Filter Labeling

The filter holders should be labeled in accordance with the naming scheme used by the PTF Science Team in order to properly identify the filters. The suggestion is to apply some type of permanent label such as a stamped tin plate riveted to filter holder frame, or a simple engraving directly into the frame. The filters should be identified by both ID and NAME as:

ID	NAME
1	g
2	R
11	Haon
12	Haoff
13	OIII

DH

Filter Assembly

Brief Description: The filter changer assembly is a mechanism which provides for motorized selection of one of two filters. One filter will always be in the field. The mechanism provides access for occasional manual exchange of the two filters which can be carried within the assembly. Motion will proceed under commands received by motor controller via TCP socket.

Items/Tasks remaining until completion:

DONE	EXPECTED DATE	TASK	WHO
X	07/28	Re-write limit switch algorithm	DH
	08/01	Drill & tap hole for motor cable shield drain wire lug	KB
X	08/01	Receive from shop 2nd Base Plate Filter Carriage Assembly	KB
X	08/01	Receive from shop 2nd Filter Frame	KB
X	08/01	Procure fasteners for nut drive assembly	KB
	08/01	RE-Design and RE-fabricate limit switch arm retainers (or simply purchase). Current ones not adequate.	KB
	08/01	Lubricate shafts to reduce shaking/vibration/squeaking (suitable lubricant for proximity to filters?) Roger comments...consider thin film of vacuum grease, which has very low volatility and thus will not evaporate. Interaction with air? This will look downward and will thus be relatively immune to dust accumulation.	KB,RS
X	08/05	Assemble filter mechanism drive assembly and frame holders (for limit switch software test)	KB
X	08/05	Check new limit switch algorithm in assembled mechanism	DH
	08/08	Modify housing for cable clearance	KB
	08/08	Receive from shop Cover Door Assembly	KB
	08/08	Procure Cover Door Hinge	KB
	08/08	Add hinged cover on the open side to stop particles/grease falling from mechanism from landing on primary	KB
	08/15	Fix limit switch #2 -- sometimes sticks such that it doesn't fully close (this may have been due to faulty limit switch arm retainer)	KB
	08/15	Investigate (solve?) stiction (heavy stiction seen in 08/04 test)	KB
	08/15	Final mechanical assembly	KB,DH
	08/15	Re-check software after final assembly (may need to be checked after limit switches in their final place)	DH
	09/05	Assemble motor controller and power supply into enclosure (need to select a suitable enclosure, probably from campus electronics store)	DH
	09/05	Fabricate and fasten name plates or engrave filter labels (see above) into frames	KB

Motion Control

• IMS MicroDrive MForce controller
  • numerous typos in IMS manual
  • when looking at RS422 port, pin 1 is at upper right. See my attached drawing for pinout description.
  • added 0.01 uF capacitors to IO lines for limit switches, otherwise noise from stepper controller could sometimes trigger limits
• 51200 steps/rev
• 1 rot of lead screw = 0.057 inch linear travel (according to KB)
• Perle IOLAN SDS4 Terminal Server
  • terminal server serial ports are 10 pin RJ-45 but most cables are 8 pin -- use middle 8 pins so subtract 1 from pinout description in Perle manual

Control Algorithm

REVISED ALGORITHM 7/28/08 To make up for the lack of hard limits, I've implemented a timeout and distance measuring function. Essentially, the controller will first locate and measure the distance of a filter exchange upon power up (measure in motor steps). It will then move between the soft limit switches and note the distance traveled as a check. If a limit isn't reached within an expected time (about 15 seconds to travel from one end to the other) then the motion will stop. An updated flowchart is attached (page 1 and page2). External software will monitor the reported traveled distances in order to monitor troubles.

There is only one limit switch at each end of travel. Normally one would utilize two switches at each end; the first switch encountered is typically the "soft limit" switch, and this is where the control algorithm would sense that it should not move further. The "outer limit" (2nd) switch is a hard limit -- if this limit is hit then power is cut, typically in a fashion that does not depend on software. DH

A flowchart of the motion control algorithm is attached (page 1 and page2). Note that this is the low level software within the motor controller. High-level software will exist as part of ArcVIEW to send commands to the controller and interpret responses. The high level functions will include the algorithm to decide if the residuals are too high.

• We need to define/refine the control algorithm. Ideally we would have a limit switch to acting as a single bit encoder to confirm arrival in position. The stepper motor drive should be very precise so the limit switch should trip at a repeatable point. If it trips early or late we can use this as a warning of mechanism problems.
• Do we have a way of measuring whether the limit switch has tripped.
  • yes, the state of the limit switch(es) can be read.
• Can we tell where the motor is at the time the limit switch trips?
  • yes, the position (in motor steps) can be read. NOTE: there is no encoder! Presumably the controller is just keeping track of steps sent to the motor.
• Can the motor controller record where it thinks the motor is when it gets this signal?
  • sure, as above, the controller is keeping track of steps and can be queried.
• The limit switch should redefine the home position each time.
  • no, this won't provide an indication of a slippage problem.
• Motor demand position should carry the filter far enough to trip the limit switch at the other end of travel, but not far enough to drive the motor into the mechanical end stop. i.e. do not drive indefinitely until limit switch trips since limit switch failure will cause motor to remain power if end stop reached. If the filter does hit the end stop then under normal conditions the motor will only overdrive briefly.
• Check that collision with mechanical end stop is harmless if motor is shut down within ~1s.
  • OK, but if limit is missed, motor timeout will be about 15 seconds
• What happens if motor is driven into end stop indefinitely. If something overheats or mechanical damage is risked then we need a second limit switch to cut the power to the motor not just provide an interrupt since this condition will only occur in the event of electrical or software failure in which case the normal signaling may not work. If this switch is tripped we have to have a way to back out of the hard electrical limit in spite of that limit being activated. Normally one arranges to interrupt the power for forward drive while still allowing reverse drive but in a microstepped motor this is probably not feasible. It may be acceptable not to have an electrical limit, but to have the supervisory software intervene to command a motor power down in the even that a timeout occurs before arrival. Dave, can you document the algorithm that is implement in some clear way (flow chart?) so that it can be reviewed. -- Roger.
• We must take precautions to make sure that the limit switch does not slip from its correct position. Apply locktite or nail varnish to secure it (or its mounting screws.

-- DavidHale - 18 Jul 2008