I am fiddling around with a few solar panels on my shed and eventually I would like to log what they are doing and maybe even monitor them over wifi.
But first I am messing arround with an arduino mini (clone) as a watt meter and this works. ( you have start simple right?) I am using an LM7805 to power my watt meter by borrowing power from the power source. (the solar panels).
The lm7805 can handle up to 35 volts which is enough for now because my panels do max 30 volts. But I would like to beable to put 2 of them in series.
For now I got the meter working with a voltage diver and an ac712 hall effect breakout board.
But when I connected my real watt/amp meter that I bought to compare results I noticed that my watt meter consumers almost 2.7 watts! That is rediculous. I only got an lcd1602A with the i2c backpack , a pro mini and the ac712. Those should not use that much. Than I noticed the lm7805 gets warm. And after googling it turns out it just dissipates all excessive voltage it does not need into heat.
I am charging some batteries with those panels and they are big enough that I can spare some power but almost 3 watts 24h a day is way to much. That is way more than my 3 watt led light out side I just turn on for like 10 hours after sun set. I could power 2 more 3 watt lights for 10 hours. What I mean to say that I do not need to go ultra low power and remove power leds and other stuff go down to 5ma or something.
So I have been searching for other voltage regulators but I seem stuck. The max they can handle seems to be 35 volts and most are less.
I did order the lp2950 5 volt version (but still waiting on them) which can handle max 30 volts which seems to close to the max 30 volts the panels can do. The max current is 100ma. Could I put 2 of those regulators in parallel so they can do 200ah?
What other regulators are there that are cheap, dont use / waste too much energy them selfse and can work with about 60 volts? I have several switching buck converters but they don't go that high either and I thought they would waste relative a lot of energy but it seems I am wrong here?
It must be possible. I am using those cheap blue plastic watt meters, 60 volt 100amps from ali x and they also have the lcd1602 inside but a custom pcb. This takes the power from source you are measuring just fine and it hardly uses any power it self. So little my arduino meter could not detect any amps being used.
Yeah, you look at your current and how much your 7805 drops in voltage and the watts (heat) gets there real fast. I would be looking for a "buck converter" something along these lines:
Try a (mains power) 5volt cellphone charger.
They are rated for ~85voltAC to ~240volt AC, but I found that some do work ok if you connect >48volt DC to the AC prongs.
Leo..
You have got something odd in your concept.
You should not power your ProMini from the solar panel, but from the battery after the solar charger.
If I where you I would use an INA 219/INA226 highside Ampermeter, it can handle as much 32Volt bias, but you can place the sensor between the two panels.
If they are identical and identically illuminated, you just can double the power.
You will find plenty of ready-made sketches for that device.
I will start that week with a similar project and will use an ESP8266 to MQTT the measurement to thinger.io, so I not only get the measure but the full trends for the battery level, the solar panel voltage and the charging current.
I was also thinking about a regular 5V power supply. Indeed designed for AC, so it surprised me when I heard this first as well, but thinking about it and understanding a bit more about how those things work it actually makes sense that it does work (the first step is usually rectification of the high voltage into DC...)
I have the ina219 which can only handle 3 amps max and I have an ACS758LCB-100B breakout board which seemed a bit over kill.
I have 4 panels of 120 watts that are 17 years old. 2 in good shape 2 got moisture behind the glass and perform less. I did manage to get 120 watts (together in parallel) on a good sunny say in oktober. Maybe they can go a bit higher in summer time so to be safe I like it to beable to handle at least 15 amps.
Anyway I have 2 meters like these from aliexpres:
I have them connected all day. They can handle 60volts and 100 amps max.
Yesterday I said I could not measure it's usage but I think is drawing about 30 milli amps.
The one that is on the battery side stays powered at night or course but the one on the solar panel side turns of. It has a connector for an external power source but when I connect the battery to power the watt meter my solar charge controler does not see the sun go down any more an never turns on my load. I believe this is because of the common ground issues. I found a solution for that using those little B1205S isolators.
But that is a bit off topic. What I wanted to do is to copy is how those blue watt meters work.
They draw their power from the power source you are measuring from anywhere between aboout 7 volts to 60 volts and they do not use a lot of power. The display is the lcd1602 which is very common in arduino world. so I thought I could do this with the mini or maybe a nano easily. The only big problem is how to get the power supply working as those watt meters.
macron0:
I have the ina219 which can only handle 3 amps max and I have an ACS758LCB-100B breakout board which seemed a bit over kill.
I have 4 panels of 120 watts that are 17 years old. 2 in good shape 2 got moisture behind the glass and perform less. I did manage to get 120 watts (together in parallel) on a good sunny say in oktober.
You did not do tell anything about your converter and the battery (power/voltage) you are using.
I still consider that using INA226 bread boards is the best solution. The current intensity they can handle depends on the shunt, just use a 0,01Ω shunt.
Additionally if you have 4 solar panels, you can monitor each of them individually.
With an ESP 8266 you will not need any display and get much more graphical information over thinger.io.
Anyhow finding switching power supplies and solar converters that perform well at low power is difficult.
Specially in Europa we sometimes full months of cloudy winter times it is difficult to keep the battery charged for 24/365 off-grid operation.
Most MQTT converters are requiring at least a couple of watts for themselves and will drain the battery during the night. Idem for switching regulators, finding one that is efficient at low-power is difficult, you will mainly not get the information and when you have found one you never can be sure to get a sustainable provider.
RIN67630:
I will start that week with a similar project and will use an ESP8266 to MQTT the measurement to thinger.io, so I not only get the measure but the full trends for the battery level, the solar panel voltage and the charging current.
RIN67630:
You did not do tell anything about your converter and the battery (power/voltage) you are using.
I still consider that using INA226 bread boards is the best solution. The current intensity they can handle depends on the shunt, just use a 0,01Ω shunt.
Additionally if you have 4 solar panels, you can monitor each of them individually.
With an ESP 8266 you will not need any display and get much more graphical information over thinger.io.
Hi and thank you for you detailed answer. After a full sunday day of trying and fooling around an magic smoke and updated to version 2 which did not have magic smoke but the lm7805 got pretty warm and the added ADS 1115 stopped reading values. So back to the bread board again. I dedcided to check the forum and got an other answer about ina's So I was searching for a short time bit well duh Try to find how much current an ina 226 can handle! it is not in teh data sheet that I can find (in a shot time) so I was mildly irritated and read you message again and read the part if the shunts hehe . So thank you ! I order 2 rediculiously large shunts from ali ( they looked small on the picture but are like 15 CM ) to try and build my own current sensor. But I think I can try this with the ina 226! So gona order some of those. thank you !
RIN67630:
Anyhow finding switching power supplies and solar converters that perform well at low power is difficult.
Specially in Europa we sometimes full months of cloudy winter times it is difficult to keep the battery charged for 24/365 off-grid operation.
Yes I noticed. In a lot of december days the 4 together only managed about 8 watts!! Was really shocked. I thought bad old blue panels. Than I asked my dad who got 10 new panels last summer of 2750 peak watts i a community project what he was doing and he said it was about 150 watts at noon. So I guess its normal tohave less than 10% in winter.
RIN67630:
Most MQTT converters are requiring at least a couple of watts for themselves and will drain the battery during the night. Idem for switching regulators, finding one that is efficient at low-power is difficult, you will mainly not get the information and when you have found one you never can be sure to get a sustainable provider.
I am charging lithium ion batterys (3S 2p 18650 and 3s 1p 6000mah a piece) and I have a Lifepo4 pack of 288 18650 batteries I would like to charge. Those cannot be charged when it is freezing.
So I am allready thinking of powering the arduino's with a lead acid battery and indeed have them disable all outputs when it is freezing. If you do max 10 watts during 7 hours of sun light youre out of options.
But than I have mains in the shed so I could make someting tofallback to mains if needed.
But I cannot even clone a stupid WATT meter that you can buy on ali for 9 euro's with shipping.
About the ESP8266. I am not going to use online services for that. I really want every thing work in a local network. I have my own home server running sinds 1996 anyways. I allready have some tests posting dht11 data. And I can run a web server in the esp8266 so I can control it from my webpages.
Like my chineese ip cams. I configured them without a gateways so they cannot go online. ffmpeg records what they see and uploads it to a server only I have control over.
I am at the point I can make and esp-01 a slave for an arduino via software serial and without the AT firmware. I tried to make an esp-01 and i2c slave but of course that is not possible.
But again I will check out the INA226 and custom shunts. that sounds like a good option. But they still only do max 36 volts.
VDISSIPATED as HEAT = VOUTSOLAR PANEL
-VINLM7805 = 23V
Let I5V = 0.08A (80 mA) (ARBITRARY GUESS !)
Let V5v = 5V
P5V = 0.4W
P = I*V = 0.080A * 5V = 0.4W
PDISSIPATED as HEAT =(2.7W-0.4W) = 2.3W
Let VDISSIPATED as HEAT = 23V
Let PDISSIPATED as HEAT = 2.4W
I DISSIPATED as HEAT= PDISSIPATED as HEAT/VDISSIPATED as HEAT = 2.4/23= 0.104 A (104 mA)
In other words, if all of your devices draw 80 mA, then they only account for 0.4W and
the other 2.3W is dissipated as heat because the input to the regulator is so high.
AFAIK
The buck converter linked early on looks perfect for your application.
Conclusion is correct (use a buck converter) but that calculation not: where does that 7V come about?? That 23 V is dropped somewhere... Not in the 7805.
What part of the input voltage do you think is excessive and what part do you need ?
Are you going to use 0v as the needed voltage ?
1V ?
2V?
You can't use 5V because it's not an LDO regulator so if you use the datasheet , anyway you cut
the cake you wind up with 7V as the correct input voltage and everything else is dissipated as heat.
Then that 23V is dropped somewhere that's NOT in the 7805... and the 7805 gets to drop just 2V, dissipating just 160 mW at 80 mA current.
Really ?
If you are putting 30V into the regulator and it's getting hot, I think it's fair to say it is being dropped in
the regulator.
That's why I said it is being dropped as heat. Do you know WHERE in the regulator it is being dissipated ?
I don't . I just know it's somewhere inside the regulator and that's why it's getting hot.
You can't use 2V because then you're saying all the power used by the OP's devices is FREE !
It all has to add up at the bottom line: (total power dissipated and measured by the OP, (ie: 2.7W))
The output is 5V , so the load draws 0.08A @ 5V = 400mW.(0.400A)
Originally you said there was only the 2V (160mW)
If you said the 2V results in 160mW (0.160A), then you are also saying it is the same 80mA drawn
by the load.
I would agree with the regulator overhead = 160mW and the load = 400mW and the excess power (due to the 23V) = 2.7W-0.160W-0.4W= 2.14W
raschemmel:
Really ?
If you are putting 30V into the regulator and it's getting hot, I think it's fair to say it is being dropped in
the regulator.
Now you're contradicting yourself.
You seem to imply that you would use a 7805 to drop 30V to 7V so you have a nice voltage to feed into the 7805.
Now you're suggesting to supply 30V to the 7805, which would give you a 25V drop, as the output of a 7805 remains of course 5V. Still no 7V in sight.
That indeed might work as the part is specified to work up to 35V. You'd just need a decent heat sink, as even at a modest 80 mA draw that results in 80 * 25 = 2,000 mW of dissipation (5/6 or 83% of the input is lost as heat).
Hi,
Can you draw a basic diagram showing how your system is connected, including batteries and charge controller please?
Also show where you want to take measurements.
Note, Most charge controller switch the PV panels out in the negative line, not positive as everybody assumes.
Have you got Blocking Diodes and bypass diodes if you are configuring series/parallel configuration?
