channels and all channels must be within 15 ADC counts of the chip-averaged gain. All the above are performed for both peak and deconvolution mode operation. There is a talk (link below) which summarises the results of some of the above tests. http://www.hep.ph.ic.ac.uk/~rjb3/PhD/Talks/APV25ProbingResults_TK_110701.pdf If you have any unanswered questions, then please get in touch... cheers, Rob ----- Original Message ----- From: "Anthony Allen Affolder" To: Sent: Monday, December 30, 2002 11:23 PM Subject: Re: APV25 testing for final production > Hi, > Mark Raymond said you may have a document listing the wafer test > criteria for the APV chip. I'm looking for the bad channel cuts during > wafer probing. I'm trying to figure out all the selection criteria for > components before shipment to the US. > > Thank you and happy holiday, > Tony Affolder > > On Thu, 19 Dec 2002, Mark Raymond wrote: > > > Hi Tony, > > > > I have passed your questions on the wafer probe test cuts to Rob Bainbridge > > who I think has a document he can send you on this. > > > > Regarding your results on the hybrid. If I was trying to diagnose what's > > going on I would want to look at the output of these chips on a scope. The > > fact that you get the same response for a 0 and 0.5 mip signal and signal > > only appears for 1 mip and greater implies the baseline is saturated > > negative. Actually I don;t really understand why the signal goes negative? > > > > The calibration circuit doesn't have any non-uniformities at the level you > > are measuring. There is nothing wrong with doing a multipoint gain measurement. > > > > It is not impossible that the chip has been damaged from cutting or that > > there is a bonding problem - you might inspect it under a microscope. > > > > Hope this is useful. > > > > Cheers, Mark. > > > > > > At 12:48 AM 12/19/2002 -0800, you wrote: > > >Hi, > > > I have been asked by the US group to compile the list of tests on > > >components previous arrival to FNAL. I've read most of the documentation > > >about APV wafer probing,but I can't find the cut values for most of > > >the > > >tests and am not completely sure what are the tests that will be used in > > >die of the new pre-production modules. If possible, could you please > > >send me or point me to a copy of the > > >selection criteria of a bad die or channel, i.e. a pedestal variation > > >larger than X, a response to calibration less than Y, and noise smaller > > >than Z, etc.? This would be a great help in understanding if our methods > > >are causing problems. > > > > > > I worked on the wafer probing of CDF for a few months > > >and understand the difficulties in grounding and filtering during wafer > > >probing. We just want to understand which problems (if any) we are > > >causing and at which period of production. > > > > > > I also have a unrelated question. I attempted to measure the gain > > >of chips on a prototype hybrid in ARC ADC vs. Mip (as defined > > >as increase of 29 ICAL per > > >Mip). the fit was done using 13 measurements between 0-3 Mip, where the > > >APV > > >response sthould be linear. The first figure shows a gain per channel in Arc > > >ADC/Mip(29 > > >cal)(top) and a chisqure for the fit(bottom). I was surprised by the 10% > > >variation in gain within 3rd chip. Excludin g the ical=0 point, the > > >chisquared is fairly good. including the ical-o the edges of chip 3 look > > >poor. Investigating farther, there seemed to be two issues. The second > > >figure illistrates these problems. The top figure shows the ADC per 1mip > > >injected, which had a response you expect. The middle figure shows the > > >response of the channels with a bad chi-squared. The response to 0 and > > >0.5 mip is the same ( causing the bad line fit). Is a ical=0 setting > > >bad, or does it indicate a large off set in the calibration curcuit. The > > >third figure is a channel with a 5-10% lower gain than the rest of the > > >channels. Again is this a concern or a known problem with nonuniformity > > >of the cal. inject capicator? Is a multi-point gain measurement useful > > >or does the uncertainty in the circuit make it pointless? the chisquared > > >of the test seemed to find channel will low response to low cal inject > > >setting. It would also find non-linear responses to charge. > > > > > >Any comment on the wafer probing and the calibration circuit is greatly > > >appreciated, > > > > > > Tony Affolder > > > > > > > > > > ------------------------------------------------------------------ > > Mark Raymond > > Blackett Laboratory > > Imperial College phone: +20 7594 7799 > > London SW7 2AZ fax: +20 7823 8830 > > ------------------------------------------------------------------ > > > > >