69 replies to this topic

[Ken Sturrock]

Hmm...so you have your Arduino on the same USB as everything else on the scope? I wonder if that may be problematic as well. Is there any way to completely and totally isolate the dew heater from everything else, both data and power?

 

I do wonder if noise being feed back from the dew heater may be contaminating the rest of the system. The KAF-16803 may just be more sensitive to this than the KAF-16200. 

 

That was my Black Swan idea (above) but, in this most recent case, John said that the Arduino was offline and disconnected. I hate it when data wrecks a cool theory...

[jhayes_tucson]

There are a lot of interesting thoughts here but I think that some of the background information has gotten completely lost.

 

1)  My system has been running with this same topology since early in 2016 when I got the camera.  I have upgraded a few components from office level stuff to wide-temperature, commercial grade stuff but none of those upgrades revealed any problem.  The system ran for two months without a single glitch with a brand new FLI-ML16200--using the same wires, supplies, hubs and software.  These two cameras are virtually identical electrically--requiring the same power and communicating through USB 2.

 

2)  I've demonstrated that the anti-dew system has nothing to do with the problem by completely disabling it and the camera still disconnects.

 

3)  I've demonstrated that power to the hub on the lower mount is not likely to be the cause of any problems by powering it down and running it passively.  Running the camera with FLIgrab allows the camera to start and work...until it disconnects just like it does when the hub is powered.

 

4)  The Arduino has nothing to do with the problem.  The original symptoms started long before I implemented the Arduino control system.

 

5)  The USB 2 hub on the OTA is unpowered.  The power requirements at that point are very low so external power is not needed.

 

6)  When I'm at the observatory, I will connect the camera directly to the computer via a short USB cord and power it using a new FLI power brick.  If it fails, it goes straight back to the factory.  Otherwise, I'll work my way through the system to find the problem in either a wire, hub, or power supply.  The ML16803 may be right agains a tolerance for voltage or cable length so it would be nice to identify that problem.  Here is how I plan to reconfigure the main control box while I'm out there.  This will take the lower hub out of the main data loop and separate dew control from everything else.  The dew system will also be powered by a linear supply for additional stability.  If you spot a problem with this configuration, let me know and I'll try to fix it.

 

John

Attached Thumbnails

• 

Edited by jhayes_tucson, 24 September 2018 - 05:26 PM.

[pfile]

maybe not related to the problem, but i'd think about powering that top hub. it has to draw power from the hub it's connected to, and there could be a lot of inductance and/or IR drop leading to low-ish voltage at the top hub.

 

rob

[Jon Rista]

There are a lot of interesting thoughts here but I think that some of the background information has gotten completely lost.

 

1)  My system has been running with this same topology since early in 2016 when I got the camera.  I have upgraded a few components from office level stuff to wide-temperature, commercial grade stuff but none of those upgrades revealed any problem.  The system ran for two months without a single glitch with a brand new FLI-ML16200--using the same wires, supplies, hubs and software.  These two cameras are virtually identical electrically--requiring the same power and communicating through USB 2.

 

 

6)  When I'm at the observatory, I will connect the camera directly to the computer via a short USB cord and power it using a new FLI power brick.  If it fails, it goes straight back to the factory.  Otherwise, I'll work my way through the system to find the problem in either a wire, hub, or power supply.  The ML16803 may be right agains a tolerance for voltage or cable length so it would be nice to identify that problem.  Here is how I plan to reconfigure the main control box while I'm out there.  This will take the lower hub out of the main data loop and separate dew control from everything else.  The dew system will also be powered by a linear supply for additional stability.  If you spot a problem with this configuration, let me know and I'll try to fix it.

 

Just curious. The "12V S FLI Cam" power source. What exactly is that? Have you checked the voltage on it? Any chance you have checked the signal coming from it? I wonder if the KAF-16200 camera has better power regulation than the KAF-16803? Just trying to think of reasons why the latter may be having problems while the former is not (currently) having any... Poor quality or non-sine signal/noisy signal might be affecting one camera more than another.

 

Anyway...I still think checking everything with an oscilloscope might help you track down issues. 

[tolgagumus]

 

 

  When the camera disconnects, it takes down my FocusBoss controller causing it to disconnect.  

John,

 

How do you conclude that the camera is taking out the focus controller and it's not the other way around or it's the hub that they are both plugged into that taking both down. 

 

I'll be at DSW on Friday. If you want I will bring and install a straight USB cable to the computer form the camera to eliminate these questions. 

[tolgagumus]

Here are the data lines.

 

John

Looking at this diagram. It looks like you have a hub plugged into the hub built in to the ASI camera. this is a no no.

[jhayes_tucson]

maybe not related to the problem, but i'd think about powering that top hub. it has to draw power from the hub it's connected to, and there could be a lot of inductance and/or IR drop leading to low-ish voltage at the top hub.

 

rob

Maybe, but there is nothing that doesn't have it's own power connected to it and it's run fine for years.  The wire running from the 13.8 VDC supply is ~14 gauge and it's short (< 10 feet) so the voltage drop will be about 0.5 V at 10 amps (way over kill).  That's 13.3 VDC on top of the OTA, which won't be a problem.

 

John

[jhayes_tucson]

Looking at this diagram. It looks like you have a hub plugged into the hub built in to the ASI camera. this is a no no.

 

Why?  It's worked fine for a log time and my system recognizes everything without any problem.

 

John

[jhayes_tucson]

John,

 

How do you conclude that the camera is taking out the focus controller and it's not the other way around or it's the hub that they are both plugged into that taking both down. 

 

I'll be at DSW on Friday. If you want I will bring and install a straight USB cable to the computer form the camera to eliminate these questions. 

 

Easy.  If I disconnect the focuser so that it's not doing anything, the camera still crashes.  I've seen the camera disconnect when there is no power being applied to the OTA at all--just like it does when everything is powered.  Conversely, if I leave the camera disconnected, the focuser connects and runs forever without any problem.

 

John

[tolgagumus]

Why? It's worked fine for a log time and my system recognizes everything without any problem.


John

In my experience piggybacking hubs is asking for trouble. I see this all the time when I'm troubleshooting.

[jhayes_tucson]

Just curious. The "12V S FLI Cam" power source. What exactly is that? Have you checked the voltage on it? Any chance you have checked the signal coming from it? I wonder if the KAF-16200 camera has better power regulation than the KAF-16803? Just trying to think of reasons why the latter may be having problems while the former is not (currently) having any... Poor quality or non-sine signal/noisy signal might be affecting one camera more than another.

 

Anyway...I still think checking everything with an oscilloscope might help you track down issues. 

 

That's the FLI power supply that came with the camera.  I have checked the voltage but not since this problem started.  When I visit next week, I'm taking another FLI supply with me for testing and to swap out.  I don't have a scope and I won't have one before I head out there the next time.  I also don't have a breakout box to hook it up.  If I get to that point, it will have to be a separate project.

 

 

 

In my experience piggybacking hubs is asking for trouble. I see this all the time when I'm troubleshooting.

 

Hubs are what allow a single data cable to be run through the mount; otherwise, you've got to run a fist full of wires everywhere.  Regardless, if the camera works by being connected directly to the computer, I'll run yet another cable for it.  I'm up to four cables so what's another cable??

 

John

[pfile]

Maybe, but there is nothing that doesn't have it's own power connected to it and it's run fine for years.  The wire running from the 13.8 VDC supply is ~14 gauge and it's short (< 10 feet) so the voltage drop will be about 0.5 V at 10 amps (way over kill).  That's 13.3 VDC on top of the OTA, which won't be a problem.

 

John

 

what i meant was the hub itself needs power, and it gets it from the "B" port, where it is connected either to another hub or a computer. so the hub is being powered by pier-side hub thru a longish USB cable.

 

rob

[choward94002]

Hmm ... unfortunately diagnosing hardware issues reminds me of diagnosing bad subs ... there are just so MANY things that can go wrong!

 

Luckily you've got three things in your favor: the problem manifests pretty quickly (versus a true Heisenbug that never shows up when you're trying to catch it), you've got someone who's able to spend some time fiddling with things (versus a plane trip) and swapping things out doesn't entail five figure bits of kit ...

 

In your shoes, I would first install the USBDevTree utility on your remote computer and see just what your topology is and who's talking to what ... you can do that remotely, and it will also do some simple diagnosis of the equipment ... that might be enough to get some info there.  Make sure you identify at least one USB 3 port on your computer ... you can install it on your laptop at home first to build confidence in it ...

 

Prior to going onsite I would bring some additional kit: several different brands of USB 3 hubs (4 port, one Belkin, one Anker and one StarTech) and several brands of USB 2 hubs (same mix).  I would also bring down some *good* USB 2 and USB 3 cables (you want the shielded stuff, ideally with ferrite beads but shielded is good ... you want the big, thick, beefy stuff with the stranded metal jacket).  Make those cables just a short as possible that will still work with your rig topology.  I would also get some good power supplies for each thing that is going to draw power: the camera, the hubs, etc (the camera power supply should be good, FLI isn't going to give you a $5 toy with an $11K camera but the other power supplies might be dodgy.  You want the power supplies that are big, heavy and expensive [https://www.amazon.c...ly switching 5v], one for your USB 3 hub and one for each other thing in your mix ... I would also bring a USB 3 optical gaming mouse [https://www.amazon.c...rds=usb 3 mouse] ... luckily all of that can be bought and shipped for a few hundred ... 

 

Onsite, first unplug everything from the computer and identify the USB 3 port.  Plug the USB mouse into the port to identify it and make sure you see the mouse as USB 3 in the utility, then plug the camera into that port with the camera power supply with one of the big, beefy USB3 cables and see if it's happy (take a bunch of flats, download, etc.), and watch to see if USBDevTree makes any noises.  If that works, then plug the USB 3 hub into the port, again plug the USB mouse into that port on the hub to make sure it's working and then the camera into hub (again using the shielded cables) and once again see if it's happy.  You may have to move plugs around different ports, the internal topology is rarely what it looks like from the case, but basically you're going to be redoing your topology using the beefy power supplies, beefy cable and different hubs while watching the utility for funny stuff and hopefully at the end of the mind-numbing permutational trials you'll have enough info to bin the camera, bin the hub(s), bin the cable(s), bin something so you can then hook everything back up and you're good.  There will be one more thing to watch out for, when you're doing your cable routing make sure that data cables don't run parallel to power cables (you'll potentially get "inductive coupling", usually not found with DC current but 'ya never know, can't hurt!), if they do you can shield them with good 'ole aluminum foil wrapped around the power cable where it's next to the data cable, or give them about 3" of distance ...

 

Good luck!  It will probably be as much fun for you as getting my HyperStar field to be flat is for me!   

Edited by choward94002, 24 September 2018 - 08:21 PM.

[rockstarbill]

If the problem will reproduce with the camera directly connected to the PC running it, its the camera. That is the test that needs to be done here. This is a fundamental problem, just use some fundamental tactics to determine the right course. 

[jhayes_tucson]

If the problem will reproduce with the camera directly connected to the PC running it, its the camera. That is the test that needs to be done here. This is a fundamental problem, just use some fundamental tactics to determine the right course. 

Agreed and that's the first test that I'm going to do next week when I'm there.

 

John

[rockstarbill]

Agreed and that's the first test that I'm going to do next week when I'm there.

 

John

Good. The whole talk about putting scopes on things, while cool from a science perspective felt like it was getting into the realm of analysis paralysis. Always go back to the basics and start there. 

[Mert]

Hope you get this item fixed as well John, you are really 

suffering a lot of backdraws since a long time now.

If I were to go the same route I'd go for seperate supplies

for each chain with a common ground point between all of them.

 

 

And thinking again on this, how is your ground (in Ohms )??

Over here in Spain, we are at about 500m above sea-level and it's

normally quit dry here, so the ground is also quit high!! ( sometimes

more then 2 - 3 Ohm )

When this happens, weird things are easily possible.

Your installation is at much more height so, what is the ground

resistance you have over there?? 

 

Hope you get it solved!!!!

Regards,

Mert

Edited by Mert, 25 September 2018 - 01:00 PM.

[jhayes_tucson]

I want to finally close out this thread with what I did and how it's worked--so far.  As I indicated earlier, I rewired my system to do two things:

 

1)  The FLI-ML16803 is now on it's own single 15', USB 2.0 cable that runs directly from the computer to the camera.  It's the same cable that FLI shipped with the camera and it has a magnetic choke at the camera end.

 

2)  I measured the voltage from the FLI power supply that runs from my electronics enclosure to the camera and it's 12.38 VDC (at the camera.)  So, that wasn't a problem.

 

3)  I swapped supplies around so that my anti-dew system now has it's own isolated 12.5V, 12A switching supply.

 

In all, I had to run two more cables through the mount.  The wiring diagram is shown below.  Before making this change, my system got to the point where I couldn't keep the FLI camera on line for more than about 15 minutes.  So far, the system has now run for two full nights without a hitch.  I want to thank Tolga for some useful discussions along the way.  I had what I call a "soft" commitment to do this change, but Tolga read me the riot act, telling me that he supports a lot of systems and wouldn't even consider running data from an astro-camera through a hub--of any kind and he was right.  We couldn't figure out why a hub should introduce this kind of problem but Tolga said that he's seen this kind of problem over and over.  So, it sure looks like the hubs were the problem.  Thanks Tolga...I really appreciate the help!

 

John

 

 

 

 

Attached Thumbnails

• 

[tolgagumus]

Thank you John,

You know how they say do what I say not what I do? Last week I installed a different FLI camera on my rig and it was disconnecting the Nightcrawler. I said the same thing. It was working fine with the MLx694. So what's the problem now? However since this was a test camera, I first thought maybe it was the camera. It was the same thing. I bypassed the hub, plugged the camera directly and boom. No problems. Luckily I discovered the issue before I left and rewired it right there.

Sent from my SM-G892A using Tapatalk

[pfile]

one possibly nitpicky question - is the 12.38V under load or open-circuit?

 

i guess i'm not surprised to hear that the hub was a problem. i do wonder if some of it is to do with the mix of usb 3.0 and usb 2.0 with respect to the camera. high-speed serial connections are somewhat tricky and to me it's a miracle that they work at all

 

in my case i have no choice but to run everything thru a hub since i'm using one of those tiny stick PCs. it's been pretty stable but maybe 1% of the time my STT is not discovered by the PC at boot time. i probably would have a better PC/USB setup if the system were remote though.

 

rob