StackShot help!

SONYNUT · Post by **SONYNUT** » Tue Nov 22, 2011 2:24 pm

i'll stick to manual

.....takes the fun out of it anyway...

ChrisR · Post by **ChrisR** » Tue Nov 22, 2011 4:07 pm

I'd be (annoyed but) not too distressed. Connectors are usually the weak points. The motors don't cost a fortune, nor the drivers. If you've lost a "pole" on the motor it would probably do something like get hot and jiggle. It may well have taken a component on the driver board with it.
I doubt Cognisys would want it to hurt you ( or any of "us") too much.

I would be using a something like a relay or optocoupler to operate the camera shuter, so there would be zero chance of electrical damage to it, short of both of them being struck by lighting.

When I was more active with bits of electronics I only ever used XLR type connectors:
http://en.wikipedia.org/wiki/XLR_connector
A tad chunky perhaps but just about lighting proof!

Hope it's sorted soon.

Craig Gerard · Post by **Craig Gerard** » Tue Nov 22, 2011 5:55 pm

I've boxed up the items and returned them to Cognisys.

The parcel was sent via Express Post International (EMS), so should arrive in 4-5 days, Cognisys will then be in a position to return a determination/analysis.

ChrisR wrote:I would be using a something like a relay or optocoupler to operate the camera shuter, so there would be zero chance of electrical damage to it, short of both of them being struck by lighting.

Good point, I'll be looking into such an arrangement. Some type of wireless device for communicating with the camera sounds good to me

Any thoughts? IR, RF, combination of a transmitter and receiver?

Cognisys have a universal IR transmitter; but the 50D does not have an IR receiver.

Craig

DQE · Post by **DQE** » Wed Nov 23, 2011 5:41 pm

As coincidence would have it, I just ordered the IR remote from Cognisys, mostly for other purposes. It seems like a useful remote and I've had bad luck with the Canon and some other remotes.

I've personally had nothing but exceptional service and good faith interactions with Cognisys - I'm sure they'll make things right.

Craig Gerard · Post by **Craig Gerard** » Wed Nov 23, 2011 7:26 pm

DQE wrote:I've personally had nothing but exceptional service and good faith interactions with Cognisys - I'm sure they'll make things right.

Phil, this has been my experience in the past. Recent correspondence with Cognisys would indicate this is still the case.

I have been assured there is sufficient isolation built into the controller to prevent any damage to any external device. During the 'meltdown' my external surge protectors saw no threat requiring a response, this would indicate details in the previous sentence are correct.

While my rail and controllers are in their hands, I have asked Cognisys to consider incorporating a more 'robust' connection for the S-Video cable or an alternative interface. That particular component of the rail/controller combo is a weak link in the chain. I won't go into particular details; but it is a documented problem. An alternative, nonproprietary solution to the current configeration should not be too difficult to incorporate.

The StackShot has made the physical aspects of automated focus stacking available to many people who would otherwise not be able to begin to consider automation in their setups. The photomacrography community would be the poorer if Cognisys were not active in it. I appreciate their products and their overall business model and fully encourage supporting their initiatives.

Craig

DQE · Post by **DQE** » Wed Nov 23, 2011 8:04 pm

I very much agree re the limitations of their S-video cables, especially since I would like to try and use the rig for some types of field macro. To do that comfortably, I will have to cobble together some sort of cable connector stabilization. At the moment, slathering on some duct tape is the best I can come up with given my modest engineering abilities.

As I've commented in another thread, I also see a need for a breakaway type of connection so that accidentally catching the cables on a bush doesn't risk pulling the rig onto the ground. Ideally, the cables would safely disconnect if a few pounds/kg of force were suddenly applied.

rjlittlefield · Post by **rjlittlefield** » Wed Nov 23, 2011 10:37 pm

Craig Gerard wrote:I have been assured there is sufficient isolation built into the controller to prevent any damage to any external device.

I concur. The manual for the StackShot controller shows a snippet of schematic about how the trigger connector is wired internally. It's just a P-channel MOSFET wired across the contacts. This makes it essentially a solid-state relay, with no opportunity to put voltage across the contacts.

As far as I can see, the most interesting opportunity for misbehavior is actually provided by the system "ground plane". By system, I mean not just the StackShot controller, but the computer and camera that it's connected to by various cables. I just now checked using my laptop and Canon T1i. With everything cabled together, the chassis of the laptop is connected to the shells of the StackShot connectors and to the hotshoe and metal fittings of the camera, including the strap brackets and the tripod mounting screw. None of this, however, is connected to any external ground. As set up right now, the whole shebang is floating about 50 volts above my house wiring ground! (When actually grounded, there's only about 65 µa of leakage current, about 50 of which seems to be provided by the StackShot power block and the remainder by the laptop's power block. This completely acceptable -- almost 10X smaller than safety standards require.)

However, if anything resembling ChrisR's lightning bolt hits any one of the components, they're all going to bounce up and down at just about the same rate. That's not necessarily such a bad thing, since it would minimize damage to any of the mutually bouncing parts. On the other hand, if I were going to use it around other potentially hot stuff, I might think about adding some real grounds just to avoid personal excitement.

--Rik

DQE · Post by **DQE** » Thu Nov 24, 2011 9:13 am

Are there concerns about creating ground loops if one adds real grounds?

ChrisLilley · Post by **ChrisLilley** » Thu Nov 24, 2011 12:32 pm

Craig Gerard wrote: While my rail and controllers are in their hands, I have asked Cognisys to consider incorporating a more 'robust' connection for the S-Video cable or an alternative interface. That particular component of the rail/controller combo is a weak link in the chain. I won't go into particular details; but it is a documented problem. An alternative, nonproprietary solution to the current configeration should not be too difficult to incorporate.

Craig - sorry to hear of your troubles. My initial response on reading your first post was that theSs-Video connector might be dislodged (later posts showed this was not the case) because of my own experience with Stackshot. Twice now, I had a situation where the motor was either not moving, or moving feebly and unable to move the platform along the rail. In both cases, it turned out that the S-Video cable was slightly dislodged from it's socket.

Craig Gerard · Post by **Craig Gerard** » Thu Nov 24, 2011 7:16 pm

Chris,

Yes, I've also experienced such intermitent connections problems. The rule-of-thumb now, for me at least; if the rail doesn't move, power off the unit immediately and check the cable, the connection and the pins.

It would appear a more robust/industrial type cable solution has not been included in the StackShot due to the fact it would increase the overall cost of the unit, the same goes for the inclusion of a simple of/off switch. The StackShot is already good value and I would have no hesitation paying more to have those features incorporated into a 'deluxe' version.

Currently awaiting a 'costing' to have the S-Video connectors modified to screw type connectors (threaded bulkhead or threaded chassis mount jack) and also to have two 1/4-20 holes drilled and tapped into the top of the rail platform to enable attachment of a double dovetail AS plate for connection to a back to back AS clamp.

*Chris, thanks for the link:
http://www.svideo.com/lockingsvideo.html

*Additional links:
http://www.thefind.com/electronics/info ... o-bulkhead

http://www.svideo.com/lo1.html (Made in USA)

Craig

rjlittlefield · Post by **rjlittlefield** » Fri Nov 25, 2011 12:35 am

DQE wrote:Are there concerns about creating ground loops if one adds real grounds?

Just the usual -- avoid two "grounds" that are actually at different potentials.

Ground loops were a huge problem back in the bad old days of non-isolated power supplies, when "ground" on a chassis often meant the device end of the neutral side of the power line. With that setup it was pretty easy to develop a volt or two difference due to resistance in the power line, which would wreak havoc with the single sided analog signals that were also common then.

These days, power supplies are usually isolated, and "ground" (if any) is provided by a separate conductor that doesn't carry power, such as the ground pin of a properly wired power outlet. Those grounds can still be at different voltages due to inductive and capacitive coupling, but the differences are more likely to be in the millivolt range.

As an amusing exercise, I just now connected a meter between the ground of my kitchen range and the ground pin of a nearby outlet (different circuit!) that has a toaster plugged into it. Turning on the toaster introduces 3 mv of difference between the two safety grounds. However the neutral side of the toaster power circuit jumps by 400 mv, even though the overly simplified schematic says it should be at ground also.

Anyway, in the case of photo setup, what we have is a "floating ground" in which the chassis of the laptop, the body of the camera, and the connector shells of the StackShot are all connected together by the shielding braids of the signal cables, but none of that is connected to the outside world except by leakage through the power blocks. It's that leakage through the power blocks that causes the connected components to float way above actual ground, as measured by a high impedance voltmeter. The wired-together components are too close to the same voltage for me to accurately measure. Meanwhile, the signals between them are several volts. I think it would be difficult to screw this up with a ground loop.

You might, however, be amused by the story of my experience whilst measuring some electrical signals inside the StackShot. I was interested in the motor drive, so I connected a scope across the motor drive connectors. No problem, good data. This was with the StackShot not connected to anything except itself. But I wanted to collect the data under a variety of conditions. It seemed like a good idea to automate the process, and the automation seemed simple. I would simply drive the StackShot from Zerene Stacker, using a camera triggered by the StackShot to snap pictures of the scope display.

Unfortunately, I failed to ask my usual question: "What could possibly go wrong?". When I ran the experiment, it failed miserably. Completely crappy data -- the motor didn't even work right. I might even have crashed the controller; don't remember for sure.

What went wrong was a form of ground loop. In order to drive the StackShot by Zerene Stacker, I had connected it to the nearest available computer, a deskside tower. But of course the tower was grounded. So was the scope. So was the StackShot, courtesy its USB connection to the computer. And so was the ground side of the scope probe I was using. In the end, as a result of all these below-the-radar ground connections, when I connected the scope to the StackShot, I was effectively short-circuiting the output of one of its motor drives. Oops! Fortunately no harm was done except to my psyche.

--Rik

SONYNUT · Post by **SONYNUT** » Fri Nov 25, 2011 1:33 am

You are probably getting emf voltage picked up by the toaster chassis...like the old waterbed heaters would do to people...sit on a water bet with a probe in one hand and the other to earth ground...you could get 50-80 volts...

do you have one of those tester screwdrivers where you put your finger on the end(you are the ground)....hold that on your toaster...

on the old waterbeds you could hold it on a person that was on the bed and light it up...

or a newer electronic pen style

SONYNUT · Post by **SONYNUT** » Fri Nov 25, 2011 1:51 am

I use mil spec plugs on the stuff I build..never have a problem

rjlittlefield · Post by **rjlittlefield** » Fri Nov 25, 2011 2:19 am

SONYNUT wrote:You are probably getting emf voltage picked up by the chassis...like the old waterbeds would do to people...sit on a water bet with a probe in one hand and the other to earth ground...you could get 50-80 volts...

That's a reasonable guess, but no. The results I'm reporting were carefully measured with properly referenced test instruments. Taking the laptop by itself... With the laptop not plugged into anything, leakage current to ground measures less than 1 µa -- the limit of my meter. Plug in its power block and the current rises to 10 µa. Leave the power block plugged into the computer but disconnect it from the power mains, and the leakage drops back to 1 µa. Likewise for the StackShot power block. Connect it to the meter but not the power mains and the meter reads 1 µa. Plug it into the mains, and it rises to 57 µa. The main source of leakage is in the power blocks. There's no magic here -- it only takes about 200 pf of capacitance to give 10 µa at 115 volts and 60 Hz.

I'm guessing your waterbed probably made one heck of a good capacitor -- especially if it had an electric heating element next to it.

--Rik

ChrisR · Post by **ChrisR** » Fri Nov 25, 2011 12:10 pm

The manual for the StackShot controller shows a snippet of schematic about how the trigger connector is wired internally. It's just a P-channel MOSFET wired across the contacts. This makes it essentially a solid-state relay, with no opportunity to put voltage across the contacts.

I don't see a download for the manual on the Stackshot site, so don't know if there's more information to be gleaned about the device which connects between the Stackshot electronics and your camera.

However, "P-channel MOSFET " says nothing much about an isolation guarantee. Commonly found fault voltages could easily break down the internal connection between its Gate ( connected to the Stackshot electronics) and the Drain/Source (connected to the camera).
The first "P-channel MOSFET " I checked, can break down with as little as 20V between Gate and Drain/Source. That's not enough protection!

Devices used for Isolation, guarantee typically several thousand volts separation. As I suggested, these might be an Optocoupler (= opto isolator) or a relay, but there are many other suitable approaches.

If Cognisys say it's Isolated then I'd tend to believe them because this is very basic stuff.
I'm sure they can tell us how, and by how much.

I feel the level of anxiety here is probably not well-founded, because the problems are so well understood and easy to deal with.
Millivolt differences between earth lines don't matter to digital circuits, which are "noise" tolerant - assuming the noise gets though in the first place.
The "400mv jump" in the toaster neutral connection is normal, there's current in the wire ( a few amps) and it has resistance (perhaps, under a tenth of an ohm), so there will be a voltage ( a few tenths of 1V) developed.