Servicing Record for new ATI/UW sonics: 1999: CASES99: ~8/15 all ATIK shadow corr. changed from 16% to 0% 9/27 0504 Spiking 10/10 0505 Spiking 0303? changed to 20 samp/s 10/14 0303 dead 10/15 0505 Had been programmed as UW array -- fixed 10/16 0303? blown cap and fuse fixed by Gary 10/17 0302 noisy v 10/27 NUW3 bad v&w 2000: OASIS2000 ~2/15 all ATIK shadow corr. changed from 0% to 20% ~3/1 0302 New V transducer NUW5 bad winds 1 or 2 channels ~3/2 NUW6 broken epoxy joint on ? arm repaired/meas. 3/5 NUW4 removed since bad (how?) ~3/10 NUW5 back from ATI (w/new board?) 3/15 NUW3 New CT transducer 3/24 NUW3 Latest sealing approach - RTV/S&P everywhere! NUW5 Latest sealing approach - RTV/S&P everywhere! 4/4 NUW6 broken epoxy joint on bottom arm repaired/meas. 4/7 NUW3 New AB transducer (was ringing, prob. because RTV had peeled from side on face). Had given speeds ~50m/s on U & W. 4/10 NUW4 Bad board->probably rain damage. Sent to ATI for repair and came back on 4/13. Sent to OASIS operation on 4/14. (Luis) 4/26 NUW7 Sonic arrived with broken epoxy joint in both arms. (Problems during transportation?) Lab. testes shown that path A was damaged. Sent to ATI the same day. Diagnosis-> burned diode at path A transducer circuit. ->The new sonic design is being applied in this unit. (Luis) 4/27 ATI 980202 Malfunction problems. A bad chip (EPROM) was identified and replaced at ATI. The EPROM software was updated. Sent to OASIS operation on 5/1. (Luis) 5/6 980202 Back to NCAR, then returned to OASIS with parity set off. 5/7 nuw7 Survived dunk test!!! Silver soldered vertical tube, new housing reassembled, measured, calibrated, sent off to OASIS 5/9 nuw3,nuw4, nuw6 Vertical tube joints broken. Disassembled, silver soldered, change housing to new design, reassembled, measured, calibrated. Should be shipped ~5/11. 5/11 nuw3 The array head on the non-orthogonal sonic #3 slightly pivots where it is clamped in the boom. To stop this motion I've placed more gasket in the front cap clamping connection. This is a temporary solution because the lack of sonics at the field. The final solution is to straightened the vertical boom of the transducer array (that become bent after the silver soldering). That solution requires new measuring of the distances of the transducers. The immediately shipping solution was discussed with Dr. Horst.(Luis) nuw6 The sonic was tested at ATI in two different days, May 10th and 11th. At 10th no problem was identified. At the following day a bad diode was identified in path A. The diode was substituted but the signal from the whole board was weak. After that, all the diodes from the transmitter circuit were substituted. Then, the working values of the sonic, such as the offset values from the calibration procedures have had improve significantly. Shipped with nuw3 ______________________________________________ Remark! This problem was discussed with Gary and a probable reason for this situation is that the diode start losing quality after something happened either with the board or the transducers. The diminishing of the diode operating quality will probably create the strange behavior of the sonic that has been verified in the last days, i.e., the satisfactory operation of the sonic in some occasions and the non operation of the whole sonic or a specific path in other occasions. (Luis) ______________________________________________ 5/12 nuw4 Transducer AB replaced. Double E-PROM replaced. Sent to OASIS Friday 12th. Programing problems caused the shipping delay. (Luis) 5/18 nuw3 Symptom: a periodic offset on the wind velocities values. After opening the boom, the CB transducer wire was not connected to the circuit board. When the sonic was shipped from Oklahoma it was working, therefore I suspect that the connection from the wire to the circuit board was weak, and broke up when the boom was openned. The wire was the shortest of all sonics. S. Semmer made a new extension and soldered it back again. I discuted this with Gary and he agrees that a weak soldering could originate a bad grounding. The result of a bad grounding is noise and velocities fluctuations. The sonic is being tested on the roof, close to a prop anemometer to exclude other reasons for the offset on the velocity values. 5/22 The test and the analysis of the data show that the sonic is working properely. Shipped to OASIS on the same day. (Luis) 5/23 ATI 980505 Symptom: There was a communication failure after operating a couple of hours in the field. At the Lab. Test, the sonic was only able to provide 3 lines of data. Diagnosis: There was a weak contact on the microprocessor slot (old board design). All the electronic contacts from the microprocessor to the circuit board were replaced by conventional solder. The RS-232 chip was malfunctioning. It was also replaced. Preventive maintenance: To avoid future degradation of the signals from the transducers, all the diodes from the transducer's transmitting circuits were replaced. New conformal coating was done. The level sensor was replaced. (Luis) 5/25 NUW4 Problems: - intermittent bias in w. Resoldered AB xducer wire to PCB. - wrong coordinates for horizontal wind. Non-orth rotatation matrix was set to factory default for Type A probe. Reentered our matrix and set to factory default. (But we had done this before, and it didn't stay set.) This problem may recur. [Gary has confirmed that the setting of factory defaults can be "flaky" generically (not just for NUW4). He'll continue tests there.] We sent this sensor back to OU. 6/10 nuw4 Silver-solder joints broken! Bart reworked by notching the tube to allow the solder to flow into the joint better. Arrays restrung, measured, sealed, calibrated, and sent back to OU. 6/10 nuw6 Like nuw4, but also saw evidence of RTV peeling on transducers. Applied a bit of (white) RTV hopefully to seal. 6/10 nuw5 Epoxy joints horizontal-to-vertical tubes broken (as expected to happen -- this was the last which was still epoxied). Bart silver- soldered using new method. Array restrung, measured, sealed, calibrated, and sent back to OU. 6/10 98???? Brought back. Found to have all weak transducers. Gary is fixing and expects to have ready ~6/23. Will also replace diodes. 6/20 nuw3 Symptom was intermittent bias in w-bar, getting worse. Path B was very weak -- probably bad BT. We see evidence of RTV pulling off the the sides for AT, BT, and CT. Will replace at least BT. Will also inspect solder joints. 6/23 980303 All transducers and diodes replaced. The w path has the new model of transducers. Although the rest of the paths have the old model, they are new. The circuit board was coated with the new poliutetane coating. Sent to Oklahoma for Oasis (Luis) 7/06 Calibration for EBEX nuw3 OK nuw4 Path A with an offset higher than the other paths. After calibration it agrees with the rest of the paths. nuw5 OK nuw6 OK nuw7 OK 980202 OK 980303 OK 980404 OK 980505 OK (Luis) Pre CME04: 4/4-9 All silicon grease applied to transducers (to fill in cracks) zeroed set to 10sps NUWs block screws replaced with stainless (except 5,6) taped transducers; taped block 980202 bad V transducer -- returning to ATI 980302 OK 980303 OK 980504 OK 980505 OK NUW3 OK (I may have replaced a transducer or two more than a year ago.) NUW4 (I may have replaced a transducer or two more than a year ago.) Need to measure array (may have swapped array with NUW7) Bad User Checksum!; Good Factory Checksum! intermittent dies NUW5 need to remeasure array after bending back Intermittent: Bad Factory Checksum! NUW6 Modified for TRAM; needs new block (then reassembly) NUW7 OK Need to measure array (may have swapped array with NUW4) (I may have replaced a transducer or two more than a year ago.)