dirkn

Sorry, I meant: http://www.nexsxd.com/ Not been updated since 2010 so no idea if still in business.

Yeah, I realise I went down a bit of a rabbit hole on this one ;-) Building my own motor controller is certainly an option- the basic design is quite simple, (setting up the timing just means knowing the mount's gear ratios and the encoder rates) but adding features such as PEC isn't straightforward. I've built a simple Arduino based focuser control for under £10 so the cost wouldn't be exorbitant either. Making it NexStar+ compatible would be an added bonus- although the NexStar aux command protocol is partly documented, most of the current feature commands aren't. It would be much easier with a working motor board to compare it to. It's been done before- there's NexSD.com in Spain who've built NexStar-compatible controllers for Vixen mounts. With PEC. It'll be interesting to see what Team Celestron can suggest, if I can avoid annoying them with my antics. The important takeaway for anybody running an Advanced VX and who doesn't love the smell of solder in the morning: You really do need to stick to an exact, regulated 12V supply when running from anything other than the Celestron 5A PSU, or a power tank you've actually measured the voltage on. I would even suggest using a 12v voltage stabilising regulator with a power tank, as an added precaution. Even then, taking precautions to avoid stressing the motors and being careful about your assembly, startup and shutdown procedures is going to potentially save you a lot of time and money: 1. Cable everything up before powering on, and never disconnect anything until you've powered down. Double-check your connections. Plug any aux connectors you aren't using. 2. Make sure your mount is balanced properly and you haven't introduced any mechanical obstructions during assembly of a complex rig. Swing your mount unpowered and with clutches disengaged to check. 3. Never shut down the power when the mount is slewing, except in an emergency to avoid injury or when whatever is about to get damaged is more valuable than a motor board. 4. Never move the mount by hand when it is powered unless you are certain the clutches are disengaged. 5. Don't toggle the power switch repeatedly. 6. Make sure all your connections are tight and free from intermittent contact. Make sure the barrel power connector socket is not loose and the split pin has is a little spread, so it makes firm contact. 7. Make sure you use a correctly sized barrel power connector, if you are not using the standard Celestron cable. 8. Turn off tracking on the hand control before powering down (a little paranoid, admittedly).

You know, I never thought of that. I'd heard of Team Celestron, but I always assumed that you had to be invited to join. I'll sign up and see what happens. In any case, the more I think about what I did to my poor Advanced VX, the more I'm convinced it might not have entirely been finger trouble. Somebody over on astronomyforum.net identified the 3.2v voltage regulator as being a Richtek rt9261-32cx. That can't actually be right, as that part is a boost converter for battery powered devices (to run 3.3v logic off of a single AA, for instance). However, an rt9161-32cx is a 3.2v linear LDO (low-dropout) regulator. It's only rated to 300/500mA. Which is pretty low, but the MK20 MCU should only draw 180mA running full tilt. Add 10 mA for the LED, 20mA for a serial TTL output and another 20 or so for each motor drive channel (very much worst case, depending on PWM frequency and input capacitance), and you're running near capacity. But that's not the major concern. This is: In contrast to the Lm1117 and clones, this part has an absolute maximum rating for input voltage of 14v. In fact, it has a _specified_ maximum input voltage of 12v. This means it's designed for operating continuously on 12v input but won't actually blow up until you go over 14v. That contrasts to a specified input of 15v for the LM1117 with a maximum voltage of 20v. As any fule kno, running something thermally limited like a regulator IC at absolute maximum rating is asking for trouble. It will work fine on a 12v regulated PSU with enough grunt (the 5A Celestron mains adaptor) . It will _probably_ work OK indefinitely on a 6-cell lead-acid battery. Especially if that battery is a little old and tired. Or is a flooded-cell rather than a gel/agm design (lower final charge voltage). When it's fully charged, the voltage will be 12.8 to 12.9 volts and just above the specified rating of the part. Not yet in the danger zone, admittedly, but most electronic engineers would be really really uncomfortable with the margin. Especially as anybody who works with Lead-acid batteries knows that a charger/PSU for 6-cell lead acid battery replacement is set up to output 13.8v for float charge, or around 14.2v for quick charging. So operating your mount from a running motor vehicle (or starting a it with the mount powered and plugged in) puts it at immediate risk of being damaged. A 4s LiFEPO4 pack (at 14v fully charged and 13.4v on load) is so close to the regulator self-destructing it's scary. Oops. Even worse, the 12V rail in the AVX is directly coupled to the motors. There is no isolation between the 12v feeding the regulator and the 12v feeding the motor controller. Now the L6226pd motor driver is rated to 52v and 60v maximum, and has internal flyback diodes, which should protect the bridge circuit from any nasty inductive spikes. However, based on the internal circuit block schematic, it is going to dump a positive voltage spike right back on to the 12v rail. Now, the chances are that a PSU or battery will absorb such a spike without noticing. However, a 2v inductive spike isn't exactly _big_. A 12v commutated motor produces some quite unpleasantly-sized sparks if it's run without suppression, and the RF interference suppression caps don't normally reduce that to below semiconductor-frying levels. Hence the need for flyback diodes. So I can see a very unpleasant pattern emerging: 1. AVX is slewing and motors are stopped by mechanical interference (OTA or accessories collide with mount/user/etc) 3. Or a naive user manually slews the mount without releasing the clutches and spins the motors in reverse. 4. Or the slew looks uncommanded/uncontrolled (unexpected meridian flip, for instance) and the naive user just turns the mount off before it does anything else scary. 5.Or the user just turns the mount off _while it is tracking_ (ruh-roh). 6. Inductive voltage spike drives 12v rail to above 14v. 7. Regulator latches up and dumps 12v across MCU. Doesn't happen every time, but the more it's abused, the more likely this gets. 7. MCU goes bye-bye. 8. If very lucky, everything recovers if left off for a while. The fix is very simple: 1. Replace regulator with something that has a higher maximum input rating. LM1117 is a good starting point. Part cost is around 50p. 2. Stick a forward-biased high-current fast recovery diode, or Schottky diode, and optionally a filter capacitor with a bleed resistor, in series with the 12v rail, after the 3.3v regulator and before the motor control chip. You lose a littles voltage for the motors (a very little, of using a Schottky), but there's no chance of an inductive spike nailing your regulator. Part cost maybe around £2 DIY, muc less if implemented on a new version of the board.

Hi, New poster here, looking for little help on a slightly crazy project. About a week ago, I was trying to set up an extension cable for my NexStar+ hand control on my new (ish) AVX mount. The short curly control cable was annoying me a little. I was also testing a new portable power supply that consists of two 4S LiFePO4 8.2Ah batteries in parallel. The batteries are HobbyKing Zippy FlightMax ones commonly used for large electric RC models. Everything was going to be connected with Anderson powerpole connectors, which, like the batteries, I've used for portable amateur radio stuff in the past. The output voltage at full charge is around 14V. The whole set-up including distribution block, wattmeter and fused inputs weighs in at less than half the weight of a traditional lead 17ah gel-cell powertank and has the added advantage of not needing regular maintenance charging when the weather's not amenable for astronomy (which seems to be most of the time around here- the new scope curse strikes again). Once I had the extension working OK, I wanted to set up a direct aux port serial interface using the interface converter designs you can find on www.nexstasite.com, but modified to use the commonly-available FT USB/Serial converter boards you can purchase from your favourite auction site or supplier of robotics/Arduino doodads. The idea was to be able to use NexRemote without a hand controller plugged in at all, and to eventually build my own SkyLink/SkyFi clone thingy out of a spare Raspberry Pi I had lying around. This was going to require a little investigation as I understand the old Celestron Aux Port Accessory for AS-GT/CG5 mounts doesn't work on AVX so the circuits for converting aux port signals to RS232 derived from it probably won't either. A little snooping of the TTL-level serial interface between the NexStar+ and the mount was going to be needed. Unfortunately, _something_ went wrong while I was setting this up. I either had the extension cable connected through the wrong RJ11 coupler (I have one which flips the connections left-to-right, which should definitely not be used here), or I managed to short something whilst trying to probe the signals on the cable. It also possible that the 14v output from the battery pack was just too high, and that the repeated power cycling of the mount while I was setting this all up upset things. In any case, I wound up with the dreaded 'No Response 16' and 'No Response 17' errors and the power light on the motor board went off. The hand control still powered up and seemed entirely happy with life, except that it was totally unable to communicate with the mount. Some further investigation showed the motor board to be very unhappy: 1. The 3.3v regulator was shutting down due to overcurrent and getting too hot to touch. 2. Replacing it with an external 3.3v supply board resulted in the microcontroller on the motor board drawing about 1.5A at 3.3v and getting very warm. 3. The motor driver IC seemed completely unaffected by any of this. So my only choice was either sending the mount back for repair or getting a spare nxw445 controller board. I was reluctant to send the mount for repair due to the expensive shipping, long turnaround times, not to mention that my overenthusiastic diagnostics were going to make it impossible for this to classify as a warranty repair. So I decided I might as well just grit my teeth and buy a new motor control board, and chalk up the(considerable) expense to a lesson on not messing with things you don't have enough experience with. I'm trying to get a quote for one now, but it's not entirely clear if they're actually available any more. I believe Celestron has designed a replacement board for the nxw445 because they've had rather a lot of DOAs and failures caused either by finger trouble (plugging the dec motor cable into an aux or handset port in the dark seems to be a popular mistake) or the motor board being oversensitive to input voltage levels. I've read forum posts that suggest Celestron specify a power source of _exacly_ 12v, not the usual 13.8v for lead-acid batteries. So while I'm waiting I thought I'd have a go at fixing the dead board myself. After all, it's already dead, and if I can resurrect it then I will have a spare, and a board I won't have to feel guilty about abusing for further experiments. I have the relevant SMD rework tools, so actually replacing the defective microcontroller is possible, just not terribly easy (LQFP-80 has quite a small lead pitch). I've managed to obtain the relevant spare parts (Freescale MK20DX256ZVLK10 ARM MCU, L6225pd full bridge driver, and an uprated 3.3v regulator) from the usual electronics supply suspects. However, I'm now a little stuck because I've overlooked an obvious issue: The new MCU won't have any firmware on it. Now, Celestron provides CFM to update the firmware via the hand control. Why can't I simply replace the MCU and then use that to load the current firmware? Unfortunately, as I understand it, for CFM to load firmware on to the MCU, it must have a bootloader installed already. The firmware upload happens via the hand control, and that's connected to the MCU via its UART 0 (serial port) input pins. Unless there is already some firmware loaded that listens on UART 0 and knows how to speak the CFM firmware update protocol, the hand control isn't going to 'see' anything connected to it. To load firmware into a blank K20 MCU (or any ARM MCU), you either need: 1. An expensive commercial firmware programmer (£100s or £1000s) for mass production. 2. A cheap in-circuit programming/debugging tool, commonly available from eBay and robotics hobbyist suppliers, that connects to the JTAG/SWD pins on the MCU. I have one of those, from some Teensy-related stuff I was working on a while back. Interestingly, the nxw445 board already has the SWD/JTAG connectors broken out on pads for a standard ARM ICP debugger connection. I just needed to solder on a 10-pin header to make that work. The remaining problem is simply: I don't have the firmware. And, although I've asked, I doubt Celestron would supply it, or even be able to. Now, I have the .cfm file for the Celestron AVX bootloader. CFM downloaded it for me the last time I ran a handset upgrade. I've even managed to reverse-engineer the file format and extract a raw binary image of the bootloader. The problem is that the reported addresses for the firmware don't really make sense based on what I know about the K20 memory map. I think there's still something missing. What I'd really like is a working K20 MCU on another nxw445 board to compare. And I'd really like to avoid doing anything to the spare I hope to obtain soon. After all, I'd actually like to get some viewing in in between all this electronic messing about. Now, given that I blew the last board up, I doubt anybody will lend me the board out of a working mount. However, it's possible that somebody either has either a dead or malfunctioning motor board from a mount on which they've replaced the board themselves, but where the MCU itself isn't actually defective. Or possibly a spare, working board they extracted from a damaged mount or kept after a repair. In any case, I'd be willing to trade the replacement board I am ordering for a working one, or be willing to repair a defective board for free, provided I can extract the firmware from it. To extract the firmware, I'd need to solder a header to the board. This should have absolutely no impact on the performance or behaviour of the board, but is obviously going to void any warranty on it. So, yes, basically I want a spare AVX motor control board for the sole purpose of sucking its brains out. The end goal, is of course, of working out a way to repair these boards without the cost and hassle of obtaining a replacement board. The individual components range from pennies to about £10 in cost, so this is viable on a small scale, for somebody unafraid of SMT soldering. I also have some ideas for increasing the robustness of the circuit and making the whole thing less sensitive to overvoltage, but that obviously requires me to have a spare board to abuse for testing. Suggestions? Should I just give up? Should you all run away screaming there's a madman with a soldering iron on the loose?