Penn Array Interfaces Internal Meeting 07may03 attendance: Dennis Egan, Bob Anderson, Melinda Mello, Roger Norrod, John Ford, Richard Prestage, Phil Jewell, Brian Mason ======================================================= 1. Weight, Balance, & Mechanical Interface Dennis and Bob summarized our current understanding of the turret as regards weight and balance. Based on the information that Tim has the weight limit of the turret is 3000 Lbs and we are currently 50% over this. The weight is the main structural concern, and balance is not as much of a conern. Balance mainly affects the pointing, which is a concern because we have a more or less static pointing model but a potentially changing suite of receivers. Richard pointed out that if these are repeatable then they can be taken out (however see previous statement re: static model-- referenced pointing will help though). The problem of cantilevering the electronics package is not solved and may be sticky. We may be able to use composites for mounting brackets etc, although mating these to the package may require some work. Bob pointed out the need for some weight management on the antenna. It was thought that at the level currently discussed (~400 Lbs) the Penn Array could be accomadated strictly from the loading point of view but there are larger issues to be addressed, and independently, we need to know what we really can accomadate. If weight or balance are limitations we could take out a receiver (S band?). Long term we may also be able to minimize distortions by relocating the turret pin to a point not opposite the receiver in use. The new structural engineer should be able to help by modelling these things. The issue of aluminum vs stainless steel was discussed. The possibility of constructing an aluminum test dewar was raised, as well as curing techniques which may mitigate outgassing & leakage. The consensus was that aluminum is likely to be a long-term operational headache. Roger will check with some people at other observatories who have used aluminum in cryostats before, but we will point the Penn group at stainless steel since it is best to be conservative with the cryogenics. The weight budget for the Penn Array is useful but currently rough, and the contribution of many smaller components needs to be allowed for (cryogen in ptc; ptc & power supply!; mounting brackets; lenses & other optical elements; lakeshore scanner; contingency for filterwheel). [Note: after the meeting Roger caught a 40 lb error in the Aluminum "cryostat total" which should be 187 not 147 lbs] Phil indicated that at this point it's our responsibility to take what Penn has provided and give fairly definite guidance as to what is acceptable, what is not, and what the limits we're working to are. ACTION: ------ -Dennis, w/Roger & Brian: over the next week look at the turret loading issue and come up with a) a weight limit for our own reference; b) a weight limit for Penn to work within. -Roger: at convenience, contact experts at other observatories having experience with aluminum dewars. -Brian: inform Penn we will go with stainless steel. Check with Penn about the possibility of mounting the electronics rack on the ceiling so as to reduce cantilever issues. 2. Electronics Interfaces/Packaging Concern was expressed over the length of the dc power lines they want to run through the wrap (some concerns are: dc ground losses, noise, low voltage, high currents). The mil-spec AMCO rack is unnecessarily heavy, in part because it is large compared to the components it must enclose, and in part because it is sturdier than we need. John volunteered us to build a custom, lighter enclosure that will meet the RFI spec, once we know what is to be enclosed. The reason for having so many programmable power supplies was not clear, and there was serious concern about the possibility of blowing up the system by setting a bad voltage. There was concern about using GPIB to control the cal diode instead of a solution more suited to realtime operation (eg, TTL from the VME crate or RS232). From the data analysis/science point of view we could live with this but it's not optimal. We also discussed the possibility of going back to having all the electronics in one rack. The pulse tube cooler power supply doesn't appear on the schematic. The electronic interface for the filterwheel is not currently specified and must fit within the scheme we are working to finalize now. (Phil points out that filterwheels were a chronic "sticking point" for the JCMT and we might be able to learn from their experience) 3. Software interface We agreed to a device-driver level interface to the penn array. These drivers currently are a) GPIB; b) the custom fiber readout card. Melinda notes that most of the Agilent devices in Penn's latest schematic also have RS232 interfaces so the cal diode (or other devices) might be controlled that way. Richard is concerned about troubleshooting & trouble-response procedures: we need to i) determine what is likely to need special attention ("special" is anything you don't do, basically, as part of your standard observing procedure); ii) ensure that there is a reasonable mechanism for doing this. Spectrometer self-tests would be an example of *un*-reasonable "special procedures". Phil pointed out that anything we need to do on any routine basis (eg loadcurves, other self-tests) we must get concrete prescriptions from them for, if not actual software in addition. Bolometer loadcurves are an example. The DAQ interface is currently unspecified and we need a documented driver; the cal diode is also only notionally specified and more details on its control are needed. We would also like to know what is in the VME crate. ACTION ------ -John, Melinda, & Brian: work with Penn to agree upon a scheme (packaging & interface) for the electronics.