First tower. Solar Dahua.

ronan

Young grasshopper
Jul 21, 2015
33
14
Hey all,

Have a look, questions/criticisms welcome. Still some work to do.

IMG_20160905_155312240.jpg
IMG_20160925_141922356.jpgIMG_20160925_142905824.jpg

The worst thing about this site is having to sign in to see thumbnails!
IMG_20160925_142041734.jpg

This is my next box. Using an old switch I have lying around and an existing 3g modem. Is a work in progress.
IMG_20160925_143529243_HDR.jpg
 
Last edited by a moderator:
excellent work, very nice.

how far out is this?
 
Thanks man,

This tower is only 1km from the AP. 2MP 1080p dahua, and 4MP dahua cams run fine on it. 10ms ping, 60mbit connection with the cheap litebeam - could be aimed a bit better.
 
Some shots from the 2MP PTZ. Note the zoom level top left. Obviously would be even better with the 4MP version kicking around these days!

Car @ 65m
Cam2.20160925_083325_1b.jpg

Car @ 210m
Cam2.20160925_083325_1a.jpg

Car @ 575m
Cam2.20160925_083325_1.jpg
 
Well i'm not 100% sure yet. The battery was discharged when i first put this up, and we've only had overcast days since then. So with that in mind, the charge controller cut power at 430am and reconnected at 830am when the battery was recharged enough.

It was fairly sunny today and i checked the battery voltage in the arvo. It was at 14.3V. So it was probably fully charged today. Will see how long it lasts tonight.
Its only a single 100AH battery, may have to get a bigger one / turn some IR off, or add a wind generator :)
 
Yeah thats what I was thinking. I have heard mixed things about it though. Physical reliability of the propellers or generator and wind availability itself is a bit iffy.

Planning on chucking wind/water meters on it tho which hook into the pi.

Actually yeah, I should have just used an old windmill! They can be bought pretty cheap when they come up..
 
Last edited by a moderator:
Nice project. Sure gives a greater FOV than mounted on a corner of a house. :)
 
Nice setup!

Couple questions: I see an RPI in the box with a relay -- What task is that doing out of curiosity? The RPI itself will almost certainly corrupt if power is kicked from it occasionally, so an in-line USB battery (mobile cell charger) would ideally help with that.

What is your battery setup? (Amp hour rating). I'm exceptionally curious with this setup and hope that you keep us updated because my tests (using 1/5th the power your setup uses) were pretty dismal with double the panels and likely double the batteries. That said, I live in a Northern region where it's pretty solar prohibitive anyways, but am still curious and may give me energy to try again. :)

Solar charger: PWM or MPPT? (If you're not getting the full potential of your batteries and are on what looks like PWM for both boxes, switch over to MPPT)


Your concept box with 3G: Tell me a little about how you plan to set that up. I only see the WIFI router and 3G modem/router, and given how cellular data networks hand out IP addresses, this may require a little more gadgetry to work successfully. Are you planning on doing live view? 24/7 stream or just local recording/playback? Monthly data limits? (If you get to this project and have questions, let me know.) I'm not sure if cell networks are structured differently where you live so this setup may vary a tad.
 
Nice setup!

Couple questions: I see an RPI in the box with a relay -- What task is that doing out of curiosity? The RPI itself will almost certainly corrupt if power is kicked from it occasionally, so an in-line USB battery (mobile cell charger) would ideally help with that.

What is your battery setup? (Amp hour rating). I'm exceptionally curious with this setup and hope that you keep us updated because my tests (using 1/5th the power your setup uses) were pretty dismal with double the panels and likely double the batteries. That said, I live in a Northern region where it's pretty solar prohibitive anyways, but am still curious and may give me energy to try again. :)

Solar charger: PWM or MPPT? (If you're not getting the full potential of your batteries and are on what looks like PWM for both boxes, switch over to MPPT)


Your concept box with 3G: Tell me a little about how you plan to set that up. I only see the WIFI router and 3G modem/router, and given how cellular data networks hand out IP addresses, this may require a little more gadgetry to work successfully. Are you planning on doing live view? 24/7 stream or just local recording/playback? Monthly data limits? (If you get to this project and have questions, let me know.) I'm not sure if cell networks are structured differently where you live so this setup may vary a tad.

Thanks,

The RPI just runs a little python script at reboot/every minute that checks if the on-board temperature is >45 degrees C and turns on the relay, which is connected to a 12V computer fan. Obviously the on-board temp is going to be much higher than outside temp, but i'd like to keep this equipment as cool as possible, since out here we regularly have 45 degree days in summer. I could probably just have the 0.08 amp/12V fan running non-stop, but its been a good python learning experience.

I just received a wind/rain meter kit from sparkfun.com so i'll play around with them, get them to work, chuck em on the tower pi, and update a website with real-time values. Long term i'd like to get into graphing some data.

Yeah good point about not shutting down properly... its still working so far!

Currently running a single 100AH deep cycle battery. Its got a picture of a guy with a fish on the front. I tried to find the recommended charge settings etc. but could not! So my PWM charge controller is on defaults, but i think it tops out at around 14.5V and cuts off at 12.5V. Those numbers sounded fairly safe. Last night however the system cut out at 1:30am. So it definetly didn't have enough charge. I'm thinking its mostly a case of just not having enough battery capacity. After doing some quick wattage calcs in my head I guess all this gear would be pulling around 50 watts/hr, will have to think about this a bit more. Would be awesome to have the pi (or something smart - switch etc.) give me live voltage readings. I have another of these batteries, so i'll wire it up for 200AH. My latitude is ~36 degrees S.

Having some drama with the 3g box. Plan is to run openWRT, or more specifically the aussie "rooter" derivitave which is built to make 3g modems work easily, and have it connect home using openvpn. It gets a dynamic public address, so i was hoping i could set it to reboot and just connect to my openvpn server (somehow)...

Picked up a TP-link MR 3020 to plug into that old router, and a 3G modem i have here. the 3g modem is "host-less" so it presents itself as an ethernet device to the tp-link. I probably could, but haven't been able to get vanilla openwrt to use this 3g modem properly (the linux usb commands are useless for this modem) - but it works sweet with rooter.

The problem with that TP-link is it has 4MB ram, and as such doesn't fit openvpn on it easily. If i made my own build of openwrt it could be done, but I dont have the time to learn how to do that right now.
Picked up a cheap tp-link WR842N to replace the old router and other tp-link, but they sent me a different version to what I ordered, and it isn't supported fully by openwrt or rooter! pretty annoying.

I can get a 4 or 8 gig/month plan here, so hoping for some kind of continual recording (even if its at lower quality) and high quality live view.

So yeah! (once i get a decent router) do you think my plan will work?
 
  • Like
Reactions: ronan
i think it tops out at around 14.5V and cuts off at 12.5V. Those numbers sounded fairly safe.

So you are cutting off the discharge with approximately 75% left in the battery.
A fully charged flooded lead acid battery should settle somewhere around 12.6-12.7v and run comfortably down to about 11.8-12v. You are barely scratching the surface of the charge if you are cutting off at 12.5v
 
So you are cutting off the discharge with approximately 75% left in the battery.
A fully charged flooded lead acid battery should settle somewhere around 12.6-12.7v and run comfortably down to about 11.8-12v. You are barely scratching the surface of the charge if you are cutting off at 12.5v

Ah, right.. well I could be mis-remembering. I'll double check
 
Ah, right.. well I could be mis-remembering. I'll double check

Roughly you have 1200WH of storage available. You estimate your draw at 50W, so you have ~24 hours of power available. You need to charge at about 12A for 10H to get the battery fully charged.
If you have 8H of daylight, you are drawing for 16 hours and should use ~66AH, you'll need about 10A for 8H to put that back in plus the 4.1A you are drawing at 50W.

Your biggest issue is going to be getting 169W of solar energy averaged across 8 hours to charge things back up. A wind generator on that flimsy tower will shake that camera enough to make it pretty annoyingly useless under any great wind. It takes very little vibration to set up harmonics in a structure. It's why you don't put cameras on high light poles. They pendulum something terrible under a stiff breeze.

Nice set up, just looks like you need to manage your power draw. What size is your panel?
 
  • Like
Reactions: ronan
Thanks,

Yeah good point about not shutting down properly... its still working so far!
No doubt a fun python learning experience! For volatile installations like this I'd probably swing the Arduino route just for the sake of eliminating OS corruption. It may never happen to you or it may happen every month--all depends on image version and what is running on it and pure luck sometimes.

Currently running a single 100AH deep cycle battery. Its got a picture of a guy with a fish on the front. I tried to find the recommended charge settings etc. but could not! So my PWM charge controller is on defaults, but i think it tops out at around 14.5V and cuts off at 12.5V. Those numbers sounded fairly safe. Last night however the system cut out at 1:30am. So it definetly didn't have enough charge. I'm thinking its mostly a case of just not having enough battery capacity. After doing some quick wattage calcs in my head I guess all this gear would be pulling around 50 watts/hr, will have to think about this a bit more. Would be awesome to have the pi (or something smart - switch etc.) give me live voltage readings. I have another of these batteries, so i'll wire it up for 200AH. My latitude is ~36 degrees S.
Others have provided good insight so far on this as well, but I'd say the more you shave off of this energy consumption wise the better. 50w/hr is a little intense and my guess is the cam is mostly responsible, but anything else that can be removed should be. The .8a/12v fan would be my first elimination despite the heat. Perhaps a lower wattage fan could work.

Having some drama with the 3g box. Plan is to run openWRT, or more specifically the aussie "rooter" derivitave which is built to make 3g modems work easily, and have it connect home using openvpn. It gets a dynamic public address, so i was hoping i could set it to reboot and just connect to my openvpn server (somehow)...

Picked up a TP-link MR 3020 to plug into that old router, and a 3G modem i have here. the 3g modem is "host-less" so it presents itself as an ethernet device to the tp-link. I probably could, but haven't been able to get vanilla openwrt to use this 3g modem properly (the linux usb commands are useless for this modem) - but it works sweet with rooter.

The problem with that TP-link is it has 4MB ram, and as such doesn't fit openvpn on it easily. If i made my own build of openwrt it could be done, but I dont have the time to learn how to do that right now.
Picked up a cheap tp-link WR842N to replace the old router and other tp-link, but they sent me a different version to what I ordered, and it isn't supported fully by openwrt or rooter! pretty annoying.

I can get a 4 or 8 gig/month plan here, so hoping for some kind of continual recording (even if its at lower quality) and high quality live view.

So yeah! (once i get a decent router) do you think my plan will work?
This is where I come in with advice/critique/sad news. The hardware isn't going to be your demise, but if you're planning on using any of that on solar, consider revising the hardware list to simplify it immensely--starting with the netgear wifi switch. Your total power for ALL networking components should be less than 3-4W. Focusing just on the networking/software, you'll find that OpenVPN will be too "chatty" on a 3G/4G connection. In other words, a constant VPN connection like that will cost you bytes per second which add up over time.

Planning and executing a cellular based IPCam setup isn't for the faint of heart :) A lot of focus with standard installations is on gigabytes, megabytes, etc... but with cellular you have to focus on bytes. Not kilobytes. Assuming OpenVPN isn't too chatty, you can still expect its keepalives to easily cost you 300-350MB/mo on low end--and that's just to keep the connection alive. Any data you wish to push through is on top of that.

Continual stream of even ~40kbps will cost you around 11-12GB/mo I'm estimating. To get an idea of what that looks like, pop open your Dahua and set it to the lowest stream it can handle. Then PTZ around to see how it fights between iframes.

This isn't to scare you away from it, just some things to consider. There are ways of doing this (minus 24/7 stream) inside of 1-2GB/mo. and still have everything sit at a beautifully encoded 1080p w/o noticeable compression. It's just a math game. Some of my 4G installs when I started doing this went over 2GB/mo but it wasn't hard to catch the data leaks to prevent it from happening again. Most I have sitting at 1GB/mo with off-site storage, 1hr FTP snaps, and live view, but again it's all about monitoring every byte to make this as efficient as possible to do everything you want/need to do.

(responded in quote)
 
I second the idea of using an arduino over the rpi. It will draw much less power and with some PWM routines in code you can modulate the speed of the fan vs temperature and save some of those valuable watts. There are so many examples online of people that use it for projects like this. You can then have it send you packets of data that contain the telemetry info (voltage, status, etc) instead of running a webserver on the tower. Have that data then populate a rpi at your head end running a webserver. The hardest thing to overcome is the expectation that solar panels are light sabers.
 
pretty dismal with double the panels and likely double the batteries
gmaster, haha, think you called it mate, looks like I should have done my homework. Last night the system cut out at 1130pm.

I live in a Northern region where it's pretty solar prohibitive anyways
What is your latitude?

You are barely scratching the surface of the charge if you are cutting off at 12.5v
Brad, Seems I remembered wrong. My low voltage disconnect was set to 11.5V. After doing some reading I decided to bump this up to 12.2. I've also disconnected some equipment until I sort out power to avoid draining the battery too heavily.

What size is your panel?
200W

Okay, you've got me thinking... I'll run through the numbers and try to understand what you're saying!

I added up or guesstimated the max power usage from spec sheets.
Fan 1W
Switch 4W
Dahua PTZ 13/23(IR on)
Dahua 4mp 4W
LBE 4 W
Charge controller 0.5W?
RPI 3.5W

Total ~ 40W night/ 30W day.

(40W *12) + (30W *12) = 840 W/day usage max
97AH 12V battery = 1164W battery capacity

My charge controller actually tells me how many total amp hours were charged the previous day, and the day before that.
Average over those two days was 37AH/day. I assume I multiply this by 12V, so

37 x 12 = ~ 444W/day of charge.
Looking up some numbers, the middle of winter only gets 3/4 the solar radiation as this time of year, so

333W/day charge.

If I add a second panel that bumps it up to 666W/day in winter vs 840W max usage. Obviously there is still a deficit.

My single panel could be under-performing, or I might have just had some overcast days that are throwing the balance out a bit.

Do you think i'd pull through if I double the panels (total 400W) and more than doubled battery capacity (260AH)? I'd basically have 3.7 days worth of battery capacity.

I have an old 60W polycrystalline panel here too. I'd like to avoid using it, but I assume there would be no issues wiring that into the system?

Thanks for running the numbers!

Your biggest issue is going to be getting 169W of solar energy averaged across 8 hours to charge things back up
Yep... So now its sinking in that a 200W panel probably only makes that under ideal conditions. Looking up some tables, it seems panels should make maybe 40% their rated capacity in the middle of winter. Add some cloudy days to that and things get pretty dire. I'm obviously making too many assumptions here, and not accounting for reduced sunlight hours during winter properly.

It will draw much less power and with some PWM routines in code you can modulate the speed of the fan vs temperature and save some of those valuable watts.
Sweet idea!

have it send you packets of data
Yeah that'd be nice.
 
Last edited by a moderator:
Hmm cheers for the insight gmaster,

Yeah i'm getting rid of that old netgear. Will have to do some research and get something more suitable (I may ask you for recommendations on this), i've seen mention of mitrotik/routerboard/netonix or davado gear, and my old work had some built for field 3g modems which i wrote the name down for. And probably go with a simpler usb modem.

Actually... maybe a raspberry pi would be one of the lowest power routers out there...?

I did not realise VPNs were that bad data wise.

Continual stream of even ~40kbps will cost you around 11-12GB/mo I'm estimating.
yeah true, just did that calc. Yeah FTP seems like a solid option.

So I'm just trying to get straight in my head how i get the remote cameras to connect to my home BI server (without using a VPN).

I suppose I can set up the cams to use the dynamic DNS built into their firmware. And then in BI i just point to the cams hostname and path?

This means i'd need to forward ports on the 3g router to allow the http or rtsp connection to the cam. Is that pretty much it?
 
Last edited by a moderator: