Hi Jeff, I was a Fermi for 3 days last week but I have made a little progress on the design of the electronics for the fast IR camera. My desire for separate Analog and Digital boxes on the dewar has gone away. I don't think that separate boxes will help very much just because of the amount of digital stuff that needs to be in the Analog box anyway. I do still think that there is an advantage in separate Digital and Analog connectors going into the dewar. My current count show 20 "wires" going into the Dewar: DIGITAL "Wires" into the Dewar (8 "wires" total) Connects to "Wire" Function TCM2620 Pin --------------------------------------- ------------- +5V Digital VDD +5V Digital Return VSS TCM2620 Master Clock CLOCK TCM2620 Master Clock Return VSS Integration Time Control INT Integration Time Control Return VSS Sample and Hold Time Select MODE Detector Integration Capacitor Select CAPSWB ANALOG "Wires" into the Dewar (12 "wires" total) Connects to "Wire" Function TCM2620 Pin(s) ---------------------------------------------- ----------------------- +5V Analog COLHI, AMPHI, ABAMPHI +5V Analog Return MUXSUB, COLLOW Detector Substrate Control Voltage (variable) DSUB Deadtime Output Control Voltage (fixed) VCLAMP Video Analog Output #1 OUT1 Analog Output #1 Return COLLOW Video Analog Output #2 OUT2 Analog Output #1 Return COLLOW Video Analog Output #3 OUT3 Analog Output #1 Return COLLOW Video Analog Output #4 OUT4 Analog Output #1 Return COLLOW To this count of 20 we should add a couple of spare pass throughs in both the Digital and Analog connectors just in case we need to get our hands on one of the analog or digital "Test" signals. From the electronics point of view, the layout of the "ROCK" card that supports the TCM2620 looks straight forward. The TCM2620 pinout is layed out in such a way that it makes the electrical design of the ROCK very easy. I currently think that 6 capacitor setups will be needed in the cold next to the TCM2620. These 6 capacitors are to bypass (or filter) the following TCM2620 pins: AMPHI, COLHI, ABAMPHI, DSUB, VCLAMP, and VDD 31 of the TCM2620's 68 package pins are not connected to the chip inside. Should these 31 socket pins be attached to a heavy thermal plane in the ROCK printed circuit board and then anchor this thermal plane to the cold finger in the dewar ? Is this kind of "all metal path" a help in carrying out the heat ? My current sketch is for 5 circuit boards outside of the Dewar. There would be 1 digital card (D_Card) and 4 identical analog cards (A_Card). The idea of 4 separate analog cards is both to control the analog ground currents and also to make the package that hangs on the dewar more compact. If we make just one large analog card then the package that hangs on the dewar will need to be more pizza box shape than the cubical shape that one can obtain by using 4 separate analog cards. I have sketches of both the A_Card and D_Card but I do not have a size estimate yet. Now for the scary part. I have added up the various current requirements and calculated the power that will be dissipated in the package that hangs on the dewar. The current heat estimate is 50.4 Watts. This will come down a little, perhaps to 40 Watts, but this is still a big number. This large heat dissipation comes about for two reasons. 1. We have 4x ADC circuits plus digital drives back to the VME environment, and 2. There is "Linear" power supply regulation right on the A_Cards and the D_Card to clean up the power supplies after their long cable hop from the VME power source. This linear power supply regulation dumps about half of the heat. My concern here is that 40 or 50 Watts in a small box requires a fan to cool it. Can a telescope tolerate a fan vibrating right on the dewar ?? How is the cooling problem handled in the other telescope cameras that you have worked with ? Dan