I have need to set up a bridge in a location visible to myself and a secondary source. No AC power is available. Where would I go to buy 'Solar power systems' with batteries and the like for a WRT54GS?
What price should I expect to pay? I'm trying to find out if it's financially viable. If so, I'll be buying and installing in the next two weeks.
You can buy solar panels and batteries from loads of online retailers. You just need to consider what current is drawn over the hours of darkness - then get a battery big enough. Also get a solar panel big enough to give the required charge and actually run the router during the day. Then work out your figures on whatever you consider it to be financialy viable against.
What does it draw at 13.8 volts? IIRC my WAP54G is just a few watts. Depending on where you are installing it, and your insolation available, the calculation is pretty simple:
Find the power consumption, call it 5 watts just to have something to talk about (confirm this before buying anything).
5 watts times 24 hours per day is 120 watt-hours.
Solar panels are rated at peak output at a nominal one sun input (1000W/M^2), and most places in the US get between 4 and 5 "peak sun hours" per day.
Divide 120 watt-hours by 4 to get 30 watts of solar panel. You want to upsize a bit to cover recharging after extended outages, so get at least 40 watts.
has 40-watt panels for $285, which probably isn't awful.
How many days of battery backup do you want? Since you are only using
10 amp-hours per day, and you don't want to use more than half of your battery capacity, you need between 50 and 100 amp-hours.
has a UB27 for $125 that looks OK.
You also need a 5 amp or better charge controller, that's probably an ASC 12/4 for $44. Check to be sure it can be set up for charging gell cells.
If you can get to the site and top off the battery water occasionally, you could save some money by using a pair of golf cart batteries for $100 and no charge regulator. HydroCaps can keep the need for battery maintenance down in this situation, but they add $50 to your initial outlay.
Call it $500 plus or minus (do you need a housing for your WRT? External antennas? Do you need cabinets or boxes to put all this stuff in? Is it subject to very high or low temperatures? Vandalism? Is there anything else you haven't told us?)
Don't forget $50-$75 for the WRT, plus whatever you are using as endpoints...
You can probably set it up for around $350-400, depending on how you're planning to mount the panel and the battery. Is it an outdoor location?
That unit draws around 6 watts, so figure around 50 watts to account for the panel not being aimed at the sun all the time, plus being able to both operate the unit and charge the battery at the same time.
" Buy a sealed lead acid battery and a controller
" A 12 volt/12 amp hour battery (144W) would be good, which will give you some buffer for cloudy days.
Too bad you can't buy a surplus roadside call box somewhere.
I couldn't find a manual on-line for this device, but I have run a Dlink DWL-900AP+ wireless on batteries. The factory supplied power supply is a 5V switch-mode wallwart. For simplicity, I built my portable device to run from 6 C cell nicads, then used a low drop-out voltage regulator that I built using a Texas Instruments chip. [If relevant, I can dig up the number.]
If you are going to use an off the shelf 12V solar system, you will need to build a DC/DC converter to reach 5V, most likely a buck configuration. If you used a LDO linear regulator from a 12V supply, more than half the power would be used in just dropping the voltage.
It would be more work on the solar cell side of the task, but I would set up a 6V system using a gell cell, then build a 5V regulator using a low drop out chip. I specify low drop out in that these devices use PMOS pass fets rather than a PNP, so they are very efficient, not to mention more stable. [The PNP in those cheap bipolar regulators sits near saturation, which leads to much power wasted in the base driver current, plus difficult circuitry to keep the PNP from entering saturation. It's ugly.] The skill required to build a switcher is an order of magnitude higher, both in circuit design and in layout (construction). The low drop out linear regulator is just the chip and a few capacitors. You would need basic soldering skills, but not much beyond that.
I know very little about solar cells themselves. My guess here is you would assemble enough to float the gell cell, and use a Schokty diode to prevent back flow when light is no longer shining on the cells.
Your first step would be to determ> I have need to set up a bridge in a location visible to myself and a
Now, I know you hang out in the homepower group, William, so you should realize those numbers don't compute.
If you want a 50amp-hour battery, you really need to be able to get a 10amp charge into it. If you're charging a 50-100amp-hour battery with a 5amp charge controller, you run a risk of never really getting a full charge into your battery, and shortening its life.
Yeah, but the tradeoff is spending another $1000 to provide the ideal charge environment for a $100 battery. Hence my suggestion to get golf cart batteries and no charge controller, which seems to work well for the Coast Guard.
FWIW, I looked up the power consumption data on the WAP54G I took a while back, and it was just under 3.2 watts from 4.2V to almost 20V. I'll characterize a WAP54G someday, but don't hold your breath.
Yeah, the OP needs to better characterize his load, his insolation, his maintenance needs, and other things, but the concensus of $500 "plus or minus 3 dB" seems to get him in the right ballpark.
The WRT54G and GS uses a switching regulator down to 3.3VDC. It will run on anything from about 4VDC to about 18VDC. The regulator will probably handle a higher voltage, but I'm worried about the input filter cazapitor.
The current drain is: Volts Amps 15 0.25 12 0.3 5 0.8 4 1.0 for about 4 watts dissipation. The actual current drain varies somewhat with hardware version. This is for my V3.0 WRT54G.
The WRT54G/GS will run quite nicely directly from a 12V gel cell. For charging, a small solar controller (i.e. Morningstar) will do just fine.
The problem is that it burns about 4 watts continuously. To run for
24 hours without a charge (required for mid-winter operation), and keep the battery at 75% charge, the 12v battery would need to be rated at 30 A-hrs min. I threw together a spreadsheet for the purpose. It was made for commercial and ham radio repeaters, but should work for anything:
to recharge a 30 A-hr gel cell in one day or less requires a 400watt solar array (non-tracking). The WRT54G will burn
94 watt-hrs/day. The solar array would need to equal that in order to prevent the battery from running down. At 25 watt solar array would do the job. oh-oh, I think I just found a mistake in the spreadsheet.
Some photos of a solar powered Wi-Fi box (not mine).
Thanks for the technical data. I think the input capacitor won't be an issue given the amount of money you will be spending on the rest of the system. Throw enough money at the input cap and the problen will be solved. I've built pulse generators for LDO testing that needed very stiff filtering, and if money isn't an issue, you can get the capacitance quality you need.
The input cap is rated way above the 13.8 or so volts you need for a decent solar system, so that's not a concern. As Jeff pointed out, the Linksys boxes have a wide-input-range power supply that draws a fairly constant power.
Lower than four watts? Depending on what the OP needs/wants (we haven't heard back from him), he could use a WAP54G, which is 25% less power consumption, but the numbers work out about the same.
Now that I think about it, wiring (or buying) the panel for 6V operation would let you save $50 by only buying one golf-cart battery (and double the charge current, which is better for the battery), and only require three HydroCaps, which saves another $20, but you are still going to need something around 40 watts of panel and that's pretty expensive.
50 watts of solar sells, in a 12 volt configuration, won't be putting out more than 5 amps. A sealed lead acid battery should be charged at a rate of not higher than C/10 (capacity divided by 10), which would be a
5 amp charge rate.
The input cazapitor is inside the WRT54G. It's a chip tantalum and I'm unable to determine the rating. Common voltages are 16V and 25V. I haven't seen one blow up yet, so I'll guess it's 25V.
Probably, but let's do a bit of thinking first. This is suppose to be an "isolated" location for one end of a bridge. My crystal ball does not envision someone sitting next to this device with a laptop. My crystal ball tells me that this will probably be used as a WDS repeater, or one end of a data collection system. If it's a WDS repeater, then including the router section in the power drain is required. If it's one end of a data collector, then the router is superfulous. Therefore, a wireless access point would do for the data collector application.
Just for fun, let's say someone built a solar powered WDS bridge based on a laptop MiniPCI card and a PIC or Rabbit controller. It's doable. The controller draws next to nothing compared to the wi-fi card.
Ubiquiti Networks SuperRange5 802.11a 400mW Receive = 350ma (1.16 watts) Xmit = 700 -> 1400ma (2.3 -> 4.6 watts) Ouch. Ok, that's for a high power 5.6GHz card.
Senao EL-2511MP+ 802.11b only card. Receive = 180ma (0.6 watts) Xmit = 280ma (0.93 watts) That's much better.
BCM94306MP - Broadcom IEEE 802.11b/g Wireless Internal Mini-PCI Card Receive = 350ma (1.15 watts) Xmit = 550ma max (1.8 watts) Sorta in the middle.
Therefore, if you use a miniPCI radio, and a very simple controller, methinks it could be done with about 2 watts average power consumption or half of what the WRT54G consumes.
I'm not sure it's worth the effort. It would be easy enough to disable the ethernet switch part of the WRT54G. The WAP54G might be a better choice for a simple bridge radio. The front panel LED's on the WRT54G could be switched off. The processor clock on the router part could be slowed down without affecting the wireless section.
There are companies that make low power wi-fi chipsets that target PDA's and cell phones:
these chips suck only microwatts in standby, the burn the usual
0.5 to 1.0 watts when active. In effect, they are very low standby power, not very low operational power.
There are also chipsets that target Zigbee sensornets that could be used for Wi-Fi.
I blundered across this white paper from TI which explains why 802.11g consumes less power than 802.11b (because the xmitter is powered on for less time):
Don't run your solar powered repeater at low data rates.
I couldn't find anything with the features of a WRT54G that was specifically designed for solar operation.
Article on building a solar power wi-fi data logger:
uses a Compact Flash Wi-Fi card. Receive = 200ma (0.66 watts) Xmit = 300ma (1.0 watt) A bit better than miniPCI but not much.
You don't want liquid electrolyte batteries inside a sealed outdoor box. The electrolyte will outgas and rot the circuitry. Hydrocaps are a great way of preserving water but still leak corrosive gases.
have to clean these at least once per year, despite the vented box. Some idiot (me) installed the main fuse and disconnect inside the box, which corroded away in about 5 years.
Much better are gel cell batteries. The vent less gases. They are not totally sealed and will vent if overcharged. If you want a fun test, put a gel cell inside a plastic bag, and rapid charge it for a while. It's a gas.
However, gel cells are losing favor and are being replaced by AGM (absorbed glass matt) batteries. They're sealed and internally recover all the vented gasses. Internal resistance is VERY low. Therefore they dont get hot. Self-discharge rate is zilch at perhaps
2% per month. Prices are coming down.
Another not so minor problem is ambient temperature. I help maintain a few mountain top weather stations and a generator. It doesn't get very cold in California, but it's still cold enough for the capacity of the battery to approach useless. My favorite screwup was to run a circuit that requires 10.0 volts on a 12V gel cell battery. When it got cold, the terminal voltage drifted down towards 10VDC under load (due to increased internal series resistance). There's probably enough headroom to run the WRT54G LDO regulator on 6VDC and still have enough to supply the minimum 4.0VDC, but I still would prefer 12V battery just to be sure. Also, 6VDC solar cells are not too common.
As for golf cart batteries, our local ham radio repeater is in the middle of a golf course. Golf carts everywhere. If you want some of the worlds deadest golf cart batteries, I can make you a deal. When the golf course is done with them, they're genuinely dead or close to dead. Using an automotive battery for stationary battery applications is a very bad idea. I'll spare you the techy details (unless somone really wants them) but the construction of vehicle batteries are quite different from stationary cells.
Drivel: One of my quaintances has a combination solar and wind power Wi-Fi, cordless phone, and cell phone repeater located on his ranch. The theory is that when the sun isn't shining, the wind is blowing. Rather than buy a commercial wind charger, he built his own. A small permanent magnet generator with a 3 blade flexible fiberglass fan for feathering. The wind really blew for a few days last month, causing the blades to almost lay flat against the sides of the alternator. That eventually sheared off all the wires and bashed in the contoller box. Sorry, no photos.
Generally good advice (although I disagree with some of the minor details):
Yes, if it was actually generating 25 watts during the day. The 25W ratings is based on the panel being in direct sunlight. You really want a 50W panel to be safe. These cost about $250.
-50 watt panel
-55AH battery (fully charged when put into operation)
Depending on the time of year (longer, sunnier days, versus shorter cloudier days), even this may not be enough.
It makes you really appreciate being on the grid, when you look at how much it costs just to provide a continuous 5 watts of power!
I wonder what they're doing with all the solar powered call boxes that are being removed from the roads (at least in California they are increasing the distance between the call boxes due to the significant reduction in usage as a result of most people having cell phones).
The battery box at the base of this tower holds a couple of golf cart batteries:
It was not only cheap, but turns out to be pretty corrosion-resistant, and except for the frogs, is holding up pretty well. 8*)
Initially it had a single 20W panel and just ran the VHF repeater and the weather station with no problems. Adding the fixed-cell phone put it into a negative power budget, so I added another 20W panel, and it seems to be holding up, though it's due to be replaced with an AGM battery, float charger, and grid-tied solar power system.
It did toast the first pair of off-brand golf cart batteries after a couple of years, I've got HydroCaps on the second set, but they aren't going to last long enough to see if that helps. 8*)