FoggyJames Posted January 2, 2016 Share Posted January 2, 2016 Hi all! Long-time lurker...and I'm surprised to find this is actually my first post! A friend has four Martin Roboscan 812s. Try not to hold it against him! When he bought them, all four worked. When I last borrowed them, we were down to one! I decided to get to work...Two worked normally, aside from failing to strike their lamps. That was down to dead thermal trips. Easily sorted. The fourth unit has a control board fault. Swap the board with another unit, and the fault follows the board. I pulled the microcontroller and swapped it with one from another board, to no effect.The fault condition is that with no address dip-switches selected, the gobo and colour wheels advance fully in one direction (counter-clockwise, I think) and then click continuously against their end-stops. Enable any of the dip-switches (either for a stand-alone mode, or a DMX address), and they do the same, but in the opposite direction. I've got as far as measuring some of the power supply components. There's a DF1502S rectifier which is seeing 26V AC at its input and giving 34.7V DC at its output. There's a second, larger, rectifier which I've not got as far as tracing easy measurement points for yet. There's a L7805CV regulator which is outputting a solid 5V. All seems sensible, on the face of it. Beyond that, the motor drive circuit appears to consist of the MC, which drives a bunch of 74HC374D flip-flops, which in turn drive ULN2803A transistor arrays. Having said that, digital electronics is not my forte, so I might have spoken a load of rubbish! There's a place selling known-good boards on eBay, but they're going for about twice what a fixture will typically sell for. I could probably fix the board given a service manual, but I guess that sort of documentation is not going to be easy to find. I guess the alternative is to take a punt on a spares & repairs unit, and hope the control board is good. Does this ring any bells for anyone? Any assistance would be much appreciated. cheers James Link to comment Share on other sites More sharing options...
Jivemaster Posted January 2, 2016 Share Posted January 2, 2016 My take on unit 4. The steppers and drivers must work because they get to their end stops, but they don't pass that in their start up mode. What, on units 1, 2 % 3 makes the chip know that the wheels have hit their stops, that is different on unit 4. I'm thinking of reed switch, sensors, magnets etc initially, or their wires and connections onto the board. Link to comment Share on other sites More sharing options...
FoggyJames Posted January 3, 2016 Author Share Posted January 3, 2016 Thanks for your reply. The issue is definitely with the board - swap the board to another unit, and the fault follows the board. There's no "detection" on the wheels. As I understand it, the board monitors current draw and senses a stalled motor (although I might have the detail of how it works wrong). It could be that that sensing mechanism (whatever it is!) has failed. cheers James Link to comment Share on other sites More sharing options...
Don Allen Posted January 3, 2016 Share Posted January 3, 2016 What is the condition of the faulty circuit board ? corrosion ? dry joints ? there is no feedback which is why the motors drive to their end stops at power on. Link to comment Share on other sites More sharing options...
FoggyJames Posted January 3, 2016 Author Share Posted January 3, 2016 Hi Don, The board looks minty-fresh! Do they just advance to the end (and a few clicks "beyond", against the stop, to be safe) and count back from there for indexing? cheers James Link to comment Share on other sites More sharing options...
DrV Posted January 3, 2016 Share Posted January 3, 2016 Hi Don, The board looks minty-fresh! Do they just advance to the end (and a few clicks "beyond", against the stop, to be safe) and count back from there for indexing? cheers James Yes, that is exactly what happens. There is no feedback, neither position sensor nor current sensing, into the MPU. By the way, there's none on the mirror motors either. All of the wheel steppers are supplied by a nominal 12V, unregulated supply which is then used to supply the 7805. It might be worth checking the input to the 7805, which is a convenient place to identify the 12V rail, for excessive ripple. It also appears on pin 10 of both of the ULN2803s if that is easier to get at. Do the mirror motors move at all? They use the other supply rail (nominally 24V). If not then the only common components are the MPU, which you've already swapped, the crystal and its capacitors, and the 74HCT238. This selects which '374 is being accessed. You said that with it in DMX mode (switch 10 OFF) it winds to the opposite end stop and then keeps going. Does it actually do this or does it just go back to position 0 (i.e. not actually against the end stop). I don't know for certain but I would expect it to move to position 0 if there were no DMX input telling it to do otherwise. Dave Link to comment Share on other sites More sharing options...
KevinE Posted January 3, 2016 Share Posted January 3, 2016 I've had problems on this age of martin boards where a bus tranceiver/buffer (or its chip select line) has stuck. Next stop is the diagram. Link to comment Share on other sites More sharing options...
FoggyJames Posted January 4, 2016 Author Share Posted January 4, 2016 Hi guys, Thanks again for your replies. I have now got my hands on a diagram! I'll have to crack out the fixtures later to do some more measuring, but I can tell you two things off the top of my head: 1) The wheels advance to the "other" stop with jump 10 "on", and keep going - so the behaviour is the same, but the direction of travel is reversed.2) I can't remember what the mirror behaviour is, as I've been testing with the mirror and motor module removed in order to access the control board. I *think* it does nothing, but I'll check later. I'm hoping to have a play with them tonight, but (if not) I'll pick it up at the weekend. cheers James Link to comment Share on other sites More sharing options...
KevinE Posted January 5, 2016 Share Posted January 5, 2016 this definitely sounds like a chip select problem or stuck data line. I can't see any other reason why a dip switch should alter the direction of a motor during its reset phase. Can you post the diagram of the main board here? Are you happy swapping SMT chips? Link to comment Share on other sites More sharing options...
FoggyJames Posted January 5, 2016 Author Share Posted January 5, 2016 Hi Kevin, Thanks for your reply. The day job has got the better of me for a few days, and the next week looks pretty hectic, but I'll pick this up as soon as I can. One thing I have quickly confirmed is that the mirror motors don't move in the same manner. The tilt motor appears to twitch very slightly (barely perceptibly) in time with the motion of the wheels, but no pan motion that I can detect. What's the ettiquette around posting diagrams on this board? This was sent to me privately, and I'm not completely comfortable with sharing it publicly without checking what the policy is first. Very happy to send it to you privately, of course. I've never done SMT work myself, but I have access to hot air stations, and can probably bribe a colleague to do it on my behalf it nothing else. cheers James Link to comment Share on other sites More sharing options...
KevinE Posted January 5, 2016 Share Posted January 5, 2016 mm I dont know. Dont upset anyone! Link to comment Share on other sites More sharing options...
FoggyJames Posted January 6, 2016 Author Share Posted January 6, 2016 I'll email it to you! cheers James Link to comment Share on other sites More sharing options...
KevinE Posted January 6, 2016 Share Posted January 6, 2016 Hi James The dip switches go straight into the micro so we'll assume they're out of the picture for the moment.The chip select decoding is done by IC03, if this fails the motors will do all sorts of daft things. If you can scope the data lines from the micro and make sure they have sensible activity on all of them, if that looks ok then I'd change IC03. If any address lines are stuck or look weird, one might be being loaded by one of the buffers, which one is anybody's guess and the best bet there would be a block change. Do make sure you've done nothing by mistake such as bent one of the micro pins as you've swapped it..its easy to do! Another thing I remembered from a while back (on a roboscan 918) was an electrolytic which had leaked very slightly underneath and rotted through a data line that just happened to pass underneath it out of sight. The data line was left floating which caused mayhem. And as always be vigilant to the effects of corrosion in general on a compact board like this and around the connectors. Link to comment Share on other sites More sharing options...
FoggyJames Posted January 31, 2016 Author Share Posted January 31, 2016 Hi all, Thanks again for your assistance with this. I've finally had an afternoon at home, and where the weather is grim enough to write-off the idea of doing anything outside, so I cracked out the scope and 812s a bit earlier! I have both a working example and our problem child on the bench. To answer earlier questions...the 12v rail is sitting at a nominal 12v on both units, and both have around 1v of ripple.The three lines to the 74HC238 don't seem to do a lot, but there is similar activity on all three lines on both units. I had a poke around inside them, and concluded that the only appreciable differences between them I could spot were... 1) On the clock input to the flip-flops, the bad unit seems to have a lot less "diversity" of activity. What was going on did change, but not as much as on the good unit. It was marginal, though. Bad unit looked more or less like this all the time:http://i612.photobucket.com/albums/tt201/foggyjames/812_bad_zpsfepmufyy.jpg Good unit varied far more, and displayed some states like this:http://i612.photobucket.com/albums/tt201/foggyjames/812_good_zpsrxjge0mc.jpg 2) Potentially far more significantly, though, the address lines out of the micro on the bad unit don't look clever. Bad unit (3.5v?):http://i612.photobucket.com/albums/tt201/foggyjames/812_bad_add_zpsbfphnnev.jpg Whereas the good unit (5v?):http://i612.photobucket.com/albums/tt201/foggyjames/812_good_add_zpsdazfjnzo.jpg Does anything jump off the page at you there, team? Would that be indicative of the bad buffer (flip-flop, right?) you mentioned, Kevin? Thanks once again for all your assistance! cheers James Link to comment Share on other sites More sharing options...
KevinE Posted January 31, 2016 Share Posted January 31, 2016 its pretty fruitless trying to analyse individual data lines other than to check there's valid data on them, ie not stuck one way or the other or half way between states. A faulty buffer or select line for example will often just cause data to go to the wrong places. Any suspect chips, just change them, they'll be cheap Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.