All Sky Camera/Weather Station build

[angryowl]

2 minutes ago, Gina said:

Very clever

Thanks Gina, not satisfied with how it handles the very centre of the circular image and this results in a REALLY stretched portion at the top of the resulting image. Sadly I don't see any other way of changing the algorithm to improve this as it needs to have a centre point from which it starts the conversion from. If I decide to use the images in my finished app I could crop a bit from the top until I start seeing acceptable star shapes.

[angryowl]

Just tried to import a master dark into MallinCam Universe software and it failed with the following error message:

Using a single image as a "master dark" works just fine. The master dark was stacked in PI from 30 images captured with the MallinCam software. Checked the resolution of the resulted master dark multiple times and it's exactly the same as the preview resolution in the software. This may prove to be a very inconvenient issue as I don't see how effective a single image used to subtract dark noise would be compared to a real master dark.

[angryowl]

Tested the bucket using two methods and both gave surprisingly similar results; Used a syringe and filled each bucket half carefully until it tipped over then recorded the ml of water used. This was performed several times for both sides of the bucket and an average was calculated from the values. The second method was using 100ml of water and a small funnel placed right above the centre of the bucket, then the water was slowly poured in, counting each tip of the bucket. The total amount of water (100ml) was then divided by the total number of bucket tips which gave the volume of water per bucket. The difference between these methods was only 0.05ml so I'd say that's pretty accurate. The final resulted volume per tip is 1.60ml.

Looking at some resources online I'm thinking a good size for the collector funnel would be 120mm diameter which with such a small bucket should record a light rain with ease and hopefully also keep up with heavy rain!

[Gina]

That sounds pretty good   I haven't tried making my own rain gauge, relying on the Fine Offset (as sold by Maplins) proprietary one which seemed alright but needed a bigger funnel to work with fine rain or drizzle.

[angryowl]

I can see the fact that I have to manually go into the MallinCam Universe software to change different settings/parameters and start/stop recording a real bummer as I'd like to have full automation on this project and full control from within the weather station application. Introducing Spy++, a  Visual Basic utility which allows viewing threads, processes, messages, and window messages. Essentiallt what can be done with it is inspect individual components (buttons, text boxes, etc) in applications and when clicked or parameters are passed to it you can visualise the actual instruction messages the system is sending to the application. Knowing these parameters can then allow one to write a VB app that sends an identical message to the one observed earlier i.e. pressing a button or retrieving/setting text in a text button. 

I shall definitely read up on this as if this works as I think it does then I'd have the ability to fully control the MallinCam Universe software from within the weather station app. 

[angryowl]

10 minutes ago, Gina said:

That sounds pretty good   I haven't tried making my own rain gauge, relying on the Fine Offset (as sold by Maplins) proprietary one which seemed alright but needed a bigger funnel to work with fine rain or drizzle.

Yeah, have seen that exact same funnel design in many manufacturers, perhaps even the exact same size! Doesn't look like it can cope very well with light rain, wonder why the decision to make them so small...

After all they're labelled as "professional" weather stations!

[angryowl]

On 23/03/2018 at 20:43, angryowl said:

I can see the fact that I have to manually go into the MallinCam Universe software to change different settings/parameters and start/stop recording a real bummer as I'd like to have full automation on this project and full control from within the weather station application. Introducing Spy++, a  Visual Basic utility which allows viewing threads, processes, messages, and window messages. Essentiallt what can be done with it is inspect individual components (buttons, text boxes, etc) in applications and when clicked or parameters are passed to it you can visualise the actual instruction messages the system is sending to the application. Knowing these parameters can then allow one to write a VB app that sends an identical message to the one observed earlier i.e. pressing a button or retrieving/setting text in a text button. 

I shall definitely read up on this as if this works as I think it does then I'd have the ability to fully control the MallinCam Universe software from within the weather station app. 

Well after further research this is a no go as the MallinCam Universe is written in such a way that it's not easy to retrieve component identifiers on the main window plus on every start you get a new instance reference which'd have to somehow be changed in the main app. Tried passing messages using the Windows Send Keys and other methods and got it to work on other programs just fine, I suppose It's just something I aheb to accept and definitely not going to invest more time in figuring this out, it will just have to be done manually.

On another note, both my main casing and funnel for the rain gauge finished printing but seems as though my slicer didn't do a proper job as towards the end of the tip of the funnel it effectively skipped about 5 layers resulting in a complete tangled mess. Got the full conical shape of the funnel printed fine though, it's just that final cylindrical bit which I added in the hopes of ensuring a straight flow of water rather than it going in all directions. Anyway printed a substitute which will be glued on top of the funnel as I'm not printing a new funnel. Also when I was removing support from within the main casing due to the casing's thisckness being only 1mm I managed to crack and eventually separate it near the top. Again this will be glued as I will not be reprinting it as it's used up 33 meters of my ASA filament.

Time to print the two arms which will be holding the rain gauge assemble and the clamp that attaches it to the main pole.

[angryowl]

I've also for a while now been having restart issues with the mini PC. This happens whilst using Windows RDP, VNC, or staright HDMI port to monitor. It's random and cannot for the life of me determine the cause of it.

When I was connected to HDMI I saw exactly what happened; I was doing nothing at all on the machine then all of a sudden it shut down, then went into a reboot loop where it would get to the Windows Logo screen sit there for a bout a minute or so then go back to the main boot sequence and start the process all over. This usually goes on for 5 to 8 times before it decides to fully boot into Windows.

Things that I suspect could be the cause of this are the following:

• A Power surge or short on the PC supply board which could be triggered by a number of factors
• The Arduino Mega has this AutoResetOnSerialConnection function which could in theory when I'm connecting my application could reset the Arduiono and the the Arduino would go through it's regular startup procedure and send a signal on the PC ON button which by default in Windows would either put the computer to sleep or switch it off. However the auto reset of the Arduino can be disabled both in code and by adding a resistor which I did and also changed the current power plan on the machine to do nothing on a press of the power button whilst the machine is on. This scenario would still not make sense as the reboot happens randomly not only when the app is connected but I thought it'd be worth mentioning
• All of the 5V and 12V PCB boards controlling the various components are grounded together, this includes the Arduino GND line. This then grounds to the aluminium enclosure.
• None of the Power adaptor I'm using have ground connections so just didn't ground any of them

Not sure if this is a software issue or hardware one, but one of the things things I'm contemplating is hookup the PC ground case to the aluminium enclosure?

Searched forums and Microsoft documentation and error reports on this but the fact that it simply does not produce any error at all is baffling and makes it hard in determining if it's a hardware or software issue let alone finding a solution.

Anyway before it can go up the pole this issue needs to be sorted. Will report back with any updates on this...

[angryowl]

The rain gauge is all glued and assembled minus the coating of the bucket and funnel with rain repellent liquid.

Did a quick test this morning as it's raining with the rain repellent liquid properly applied to the piece of borosilicate glass that's going to be my printer bed replacement. It works quite well, the majority of water droplets slides off the glass instantly and most of the ones that remained began to slowly also make their way down. So I would say properly applied to my rain gauge this will do a good enough job in gathering most of the water droplets. Enough of rain...Let's move onto something else, the anemometer!

Received my handheld digital anemometer in the post this morning and will be building my array of four server fans(Delta PFC0612DE) powered by a monster power supply as each runs on 12V and takes 1.68 amps! The reason I'm using four fans covering the entire span of the cups instead of just one pointed at the cups is because wind doesn't blow like that, it blows over the entire surface of the anemometer's cups, front and back, so that needs to be accounted for when taking a measurement. But you people probably already knew that so don't know why I just wrote that. Anyway....Will be posting back with measurements and as always accompanied by pictures.

[angryowl]

Right, so just realized I sold two of the server fans, so I don't think I'll work with just the two I have left. So I found a fan blade from an old fan I had dismantled, It's quite big and easily covers the entire anemometer. It'll spin with a motor from a cordless drill which will be powered by a 12V 5A AC transformer being rectified by a full bridge rectifier. Tried switching power supplies as high as 6A but they just kind of send a short pulse to the motor then stop then send a pulse again. I'm no expert in the workings of switching power supplies but I think what was happening was the motor itself draws a lot of amps with no load, but with a load (fan blade) I reckon it was trying to draw a huge amount of amps and the power supply has some sort of protection embedded to prevent this. Again this is just an assumption and just mentioned as to why I'm using the transformer and rectifier combo.

Some pics of the setup:

The fan blade in centered in both directions on the cups of the anemometer and the distance to the anemometer I don't think is critical. The rotations per minute will be measured in the Mega sketch then the handheld electronic anemometer will be placed exactly in the same spot as the anemometer and record the wind speed. From this a constant of double type (number with one or more leading digits) can then be deducted which will then be used in the Mega sketch to turn those rotations per minute into wind speed as mph.

[angryowl]

The average of several RPM values recorded by the Mega was recorded as 57.245 RPM at the "fan's" full speed. This was four distinct measurement sessions of several values each then averaged out.

With the handheld digital anemometer placed in the exact location as the other one and the measuring unit set to mph, the average was recorded as 8.5375 mph. This was again four distinct measurement sessions of several values each averaged out.

With these two numbers and with a bit of simple math a "constant" value was determined to be 6.7052. If the current RPM value recorded by the anemometer is then divided by this constant value the result is a value in mph. Simples

Inputting this constant value into the Mega sketch and doing several runs, the current wind speed values obtained were:

• A minimum of 7.46 mph and a maximum of 9.10 mph 
• A Mean or Average of 8.34 mph
• A Median of 8.35 mph
• A Mode of 8.65 mph

I do realise there are countless variables to be taken into account when doing such measurements and expecting very precise results, but I think for a simple wind vane I am happy with these numbers and will just use this in the final application.

[angryowl]

Tested the rain gauge with 100ml of water today. I used a small improvisation in which I had the water in a normal plastic bottle and poked a few holes through the cap using a tiny needle. This was to sort of simulate more of a rain effect rather than a stream of water through the funnel. 

Right, so measured 64 tips of the cup using the entire 100ml. Which gives me 1.56ml of actual water per half cup compared to my previously 1.6ml measured amount. I think that's quite a small error and should be fairly accurate in the field measuring real rain.

Any thoughts on whether I'm correct and it really is that accurate or I'm completely off the rails?

Also are any others out there with DIY rain gauges and how did you calibrate them? What difference did you get in the measured amount versus when you actually tested the fully assembled device?

Oh, and by the way currently printing the very last piece for the weather station. Now just need to sort out the PC restarting itself randomly and then up the pole it goes!

[Gina]

That is accurate

[angryowl]

1 hour ago, Gina said:

That is accurate

Yeah I'm hoping it'll be this accurate when mounted but not sure how long the rain repellent layer on the funnel last for. I suppose I'll just have to see. By the way do you think it would be a good idea to place a metal mesh over the funnel? This is so that any leaves and such don't fall directly inside the funnel but stay on the mesh and hopefully a gust of wind might blow them off?

[Gina]

Yes, I do.  Might also be a good idea to arrange some stiff wires round the edge pointing upwards to deter birds from sitting on it and leaving their "calling cards" when they depart.

[angryowl]

So alter many failed attempts at reinstalling drivers and checking uninstalling Windows updates the PC issues still persists. the short story is it randomly restarts and also randomly when turned on it will just sit there at the Windows log screen for hours (I tested) without the usual revolving circles indicating that Windows is indeed booting up your files. Now to the more advanced solutions for fixing this:

1. Try a manual Startup Repair. This can be accessed from withing Windows (when it's booted up of course) in the Advanced Startup options menu I think. So to summarize I tried its own automatic repair which failed, then went on to try and use a command prompt to try and use automatic windows backups to restore my registry files, but turns out my backup registry files were completely empty, 0 bytes! So that stop there! For anyone interested in the procedure it's here (works the same for windows 7 and 10)
2. I am now currently performing a clean Reset of the PC with the option to completely cleanup the files and apps and essentially I should be left with a brand spanking new (working!) Windows 10 installation.
3. My last and final option would be to do a Windows 10 reinstall using a manufacturer specific ISO from here . A tad bit confused at the end of the page, step 5, where they mention several drivers critical to correct operation will need to be installed, dunno if those are included in the ISO or installed automatically or I'd need to search for them.

For anyone interested the PC dongle that I have is this one, it's the one without an RJ45 port.

Will update once the reset is complete....

[angryowl]

Right, so the reset went through successfully so I'll reinstall all of the required applications and then do a few tests to see if the issue still persists. I like to think it solved it, if that were true then I could finally fit the enclosure lid and seal it off. Will report back

[angryowl]

Seems like the reset did it, no more restart loops! YAY 

[Dr_Ju_ju]

Hopefully your restart has fixed things, but I've had the same sort of issues you've experienced, which eventually turned out to be problems with disk drives (spinning rust & ssd's) whether it be bag blocks, corrupt data (files) etc.

Replacing was the best fix, which also gave peace of mind.

[angryowl]

7 minutes ago, Dr_Ju_ju said:

Hopefully your restart has fixed things, but I've had the same sort of issues you've experienced, which eventually turned out to be problems with disk drives (spinning rust & ssd's) whether it be bag blocks, corrupt data (files) etc.

Replacing was the best fix, which also gave peace of mind.

Yeah, if this was a regular PC/laptop with replaceable HDD then yes I could have replaced the HDD making sure the issue would never occur again if indeed it was bad blocks on the HDD. But seeing as the stick has 32GB of eMMC, that is sadly impossible and it's a shame but it is what it is.

Appreciate the advice though. 

I may be wrong and it might just appear the issue was fixed and then turns out the eMMC is corrupt as you pointed out and I'm completely f****d. 

[angryowl]

Ohhh, here we go...

The reset did nothing, upon further inspection I had the idea that the power supply for the PC stick could be showing signs of failing, so went ahead and replaced that, restarts still happen!!!!

At this point I'm stumped really. 

There's this one thing which may cause it; all of my power supplies are all connected and they all use the incoming power cord to draw power. None of them have ground capabilities so the only thing this was grounded to the main aluminium case was the 5V and 12V outputs. I'm no electronics wizard, but is this the right way of doing it? Anyone on here have any idea what might possibly be causing this? Should I be changing something in my current configuration? Is it the power adapter the most likely cause here, or is Windows still a viable option?

Guess the next thing to do it take out several multi meters and have points all overt he place and when a restart happens I at least have a better understanding of this??? 

[angryowl]

Any help or suggestions on what I may be doing wrong or things to try out would be much appreciated at this point as I'm stumped!

[AngryDonkey]

You've probably looked at this already but one of the models (I think this one https://www.amazon.co.uk/Sumvision®-Officially-Licensed-Bluetooth-Streaming/dp/B01LQ8J1H2) has a know problem with a USB driver which leads ti blue screen of deaths (and probably a restart). On the Sumvision website there is a support article for this and a custom ISO download for Windows 10 to fix it. Not sure if this applies to your model though.

[angryowl]

1 minute ago, AngryDonkey said:

You've probably looked at this already but one of the models (I think this one https://www.amazon.co.uk/Sumvision®-Officially-Licensed-Bluetooth-Streaming/dp/B01LQ8J1H2) has a know problem with a USB driver which leads ti blue screen of deaths (and probably a restart). On the Sumvision website there is a support article for this and a custom ISO download for Windows 10 to fix it. Not sure if this applies to your model though.

Thanks AngryDonkey for the suggestion, although this is not my model it's pretty close and I think Sumvision kind of use the same motherboard with the Cherry Trail series with slight modifications.

So this USB driver issue might apply to my stick as well so much appreciated for the heads up on this as this gives me something else to try.

I do also have the Windows 10 ISO for my stick but have not tried it yet. Trying to figure out if it's a hardware (power supplies, etc) or the PC itself.

[angryowl]

Have just finished the ISO Windows 10 reinstall as recommended by the manufacturer and all went well, just reinstalling the various pieces of software and time will tell if this fixed the reboot issue.

I really do hope it did as I'm almost finished with the rest of the build and it's so disheartening to have this happen just now