Uninteruptable power supplies - and CCD cameras

[Tonk]

My mount and cameras + other devices is controlled by an "on-board" Intel NUC computer. The rig is going to a remote site that uses solar charged batteries to provide mains power. Therefore there is a real risk of a power failure at various time for various reasons. So I'm building a custom UPS set up to at least ensure that the NUC computer running Windows 7 has a fair chance to shut down cleanly.

This is the UPS board I'll be using (its on order via Amazon UK - delivered from the USA - expecting it next week) - http://www.mini-box.com/OpenUPS2

You have to obtain your own batteries (x3) which is no big deal though they have to be LiFePO4 chemistry and not Li-Ion. You will also need to source an appropriate sized project box and various connectors (USB, Power SW, Power In, Power Out). The other setups that have to be done are:

1) Set the NUC BIOS to auto-boot on detecting power (though the OpenUPS2 device can do the external power detection and send an "on" pulse to the NUC as an alternative method)
2) Connect a pair of wires to the Power SW header on the top face of the NUC mother board (the two red pins on the "front panel" header) - my NUC has the NUC5i5RYB motherboard - http://www.intel.com/content/dam/support/us/en/documents/boardsandkits/NUC5i5RYB_NUC5i3RYB_TechProdSpec04.pdf - I've brought the wires out of the NUC box via a ventilation slot. These are connected direct to the OpenUPS2 board to the P6 header so that OpenUPS2 can command the NUC to shutdown (or start up).
3) Connect the OpenUPS2 USB header (P3) to either a spare external USB socket on the NUC (all 4 are USB 3.0) or to one of the two internal USB 2.0 headers on the NUC motherboard. OpenUPS2 uses this to detect when the NUC motherboard is powered (via the +5v line) and to allow the OpenUPS2 configuration software to let you set the rules for when the power off pulse gets sent to the NUC computer and to set the voltage boost value (the NUC is rated for 19 volts but does run OK at 12 volts - you can chose the voltage for your computer). The rules could be a measured time-out after main power loss or on a battery voltage threshold.
4) Configure Windows to force shutdown without having applications blocking the request (usually the block is done with a dialog to warn the user to save data/documents - if not dismissed then the shut down gets stalled until the UPS batteries run out - not what you want). You need unattended guaranteed shutdown to keep Windows in a clean state. This involves using the Windows registry editor to add the "AutoEndTasks" parameter and set its value to 1 (see http://www.theeldergeek.com/close_applications_not_responding_at_shutdown.htm ) .

In the absence of the ordered OpenUPS2 board I've tested 1), 2) (using a momentary shorting switch in place of the UPS) and 4) and it all works - the NUC cleanly shuts down Windows and turns off.

Now here is my question - as well as the computer, should you control the shutdown of CCD cameras (I have an active cooled QSI) so say control the warm up of the camera before shut down. I'm hoping CCD cameras can tolerate power loss at any point without damage - as otherwise I need to get a little too sophisticated - needing to write some software to interrogate the OpenUPS USB port (it reports the UPS state) recognise power has been lost and the UPS is heading for time-out shutdown and somehow yank the CCD control from Sequence generator pro and command the camera to warm up (unless there is a way to command SGP to do all this for you!)  

Any thoughts on the camera question?

(rig before NUC hacked to connect up Power SW cable - )

[Gina]

I don't like LiPos - they have a tendency to catch fire! Or even explode

[Davey-T]

I always warm up my QSI683 before shutting down but read somewhere that it was built to withstand sudden power loss.

Nowt wrong with LIPos as long as you treat them nicely 

Dave

[Tonk]

"I don't like LiPos - they have a tendency to catch fire! Or even explode " 


..... take care here LiFePO4 is not the same as LiPo - different chemistries (I'm a chemist so this is right up my street ). They have distinctly different characteristics. The main characteristic of LiFePO4 is its ** safety **  and long life! - ideal for unattended UPS applications

LiFePO4 - https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery

"One important advantage over other lithium-ion chemistries is thermal and chemical stability, which improves battery safety.^[9]LiFePO4 is an intrinsically safer cathode material than LiCoO₂ and manganese spinel. The Fe-P-O bond is stronger than the Co-O bond, so that when abused, (short-circuited, overheated, etc.) the oxygen atoms are much harder to remove.
...
As a result, lithium iron phosphate cells are much harder to ignite in the event of mishandling (especially during charge) although any fully charged battery can only dissipate overcharge energy as heat. Therefore, failure of the battery through misuse is still possible. It is commonly accepted that LiFePO4 battery does not decompose at high temperatures"

LiPo (Li polymer) - https://en.wikipedia.org/wiki/Lithium_polymer_battery - here the polymer refers to the polymer battery casing - the underlying chemistry is standard Li-Ion

"LiPo cells are affected by the same problems as other lithium-ion cells. This means that overcharge, over-discharge, over-temperature, short circuit, crush and nail penetration may all result in a catastrophic failure, including the pouch rupturing, the electrolyte leaking, and fire"


So LiFePO4 batteries is the better choice made by the OpenUPS2 designers. Its worth pointing out the prior OpenUPS board (still available) supports Li-Ion (same as Li-poly/LiPo), LiFePO4 and lead acid but consequently comes with no on-board battery cell holders. You have to configure the board parameters to match the battery used. Also the newer OpenUPS2 board can be hacked (a resistor needs changing and different  charging parameters uploaded) to support Li-Ion but I don't recommend this due to afore mentioned Li-Ion battery safety issues

[Tonk]

"I always warm up my QSI683 before shutting down but read somewhere that it was built to withstand sudden power loss. "

 

Cheers dave - I'll check this out and if OK I'll set the UPS to deal only with the NUC shutdown - that's just wires and no extra software

 

Update: Typical - I use google to search for "QSI683 sudden power loss" and the best hit is .... you guessed ... this thread!

[Gina]

15 minutes ago, Tonk said:

"I don't like LiPos - they have a tendency to catch fire! Or even explode " 


..... take care here LiFePO4 is not the same as LiPo - different chemistries (I'm a chemist so this is right up my street ). They have distinctly different characteristics. The main characteristic of LiFePO4 is its ** safety **  and long life! - ideal for unattended UPS applications

LiFePO4 - https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery

"One important advantage over other lithium-ion chemistries is thermal and chemical stability, which improves battery safety.^[9]LiFePO4 is an intrinsically safer cathode material than LiCoO₂ and manganese spinel. The Fe-P-O bond is stronger than the Co-O bond, so that when abused, (short-circuited, overheated, etc.) the oxygen atoms are much harder to remove.
...
As a result, lithium iron phosphate cells are much harder to ignite in the event of mishandling (especially during charge) although any fully charged battery can only dissipate overcharge energy as heat. Therefore, failure of the battery through misuse is still possible. It is commonly accepted that LiFePO4 battery does not decompose at high temperatures"

LiPo (Li polymer) - https://en.wikipedia.org/wiki/Lithium_polymer_battery - here the polymer refers to the polymer battery casing - the underlying chemistry is standard Li-Ion

"LiPo cells are affected by the same problems as other lithium-ion cells. This means that overcharge, over-discharge, over-temperature, short circuit, crush and nail penetration may all result in a catastrophic failure, including the pouch rupturing, the electrolyte leaking, and fire"


So LiFePO4 batteries is the better choice made by the OpenUPS2 designers. Its worth pointing out the prior OpenUPS board (still available) supports Li-Ion (same as Li-poly/LiPo), LiFePO4 and lead acid but consequently comes with no on-board battery cell holders. You have to configure the board parameters to match the battery used. Also the newer OpenUPS2 board can be hacked (a resistor needs changing and different  charging parameters uploaded) to support Li-Ion but I don't recommend this due to afore mentioned Li-Ion battery safety issues

I stand corrected   Thank you for the info

[Davey-T]

Both my batteries are lifepo4

 

Dave

[hughgilhespie]

Hi Tonk,

I am impressed by your project! I am also considering using an NUC or similar so that the computer, with its USB ports is on the mount with the scope, camera, etc. I am - slowly - automating my 2.2 metre Pulsar dome and, after the mechanics of rotating the dome and opening the shutter are complete, this seems a logical next step. I would use WiFi to connect to the dome control electronics, so the only cables to the mount and NUC would be a 14 volt UPS for the mount and NUC and a Cat 6 to connect to my network.

I have a couple of questions about using an NUC if you don't mind. Firstly I assume some sort of SSD will be used with the NUC. Does the board have the necessary mechanical connectors or do you have to build an enclosure that takes the NUC and the SSD?

Second question is about the local display.  I would probably install Win 7 Pro as the NUC OS and then use Microsoft Remote Desktop Protocol to allow remote operation of the NUC but I have seen comments somewhere that says you also need a local display attached to the HDMI port. Ideally I would like to operate the NUC remotely without a 'local' display but have the option to connect a local display when working in the observatory. My dome is in my garden, so easy enough to access if needed. 

I would love to learn how you have addressed these points and I wish you every success with your well thought out project.

Regards, Hugh

[Tonk]

Hugh - when you order a NUC from say an integrator company such as http://www.box.co.uk/ you buy the bare bones kit with various extras such as on-board SSD "disk", memory and OS - you get a list of options. They fit the extras into the NUC box, install the OS, NUC device drivers and test the system before sending on to you. You can chose NUCs with SATA disks or SSD (case sizes differ)

I chose a NUC with 4Gb memory (no plans to process images on the NUC) and an on-board 250Gb SSD.

I didn't select an OS as I could get a "refurbished" Windows 7 install disk [these can occasionally be found on Amazon market place! - Microsoft collect OS license keys from companies that dispose of old computers and rerelease these to OEMs for reuse at a much reduced price - some OEMs resell these disks and license codes for about 30% of the price of a new "first time used" OS disk]. If you do go the refurbished OS route then you will have to collect all the device drivers from the Intel website - its can get tricky finding out what your NUC board needs - e.g. SMBUS drivers, IR receiver drivers, on-board wifi etc. Added complication is old Windows 7 install disks don't know about USB 3.0. I had a small problem when installing from an external DVD plugged into the USB 3.0 port. After one of the numerous machine reboots the installer commands, the USB 3.0 <-> DVD link was lost (dialog pops up asking for installer location - DVD drive is not on known list of locations - darn!). The solution here is very simple - unplug the DVD USB cable and reconnect. The boot loader gets the "plugged in" signal and enables the USB link - you can now carry on. 

Beyond that I've elected to add an additional USB 3.0 external Wifi dongle with dual aerials (a Netis AC1200 Wireless Dual band USB adapter). Windows 7 happily uses both the on-board wifi and the external one at the same time. This is a nice bit of redundancy in case one or the other gets positioned badly by the mounts current position + plus my NUC is sandwiched between two aluminium plates. I've also positioned the Netis aerials at 90 degrees to each other as that keeps the received signal strength reasonably constant no matter were the scope is pointing. This might not be an issue for your set up (your transmitter is close) - but for me I don't know where the remote site transmitters are placed for the "shed of scopes" - I'm taking no chances  

[Tonk]

As for the display - I only attached one for the initial OS install and getting base level apps installed. Once I set up TeamViewer AND Google Remote Desktop I removed the monitor and have since only used the remote desktop links. Again you need to ensure 2 or more alternatives in case the one of the other is having comms issues (it occasionally happens that one remote desktop app wont connect while the other is OK)