Community discussions

MUM Europe 2020
 
WirelessRudy
Forum Guru
Forum Guru
Topic Author
Posts: 3089
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 4:27 pm

rb750UP is connected to solar and 300A/H gel batteries
Voltage during the day is 13,7-14V mesured on the rb750UP

2 x rb433AH both with one R52Hn radio transmitting at full default power in 802.11a/n mode with all rates allowed (default + advanced setting).
Both get their PoE power over approx. 10 meter standard utp cable.
Both units give a reading of approx. 12,5V when the rb750 reads 13,7V. So a drop of roughly 1,2V. This is during the daytime hours with bright light or sunlight.

Since a couple days had lots of problems with one of the rb433's disconnectiong or loosing ethernet connectivity.
After checking for water ingress at any point, checking and replacing the cable connectors I replaced the board for a before used rb433AH.
Initially it worked fine but after sunset the voltage dropped on the batteries to a reading of 12,1 or 12,2 at the rb750UP.

One of the rb411AH boards has now a power drop of 0,4-05V and reads 11,5 to 11,7V and works normally with some 10 clients connected (and both chains working at default power)

The other rb411AH tho, with some 45-50 clients attached still had a voltage drop of 1,2 wich resulted in a reading in winbox on only 11V. I even saw a 10,9 or 10,8V at times.... (I presume lower ones are not even noted because board than either crashes or ethernet connectivity is lost.)

Looking in the rb750UP I see the amps fluctuate a lot of the feed to the troubled rb411AH. The board also started loosing its ether connectivity and only by setting the radio power to 17dBm (in stead of the 'default' which is 23dBm) I got the board back stable.
Looking in the log I found some alerts (I have a script for that) messaging alarming low voltage.


So:
Is a voltage drop of 1,2V normal for approx. 10mtr normal utp cable?
Would a voltage in the region of 11V make the ethernet interface become unstable?
Is the fact that many clients are connected having an influence compared to a lower used radio?
Is the amount of traffic clients produce having any effect on the power consumption of the radio/board?
Is the established connection rate between AP and clients having any effects on the power consumption?
(More power consumed the lower the voltage?)
(When I started to mess with the connection rates and set only some rates. This way forced many clients to even in idle status to maintain higher rates. It made the board inmediately crash......)

My solution is going to be to upgrade the instalation into a 24V system, with the risk on very sunny days the voltage might becomes too high?

Anybody any better insight? Or advices?
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
JackANSI
Frequent Visitor
Frequent Visitor
Posts: 71
Joined: Wed Apr 03, 2013 6:52 pm

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 7:35 pm

If the PoE out of the RB750UP goes through any semiconductors (for switching or reverse protection) you'll see some voltage drop (look up "forward voltage drop") just inside the box. You could try powering the RB433AH's directly instead of going through the RB750UP that will tell you if its just a voltage drop issue or something deeper. Remember as voltage decrease, the current required will increase (to maintain the given output for the wireless).

I find MikroTik's choice to use something other than the standard 802.11af PoE unfortunate as you would have ample voltage and current under 802.11af.

Also be aware that in MikroTik's manual for PoE-Out states:
Overload protection
when port is on (PoE out is enabled) port is checked for overload, If that is detected to avoid hardware damage of powered device or powering device PoE out is turned of on this port.

With new firmware after overload is detected starts 120 second countdown after what PoE Out feature will be turned on again to see if environment has changed and device connected can be supplied with power again. That is important for configurations that are not connected to mains (solar installations, equipment running on batteries due to mains failure) so that when voltage drops - overload will be detected and connected devices turned off. After a while when voltage level returns to usual operating value - connected equipment can be powered on again.


Note: PoE-Out controller firmware version 2.0 allows 1A load on port with 2.2A of total limit on all ports
Make sure you're not exceeding the 1A/port and 2.2A/total limits.



To calculate the voltage drop over a wire:
http://www.calculator.net/voltage-drop-calculator.html

Remember: in MikroTik-land, positive is on the blue pair and negative is on the brown pair so select 2 conductors in the calculator

I've noticed some cheaper ethernet wire on the market that is using 25AWG instead of the usual 24AWG
The drop over 10m of 24AWG is 7% (.84v), but over 25AWG it is 8.83% (1.06v) (assuming 12V 1A). Make sure you've got good quality ethernet cabling. If you've got that super cheap "copper bonded aluminum" cabling you're riding somewhere around 22% loss over 10m.

This also does not take into consideration the voltage drop over connectors or junctions. A crimp style ethernet cable has two junctions (pin to wire) and two connectors (pin to pin) because you've got one on each end of the ethernet cable. So you can probably take a bit more off for those. So I'd have to guess that anything between .8v and 1.4v drop over 10m could be considered "normal".

Depending on the site, I would try:
1. Running the power to the RB-750/RB-433's separately over something like 12AWG, that would cut your voltage losses by almost 80%.
2. Putting some kind of RF noise suppression on the power pairs near the devices (ferrite beads) as I don't know if MikroTik would have put noise suppression on the power lines inside their devices and it may be stray RF noise that is causing some of your headaches. (or make sure your shielded cable is properly terminated, you are using shielded cable right? ;))
Last edited by JackANSI on Thu Apr 25, 2013 7:41 pm, edited 1 time in total.
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 7:40 pm

Why don't you put TWO batteries in series to get 24 - 28VDC and a solar of 24V or 2 x 12V in series?
You will be MORE stable, PLUS the batteries in the case of 24VDC, could go down (due to discharge) even at 14VDC and then "cut them down".
Your way, IF the battery discharges a BIT (let alone the drop on the ethernet cable), you are DOWN.

EDIT: The DC power when entering the input jack of RB750UP, passes through diodes to go out to the PoE Ethernet RJ-45, so you have a droped voltage on them too.


.
Last edited by hipro5 on Thu Apr 25, 2013 7:54 pm, edited 1 time in total.
 
JackANSI
Frequent Visitor
Frequent Visitor
Posts: 71
Joined: Wed Apr 03, 2013 6:52 pm

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 7:54 pm

Why don't you put TWO batteries in series to get 24 - 28VDC and a solar of 24V or 2 x 12V in series?
You will be MORE stable, PLUS the batteries in the case of 24VDC, could go down (due to discharge) even at 14VDC and then "cut them down".
Your way, IF the battery discharges a BIT (let alone the drop on the ethernet cable), you are DOWN.
Yes 24v would be better, but depending on his setup, switching to 24v could mean anything from half the run time, or spending a pile of money to double up.

I also wouldn't recommend running a pair of 12V batteries connected for 24v down to 14V.. thats going to damage them.
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 7:56 pm

I also wouldn't recommend running a pair of 12V batteries connected for 24v down to 14V.. thats going to damage them.
That was an example..... :D
Yes 24v would be better, but depending on his setup, switching to 24v could mean anything from half the run time...
Nope........Running at 24VDC, RBs draw almost the SAME wattage as they do running at 12VDC - They DON'T draw the SAME CURRENT, they draw the SAME WATTAGE due to switching mode regulator circuits they have build in. ;)
 
JackANSI
Frequent Visitor
Frequent Visitor
Posts: 71
Joined: Wed Apr 03, 2013 6:52 pm

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 8:51 pm

I also wouldn't recommend running a pair of 12V batteries connected for 24v down to 14V.. thats going to damage them.
That was an example..... :D
Yes 24v would be better, but depending on his setup, switching to 24v could mean anything from half the run time...
Nope........Running at 24VDC, RBs draw almost the SAME wattage as they do running at 12VDC - They DON'T draw the SAME CURRENT, they draw the SAME WATTAGE due to switching mode regulator circuits they have build in. ;)
But if he has an odd number of 12v batteries he'll still have to put money in at least there (or lose run time), possibly on the solar panel side as well (hopefully the charge controller he has will do 12/24v) 8)
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 9:43 pm

Thanks JackANSI. :D

Yes he has to buy another solar panel plus another battery - BUT afterwards he will encode NO MORE problems with his hardware due to power. :)

A 2.3Ah (2x12V batteries in series) can handle 4:30 hours before shutting down an RB750UP, an RB435G with a DBII N pro on it.
He has 300Ah :shock: battery (which probably will not be JUST one so he could "separate them" to get 24V out of them).
He could ALSO have more than one solar panels to charge them - so he could split them too.
Supposenaly that afterwards gets 150Ah total, with his config, he could be uptime, WITHOUT EVEN charge the batteries, for ~ 200hours thus 8 FULL DAYS - not bad...
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Thu Apr 25, 2013 10:11 pm

My solution is going to be to upgrade the instalation into a 24V system, with the risk on very sunny days the voltage might becomes too high?
DON'T worry.
You can feed the RB750UP with 30VDC... ;)
 
WirelessRudy
Forum Guru
Forum Guru
Topic Author
Posts: 3089
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: Battery feeded PoE voltage drop over 10m utp normal?

Fri Apr 26, 2013 12:54 am

If the PoE out of the RB750UP goes through any semiconductors (for switching or reverse protection) you'll see some voltage drop (look up "forward voltage drop") just inside the box. You could try powering the RB433AH's directly instead of going through the RB750UP that will tell you if its just a voltage drop issue or something deeper. Remember as voltage decrease, the current required will increase (to maintain the given output for the wireless).
The reason for recently starting using the rb750up is to have the remote power cycle option. Before I always had to jump in the car to have a 20 mins drive to the tower, usually in the middle of the night, to power cycle if a units 'hangs'. (Which sometimes does happen. Not a lot, but always at very unconveniant hours....)
I find MikroTik's choice to use something other than the standard 802.11af PoE unfortunate as you would have ample voltage and current under 802.11af.
I always wondered why MT (and others like ubnt) are using 'non standard' voltage. But if you work with batteries and solar you are actually glad they didn't. It keeps installations simple and cheap.
Also be aware that in MikroTik's manual for PoE-Out states:
Overload protection
when port is on (PoE out is enabled) port is checked for overload, If that is detected to avoid hardware damage of powered device or powering device PoE out is turned of on this port.

With new firmware after overload is detected starts 120 second countdown after what PoE Out feature will be turned on again to see if environment has changed and device connected can be supplied with power again. That is important for configurations that are not connected to mains (solar installations, equipment running on batteries due to mains failure) so that when voltage drops - overload will be detected and connected devices turned off. After a while when voltage level returns to usual operating value - connected equipment can be powered on again.


Note: PoE-Out controller firmware version 2.0 allows 1A load on port with 2.2A of total limit on all ports
Make sure you're not exceeding the 1A/port and 2.2A/total limits.
I see 200 to 400 milliamps on each of the 2 connected units, sometimes 500. So that should be ok.[/quote]


If you've got that super cheap "copper bonded aluminum" cabling you're riding somewhere around 22% loss over 10m.
I found that out in another instalation. :shock:
This also does not take into consideration the voltage drop over connectors or junctions. A crimp style ethernet cable has two junctions (pin to wire) and two connectors (pin to pin) because you've got one on each end of the ethernet cable. So you can probably take a bit more off for those. So I'd have to guess that anything between .8v and 1.4v drop over 10m could be considered "normal".
Hmm, OK. I'll get that. Strange I never had that issue before. Maybe because first I worked with one alraedy 'sick' board (Ether1 already didn't work anymore and ether2 started to play up.) To solve the power cycle problem I did install the rb750up which than actually gave me the new issue of low 'end of the line' voltage?
Depending on the site, I would try:
1. Running the power to the RB-750/RB-433's separately over something like 12AWG, that would cut your voltage losses by almost 80%.
I started to use the rb-750up to give myself the remote power control and a switch/bridge in the centre also gave me a bit more options for extra units or hooking up a monitor laptop.[/quote]
2. Putting some kind of RF noise suppression on the power pairs near the devices (ferrite beads) as I don't know if MikroTik would have put noise suppression on the power lines inside their devices and it may be stray RF noise that is causing some of your headaches.
Well, I'm not a elec. tecnician. Tell me more about this. I always asked myself why some device come with these heavy round things around the cable (or phone cable). Is this to suppress noise on the cables? Where is this noise coming from?
(or make sure your shielded cable is properly terminated, you are using shielded cable right? ;))
No, I am NOT using shielded cables. Never did and never have been aware of issues resulting. Most of the MT plastic outdoor units have hardly options to put an earth connection on (Groove). Or, when units are fit with the standard non-shielded watertight box connectors, what is the use of a shielded cable if the shielding stops at the connector? (I`m have been told by some elecs. that in case of shielding both ends need to be earthed. If one end is not, or earth is poor, the end situation is worse for data than using non-shieldd cable . In Spain, with its very, very dry soils in the summer, proper earth is not something to find easy. Also, my believe is that earthing high points also attracks lighting strikes. Something I'd really try to avoid.) I know the 'books' keep telling me to earth properly and than protect the device against lightning. But the books are not written with 'economics' in mind. Nowadays clients rates are so low a € 100,- instalation takes me almost a year to make some money. And the competition is still undercutting my rates.....

I always worked with cheap indoor cable and usually the cabling+antenna installation has a longer life span than the usual client stays in the books.... I have some clients that still run on the original indoor, unshielded, not earthed CPE antennas without any problems for almost 5 years now. We started using cheap copper coated alu cable now half a year ago for simple CPE installations and indoor building cable runs. If the cable is handled with ease during the install (dont 'kink' or fold them!) they give us no problems so far and for € 35,- /305mtrs this is realy economic pricing...[/quote]
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
WirelessRudy
Forum Guru
Forum Guru
Topic Author
Posts: 3089
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: Battery feeded PoE voltage drop over 10m utp normal?

Fri Apr 26, 2013 1:44 am

Thanks JackANSI. :D

Yes he has to buy another solar panel plus another battery - BUT afterwards he will encode NO MORE problems with his hardware due to power. :)

A 2.3Ah (2x12V batteries in series) can handle 4:30 hours before shutting down an RB750UP, an RB435G with a DBII N pro on it.
He has 300Ah :shock: battery (which probably will not be JUST one so he could "separate them" to get 24V out of them).
He could ALSO have more than one solar panels to charge them - so he could split them too.
Supposenaly that afterwards gets 150Ah total, with his config, he could be uptime, WITHOUT EVEN charge the batteries, for ~ 200hours thus 8 FULL DAYS - not bad...
No, wrong. Let me explain:
I have multiple panels (4 to be exact) and yes I have 4 batteries. So yes, I can make a 24V installation. But, before I had a windmill only producing 12V, with only 2x 105AH batteries and only 2x 50Watt panels. Hence the 12V setup. But in case of long spells of bad weather I still needed to run a generator with a battery charger. The latter is widely available and cheap in 12V versions where the 24V is suddenly much more expensive and usually to big to fit in the small space I have available.

AND, last but not least, with a full solar installation in the very long sunny summers here I end up having voltages on the wiring close to the max the routerboards can handle. I had some issues with another installation in the past coming from too high voltages....(Most boards have 28V as maximum that is allowed. But charging solarpanels can put this, or even higher, voltages on the battery and thus on the feed of the boards (the usual solar controller doesn't have a adjustable limiter. If they already have a separate 'consumer' voltage limiter than that is usually set somewhere around 28Volt. Several guys on this forum mentioned this accompanied by the question to MT why the boards could not be made to handle some extra volts to make it easier to work with 24V solar/wind controlled installations.)

The original 210AH combined battery set ran with 3 boards and 4 radio's in total and gave me a 'tested' life span of some 60 hours max.. The voltage of the batteries dropped then back to below 11volts (new gel batteries) and the controller would than shut down the consumer output. And then, it needed at least 12,4v input to release the consumer output again. That could mean my tower would be deprived of supply well into the next morning! Even when the sun was shining! (The panels output is initially fully consumed by the batteries before the voltage really started to run up.)
You have to remember that althoug a certain amount of AH compared to the consumption gives certain life span, when the sun's pops up again lots of the produced amps are consumed by the batteries. And most panels only deliver their max output only twice a year, during an hour a day max. under optimum circumstances. My panels only get the first sun somewhere around 10am anyway. The sun drops behind the mountains already at around 5pm in the winter. So, these panels have only 7 hours of sun and probably only 2-3 hours with real production.
Spain is usually very sunny, but during the winter, with its already short daytimes, we can sometimes have 4-5 days without any sun at all. Yet again, in winter batteries are only charged some 6 hours a day by the panels where the rest of the day they are beeing 'eaten' . So, 2 or 3 days without sun is taking a lot.
(And theoratical calculations are not always the same as the 'real thing'....)

The new extended (more panels, more batteries) 12v setup worked fine for some weeks until I ran into these new problems this week. So only now I found the voltages on the boards were really low.... So only now I am considering re-designing the setup into 24Volts. I'm planning to run camera's on site so by always having considerable users consuming combined with the now discovered powerdrop I should not have to worry too much against over voltage....

P.S.
For the interesting reader: My initial set of panels was facing some 20 degrees east of due south. So they did best during the morning ours. Usually the microclimate here in the mountains give afternoon clouds where the day usually starts sunny.
My second set of panels is now facing with 25 degrees west of due south. This gives me longer sun coverage in the afternoons (provided the sun shines) and there for consumers start 'eating' some 2 hours later from the batteries than before. And off course it boosts up the whole production. Around noon I now got twice the production as before. It made a hell of a difference, I'll tell you.
(Be aware that two separate controllers are needed. Each azimuth group needs its own and both are independingly connected to the batteries.)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
cevcrd
just joined
Posts: 1
Joined: Tue Jun 05, 2007 4:20 am

Re: Battery feeded PoE voltage drop over 10m utp normal?

Fri Apr 26, 2013 5:51 am

Here my 2 cents!

We have several 12V sites and for different reasons, out of our control, some sites are solar with 12V setup and others are setups with 12V inverter/charger system that we can't have access to the 120VAC, leaving us only with a direct connection to the 12VDC from the batteries.

The best efficient way is to power direct from the batteries, but as you clearly explains, the voltage in a 12V system with some UTP cable seems to be very fluctuating and leave too little voltage/amp margin for the equipments. When the traffic increase, the boards demand much more bursts of power and in these moments the boards lock-up.

In your case seems to be that you have a very good solar setup, so I would let that part as is and try to use DC-DC converters. In our case, we converted to 24VDC with a Tycon Power DC-DC converter, as the TP-DCDC-1224. This unit have a 75% of efficiency, so is not so good if you have limited batteries, but they're affordable and of good quality.

We started using these for the most power demanding routerboards and this converter can supply around 18W, so you can connect to a 750UP and power the RB435G also, that way you can reboot from the 750UP. We very heavily loaded APs we only use one TP-DCDC-1224 for each AP to be sure that the power is very stable and reduce problems almost to zero.

In our experience, the lookup of the RBs are very related to the quality of the power, in our case, we have very few late night visits to our cells for lookup reasons after the 24V upgrade. If the site is too far or difficult to visit, we put another 750UP if necesary for reboot, because a trip is much more expensive for us.

This units work very good for us and the stability of the RBs are excellent since they have a very stable power with very little drop in voltage with cables up to 30M (in our case). And to be honest, in some cases we have desisted of using the 750UP.


CEVC
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Fri Apr 26, 2013 9:35 am

Ohhhh.........shift!....... :(

Your solar system is NOT WELL designed - who did that?
To be exact.
There are THREE "cheep" - could say - solar panel selling widelly around Europe.
Chineese (China) and are cheeper - they could provide you on a fully sunny European summer day ~80% of their nominal output power.
Siemens (Germany which are made in China too) - they could provide you on a fully sunny European summer day ~92% of their nominal output power.
Sanyo (Japanese) - they could provide you on a fully sunny European summer day ~99.5% of their nominal output power. - that's why Sanyo are more expensive.

Now

Let's suppose that you NEED ~ 50Watts from your setup (thus ~1000Watts per day for 20 hours without proper sun) and ~ 1800 - 2200 Watts per day to charge your battery setup. - or ~1200 Watts for your setup.


WITH the SANYO solar panel

You NEED on summer days with the SANYO solar panel ~ the total drawing wattage x 1.2 = 1200 x 1.2 = 1440 WATTAGE to provide from the solar panels PER DAY to "fully charge" them.
You NEED on WINTER days (were sunny days are less PLUS the sun has more distance from earth) with SANYO solar panel - ~ the total wattage x 2.5 = 1200 x 2.5 = 3000 WATTAGE to provide from the solar panels PER DAY to "fully charge" them.
The above numbers are for Greece which has sun - we are almost in the same situation with you in Spain.

So if you say that you have ONLY 4 x 50Watts solar panels, they are NOT enough to charge the batteries in summer and you want to do it in winter too?

For our setup (ONLY to backup systems in case of AC loss for a couple of hours) we use 2 x 50 Watts SANYO solar panels just to charge 2.3A/h (internal system batteries + 7A/h (external batteries) thus ~ 10A/h total and you use 200Watts total wattage solar panels to charge ~ 300A/h batteries?

As you say, in winter you will have a major problems there plus they are NOT even enough to charge the bateries in summer times. :?

PLUS that if you DON'T have any AC voltage up there to charge the batteries and you are ONLY with the solar panels to do so, you'd better put up there an automatation for the panels to "follow" the sun (electronic/mecanical with rotors).

P.S. with ONLY the internal batteries of our system thus 2.3A/h and ~ 15Watts drawing wattage, we see each battery to go down to 12V (from 13.8V) AFTER 3 hours.

You just DON'T CHARGE your batteries. :(

EDIT: Your solar charging system IS NOT WELL designed from the suplier that you are buying it.


.
 
n21roadie
Forum Guru
Forum Guru
Posts: 1895
Joined: Fri Aug 07, 2009 10:36 pm
Location: Limerick,Ireland

Re: Battery feeded PoE voltage drop over 10m utp normal?

Fri Apr 26, 2013 2:15 pm

....................... But, before I had a windmill only producing 12V, with only 2x 105AH batteries and only 2x 50Watt panels. Hence the 12V setup. But in case of long spells of bad weather I still needed to run a generator with a battery charger. The latter is widely available and cheap in 12V versions where the 24V is suddenly much more expensive and usually to big to fit in the small space I have available...............
I have been considering designing a “Hybrid” system and so far here are my conclusions:
(1) Use 12V Solar – 24V system charging voltage is very close to RB cut out voltage and have read posters complaining about this and why state – 8-30v but cut-off at 28.3?
(2) DC Voltage cable loss – Use 7core trailer cable - Wire Size: 7 x 9/0.30mm (0.65mm sq) from battery bank up close to devices and then from 7core use 3 x pairs which will power 3 devices plus one spare core?
(3) Solar panel(s) for maximum efficiency solar tracking needs to be used and design of such to work effectively and survive storms at a exposed location could be a big task,
(4) Solar panels in remote locations have a risk of been stolen by thieves and there is a risk of damage due to debris colliding with the panel(s) during a storm and replacement cost(s) may have to be budgeted during design stage,
(5) Use vertical axis wind turbine instead of the more popular horizontal axis designs and while less efficient they have a much better chance of survival of turbulent windy conditions,
(6) To reduce or eliminate journeys to a site just to re-cycle power due to lockup, install a GSM relay for this or network relay
…………………………………………………………………
N21roadie,
Network 100% MT for Now?
 
WirelessRudy
Forum Guru
Forum Guru
Topic Author
Posts: 3089
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: Battery feeded PoE voltage drop over 10m utp normal?

Sat Apr 27, 2013 1:58 am

Ohhhh.........shift!....... :(

Your solar system is NOT WELL designed - who did that?
To be exact.
There are THREE "cheep" - could say - solar panel selling widelly around Europe.
Chineese (China) and are cheeper - they could provide you on a fully sunny European summer day ~80% of their nominal output power.
Siemens (Germany which are made in China too) - they could provide you on a fully sunny European summer day ~92% of their nominal output power.
Sanyo (Japanese) - they could provide you on a fully sunny European summer day ~99.5% of their nominal output power. - that's why Sanyo are more expensive.

Now

Let's suppose that you NEED ~ 50Watts from your setup (thus ~1000Watts per day for 20 hours without proper sun) and ~ 1800 - 2200 Watts per day to charge your battery setup. - or ~1200 Watts for your setup.


WITH the SANYO solar panel

You NEED on summer days with the SANYO solar panel ~ the total drawing wattage x 1.2 = 1200 x 1.2 = 1440 WATTAGE to provide from the solar panels PER DAY to "fully charge" them.
You NEED on WINTER days (were sunny days are less PLUS the sun has more distance from earth) with SANYO solar panel - ~ the total wattage x 2.5 = 1200 x 2.5 = 3000 WATTAGE to provide from the solar panels PER DAY to "fully charge" them.
The above numbers are for Greece which has sun - we are almost in the same situation with you in Spain.

So if you say that you have ONLY 4 x 50Watts solar panels, they are NOT enough to charge the batteries in summer and you want to do it in winter too?

For our setup (ONLY to backup systems in case of AC loss for a couple of hours) we use 2 x 50 Watts SANYO solar panels just to charge 2.3A/h (internal system batteries + 7A/h (external batteries) thus ~ 10A/h total and you use 200Watts total wattage solar panels to charge ~ 300A/h batteries?

As you say, in winter you will have a major problems there plus they are NOT even enough to charge the bateries in summer times. :?

PLUS that if you DON'T have any AC voltage up there to charge the batteries and you are ONLY with the solar panels to do so, you'd better put up there an automatation for the panels to "follow" the sun (electronic/mecanical with rotors).

P.S. with ONLY the internal batteries of our system thus 2.3A/h and ~ 15Watts drawing wattage, we see each battery to go down to 12V (from 13.8V) AFTER 3 hours.

You just DON'T CHARGE your batteries. :(

EDIT: Your solar charging system IS NOT WELL designed from the suplier that you are buying it.


.
well, done it all by myself and it worked fine the last 2.5years. Only when we have 2 consequtive overcast days and posibly more to come I run a generator to boost the batteries up. In these 2.5 years I had to run the generator in total maybe 7 days.... The rest of the time I never had any issues. According the solar charger the batteries are normally full and only after some bad weather days I see a real drop in charge level.

Theoratical calculations are not always the same as the real world. When calculate a system you need to go from a lot of presumptions so you try to stay on the safe side and build margins in at all ends.. in reality it hardly ever works out that bad......

I started this topic because now I only found one board not beeing very stable at the low voltages it is presented. But its now up for 3 nights again and although I see 10,6V on the board over night it doesn't drop anymore... Still considering upgrading to 24V but am a bit worried about the over voltage issue that now might start troubling me on long sunny summer days...
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
WirelessRudy
Forum Guru
Forum Guru
Topic Author
Posts: 3089
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: Battery feeded PoE voltage drop over 10m utp normal?

Sat Apr 27, 2013 2:09 am

(3) Solar panel(s) for maximum efficiency solar tracking needs to be used and design of such to work effectively and survive storms at a exposed location could be a big task,
I think it is cheaper and more robust to buy more (small) panels. We have regurlarly severe gusts so a tracker would have to be very sturdy and thus expensive....
(4) Solar panels in remote locations have a risk of been stolen by thieves and there is a risk of damage due to debris colliding with the panel(s) during a storm and replacement cost(s) may have to be budgeted during design stage,
I had some steel frames made and the panels are enclosed by them in the workshop (welded a frame around). After that the frames are welded at least 3 meters high on my tower, which is fenced and the site has no power. So to steel them you need battery operated disk or torch on a ladder at hight after you climbed over a fence. I don't consider it to be a big risk. But nevertheless, security camera's will be installed soon...
(5) Use vertical axis wind turbine instead of the more popular horizontal axis designs and while less efficient they have a much better chance of survival of turbulent windy conditions,
The windmill died indeed.... :(
(6) To reduce or eliminate journeys to a site just to re-cycle power due to lockup, install a GSM relay for this or network relay
That's going to be my next project. Just ordered 6 G3 modems. I only need to find out how to do this. Probably a new topic coming up... :)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
JackANSI
Frequent Visitor
Frequent Visitor
Posts: 71
Joined: Wed Apr 03, 2013 6:52 pm

Re: Battery feeded PoE voltage drop over 10m utp normal?

Sat Apr 27, 2013 8:15 am

2. Putting some kind of RF noise suppression on the power pairs near the devices (ferrite beads) as I don't know if MikroTik would have put noise suppression on the power lines inside their devices and it may be stray RF noise that is causing some of your headaches.
Well, I'm not a elec. tecnician. Tell me more about this. I always asked myself why some device come with these heavy round things around the cable (or phone cable). Is this to suppress noise on the cables? Where is this noise coming from?

You only want DC on your power pairs (blue and brown), anything else is something you don't want that could cause problems.

Ethernet devices usually have a transformer on each end between the connector and the line driver. Sometimes its a black boxy looking thing within a couple cm of the jack, other times its built directly into the jack itself to save room and simplify construction. This device, more often than not, removes common-mode noise (like the interference I'm speaking of) on the orange and green pairs only.

To put it simply, any length of wire has the potential to be an antenna. Adding ferrite beads adds impedance to high frequency signals.. forming a kind of low-pass filter. These high-frequency signals come from just about anything that isn't completely DC. Other transmitters for AM, FM, TV, other communication hardware like amateur radio, cell towers, terrestrial links, and satellite links, etc..

This noise can get into digital devices and cause errors anywhere along the process of reading and transferring bits of data. These errors can cause anything from a re-transmit to a re-boot (or freeze).

Being that you're running off a comparatively noise-free power source, its probably safe to assume its not a bad/noisy power supply. So if noise is an issue it would be coming in through the ethernet cabling. But still, noise is a stretch in your situation. I only mention it to be thorough. Getting adequate power to your devices is always the first step.



According to MikroTik's documentation:
Port eligibility detection for poe out
auto-on mode is considered safe, it will check if resistance on the port is within range and only then enable PoE out on the interface. The range is 3kΩ to 26.5kΩ
So any relay or relay driver circuit that presents a resistance in that range could possibly fool the MikroTik into turning the relay on, then the relay could apply power to your devices, skipping the voltage drop in the RB750. If I had a RB750UP I'd play around with it and put something solid together that I could tell you how to build, but I can't so you'll have to come up with that part...



You could try to replace the power handling duties of the RB750 with this little gem for your power cycling needs. Direct relay switching would give you over .5v extra that you don't have with the RB750. Granted you lose a little in the way not having a one-stop shop to check on data and power, but getting more power to your 433's ~might~ solve your problem. I have one of the 802.3af versions for power cycling a few IP cameras we have on site and it works great. But relays draw a fair amount of power when on, so you'll burn through your batteries faster, keep that in mind.
 
hipro5
Member Candidate
Member Candidate
Posts: 161
Joined: Sat Feb 09, 2013 11:48 am
Location: Hellas (Greece)

Re: Battery feeded PoE voltage drop over 10m utp normal?

Sat Apr 27, 2013 9:46 am

My calculations were done based that your systems are drawimg their MAX wattage in case of full traffic, which actually never hapens. ;)
But as I think and work: "Better be prepaired that down" ... :)

Who is online

Users browsing this forum: No registered users and 74 guests