I have an off-grid cabin in south-western Wisconsin. I have 3150 watts of solar, a classic 150, a Magnum 4448pae inverter, and a 390ah 48v battery bank. All of that is on/in a shed that I purpose built for it, so that the cabin wiring is completely conventional.
I only occupy the cabin for two or three weekends a month at most, so there is a lot of excess solar. Today was a sunny day and the charger went into absorb at 8:15am and into float at 10:45am. In the winter, we plan to heat to ~45f so that we don’t have to winterize every time we leave. The heater will be a direct-vent propane unit that doesn’t require electricity.
I would like to use my excess solar to run a 1500w oil filled heater in the cabin. I would do this to reduce my propane usage. When I’m at the cabin, I would turn off the heater so that I would have the full capacity of the inverter (4400w continuous).
My plan:
- Heater is 1500w. I don’t have it yet, but I’ve read that it actually uses 1250w. My inverter is 93% efficient at this load, so I’ll probably pull about 1344w, and another 50 for other electronics in the cabin.
- Use Aux1 waste-not high mode
- 12v signal would go through two unused pair of a Cat5 cable to the cabin, about 120 feet away. 1 pair for positive and 1 for negative. I’ll double check the voltage drop, but my relay coil works down to 9v.
- 12v DC relay (not SSR), Omron G7L-1A-BUBJ-CB. SPST rated at 30A for 120v. The data sheet shows that it’s rated for 100,000 switches minimum (electrical, 1 million for mechanical)
- 10 minute delay. If it’s a cloudy day it may switch on and off all day, potentially for eight hours, so this would be about 48 switches per day, worst case. I would only use this for less than half a year, so that should be about an 11 year life expectancy.
- The relay coil pulls 158ma, the classic aux 1 is rated at 200ma. I will test to confirm the current.
- 1 second hold. Since I won’t be there, I won’t have voltage sags from turning on appliances and it’s more important to finish charging. Maybe I can go longer?
- Not sure about voltages. Maybe -.3 low and -.1 high? I suppose this is just one that I have to play with, depending on my load?
- I will be using an outlet that is directly connected to the breaker panel with 12ga wire on a 20a breaker.
My understanding of waste-not, is that initially when I get to the high set point in absorb that it will turn on Aux1. At that point in the morning, my panels won’t be producing enough electricity yet to maintain absorb + load, so the voltage will sag to the low set point and turn off Aux 1. Then ten minutes later it would try again. It would repeat that until the sun was high enough to support the load, and then it would stay on until the sun went down or a cloud went by. At the end of the day, it may bounce a few more times before finally giving up.
What’s wrong with my plan? I hope it’s not everything. 🙂 Thanks!
Let me know if you get the Aux 1 waste not to work . I messed around with it a little bit but never seem to have gotten it to work like I thought it should.
I have two Classics so have a second free AUX 2 and use that in Waste Not to a SSR to pwm AC resistance heat loads and that works perfect. It will take just the extra that battery doesn't need when in Absorb and Float and send it to the load.
But there may be some issue with the inverter trying to keep up with the fast pwm switching. But I keep using it .
I have some of this documented in a few videos I made and they are on my youtube channel. This is one video but there are others
https://youtu.be/qj97LjGcj1Y
Larry
I’ll let you know how it goes! I’ll probably try a prototype in about a week and a half. I’ll be there watching the numbers as the sun comes up, watching out for unexpectedly rapid switching and falling out of absorb or float. Should be fun. 8)
Some space heaters have two power level settings too in case the full load you have there is too much.
My radiator style heater has toggle switches on it to select the power levels.
I think you can actually make it do three different power levels.
Larry
Yeah, mine has one switch for 600, another for 900, and both for 1500. The thought had crossed my mind of getting really fancy and re-inventing the waste-not mode with an Arduino. I could read in the current charge mode and voltage and basically replicate all the same logic that Aux1 waste-not uses, except that I could use two relay outputs to the heater to first try 600, then 900, and lastly 1500. It would always just back off to whichever setting supported the low voltage setting. That way I could get most of what Aux2 PWM can do, but with standard AC. I could even set up two heaters so that I could go through all the various possible power levels. 600,900,1200,1500,1800,2100,2400,3000.
That’s just daydreaming at this point though. Focus, focus!
Regarding the load, do you think that 1500w would be enough to lower the voltage and take it out of three stage charging, or were you thinking more about the morning and late afternoon periods only? Also, at what point does it actually knock it out of three stage charging? Is it after a set period of time when it’s below the necessary voltage? Does it bounce right back into absorb if it momentarily drops to bulk?
Thanks!
Here’s what my charging looked like yesterday by the way. Power begins flowing at 7:15am, the peak is when it switches to absorb at 8:15am, and the minor drop off is when it switches to float at 10:45am. Power stops at 4:30pm.
are you able to remote monitor your system when you are away ?
I mean other than the my midnite .
Hard to get any real info from your mymidnite screen shot since if you didnt have a load on it then it is tough to know what the available power was after it went to Absorb and Float.
yes it will change pretty quickly from Bulk back to Absorb once the battery voltage climbs back up to absorb setpoint.
when it is in float it will go to float mppt if it drops below float voltage setpoint.
If you are good with Arduino code you may want to look at Will's project - he wrote some code to control the Classic via modbus . He had it set up to charge his electric car. I forget now exactly what it was doing. You can see that and other projects in Open Source section of forums here.
I think if you wanted to control things - might be easier to do using Node Red running on a raspberry pi .
larry
The Arduino/Raspberry options are pretty cool. I looked into modbus just a little and I think I could use that eventually.
For monitoring, I could set up a static IP address to use the desktop app remotely, but I haven’t actually even tried it locally yet. Does it have any advantage over MyMidnite other than more instantaneous readings? I haven’t tried it because I don’t have a PC with me when I go to the cabin. I guess I’ll bring one next time just to try it out.
I attached another screenshot. This is from a couple weeks ago when I was down there and playing with turning on a 1500w heater, basically manually using the opportunity load. You can see the plateau at around 1700w when I had that on (along with a few fans and lights). The peaks were when I was running the microwave, and the peak power production says 3000w. That seems pretty realistic for 3150w of panels in the winter.
I got my heater and measured it at 1290w. The part that I don’t understand yet is how much a load like this will pull down the voltage, if at all. I would think that the Midnite could maintain the desired absorb/float voltages as long as there is excess power available. Is that true though? If there are 3000w available, will a 1290w load pull down the voltage or will that only happen when the load plus charge exceeds the available power?
Thanks for all your help. :)
Quote from: hansenjw on January 23, 2022, 04:49:32 PM
The Arduino/Raspberry options are pretty cool. I looked into modbus just a little and I think I could use that eventually.
For monitoring, I could set up a static IP address to use the desktop app remotely, but I haven’t actually even tried it locally yet. Does it have any advantage over MyMidnite other than more instantaneous readings? I haven’t tried it because I don’t have a PC with me when I go to the cabin. I guess I’ll bring one next time just to try it out.
I attached another screenshot. This is from a couple weeks ago when I was down there and playing with turning on a 1500w heater, basically manually using the opportunity load. You can see the plateau at around 1700w when I had that on (along with a few fans and lights). The peaks were when I was running the microwave, and the peak power production says 3000w. That seems pretty realistic for 3150w of panels in the winter.
I got my heater and measured it at 1290w. The part that I don’t understand yet is how much a load like this will pull down the voltage, if at all. I would think that the Midnite could maintain the desired absorb/float voltages as long as there is excess power available. Is that true though? If there are 3000w available, will a 1290w load pull down the voltage or will that only happen when the load plus charge exceeds the available power?
Thanks for all your help. :)
It really depends on how much extra power is available if it will pull down the battery voltage or not. Since the light intensity is different every day it is hard to predict what will happen - and of course clouds move in and out during the day.
It also depends on how cold your batteries are .
The best monitor is what I am using now which is using a raspberry pi that gets the Classic Data via ip and then sends it out as mqtt to mosquitto server on the pi . That way more than one program can access the Classic data at the same time off mqtt. Otherwise the modbus is limited to one connection at a time. So with the Classic Monitoring app I can view data ( this app will work for both mqtt and modbus) , then connect to the pi with a computer and I have grafana running on that so there are no limits to custom graphs that you can make. For remote you could sent to the mqtt data to an outside server and then subscribe to that anywhere you are.
Look here - installs are all very well documented
https://github.com/ClassicDIY
Also I have a topic on the grafana along with the dashboard code too.
http://midniteftp.com/forum/index.php?topic=5741.0
I also have node red running on the pi so that can be used for controlling things too though I haven't gotten that far yet. I would like to because I wanted to do sort of what you want to - turn loads on and off according to available power. What I would like to do is have priority loads and then when they are satisfied turn other on.
Larry
I have an update. It's installed and running! I'm currently only running the heater at 600w. It's working well!
The one issue that puzzled me a little is that I initially set my voltages at -.2v and -.4v but waste-not never came on. The voltage displayed on the charge controller would generally only get up to .-2v from the absorb setting, so it seems like the voltage needs to be set lower but not equal to the actual voltage. So I changed it to -.3v and -.5v and it seems to work. It goes through absorb and transitions to float just like normal, so I assume that's good.
Today was a cloudy day and it ran from 9:30-2:40 with just one gap where it got too cloudy and it dropped out nicely for that. 600w seems to be good on a cloudy day for my system since if I had put it at 1500w it may not have run at all today. I'll have to decide which makes more sense for me. Since my goal is to lower propane usage while I'm gone, I'll have to experiment with seeing whether 600w for longer or 1500w for shorter results in more overall savings.
Regardless, it has been a pretty successful experiment, and certainly motivates me to try going further with a multi-stage approach. Of course that makes it a LOT more complex, so we'll see!
Thanks for the update and those settings - I will give them a try for AUX1 waste-not.
I would like to use it in conjunction with temperature controller to turn a space heater on and off.
So I will have the Waste Not turn on SSR to power load, but the heater will also have thermostat control that can shut it off when it gets up to temperature.
I know space heaters have a thermostat but I am going to use one with wired temperature sensor that can be placed where ever I want it .
Larry
I thought of a much less complicated way to get multiple levels of heat output. I was previously thinking about hooking up an Arduino to the classic to read the voltages and set points through the serial port, but that seemed fairly complex. Instead I realized that I can probably just reduce the delay and hold to very low values, like a tenth of a second each. Then it would basically be a near real-time “has excess†logic switch. Then on the heater side I would have an Arduino that would just read that value and set up timers. When the state initially went to on, I would turn in the relay for the 600w element and start my own hold timer of a second. If it went off, then I would turn off the 600w relay and then start my 10 minute delay timer. However, if 600w held successfully, then I would switch to the 900w relay and try that. If the excess switch turned off, I would switch back to the 600w and try 900w again in ten minutes. Since the actual classic delay was really low, as soon as I switched back to the lower wattage, it would turn on again so I wouldn’t drop back to zero.
Basically this would involve a single Arduino and two relays to control the two heating elements on the heater.
Still complex, but lots easier than reading the registers on the classic.
Quote from: hansenjw on March 10, 2022, 11:30:04 AM
I thought of a much less complicated way to get multiple levels of heat output. I was previously thinking about hooking up an Arduino to the classic to read the voltages and set points through the serial port, but that seemed fairly complex. Instead I realized that I can probably just reduce the delay and hold to very low values, like a tenth of a second each. Then it would basically be a near real-time “has excess†logic switch. Then on the heater side I would have an Arduino that would just read that value and set up timers. When the state initially went to on, I would turn in the relay for the 600w element and start my own hold timer of a second. If it went off, then I would turn off the 600w relay and then start my 10 minute delay timer. However, if 600w held successfully, then I would switch to the 900w relay and try that. If the excess switch turned off, I would switch back to the 600w and try 900w again in ten minutes. Since the actual classic delay was really low, as soon as I switched back to the lower wattage, it would turn on again so I wouldn’t drop back to zero.
Basically this would involve a single Arduino and two relays to control the two heating elements on the heater.
Still complex, but lots easier than reading the registers on the classic.
Do you have room for more PV ? Ever think about doing direct dc PV . Power goes into resistance heater . All that is needed is a thermostat in case your area would get too warm. No need for batteries, controller, inverter.
That is what I am doing right now. I just need to put in a relay to switch from water heater to space heater. Right now I just flip the switch manually. I could leave them both on but the two loads are matched to PV independently so if I turn them both on at same time the power from direct dc pv goes down.
Larry
Yeah, I’d do DC except that I have my solar and batteries in the shed and I want the heat in the cabin.
Regardless though, I’m using Aux 2 for the Whizbang JR, so I’m left with Aux 1 and no simple way to modulate the power output.
It’s too bad we can’t get PWM on Aux 1!
Quote from: hansenjw on March 10, 2022, 01:52:10 PM
Yeah, I’d do DC except that I have my solar and batteries in the shed and I want the heat in the cabin.
Regardless though, I’m using Aux 2 for the Whizbang JR, so I’m left with Aux 1 and no simple way to modulate the power output.
It’s too bad we can’t get PWM on Aux 1!
Wouldn't it be possible to have the Arduino do pwm ? I am not enough of programmer to tackle that.
Maybe you are ? Maybe possible to take the pwm part of code from an arduino solar charge controller project and have it do the pwm control of a load ?
https://hackaday.io/project/159708-arduino-pwm-solar-charge-controller
Yeah, an Arduino could do PWM. I suppose that would give me infinitely variable control of the heat output, but I would need to rectify the AC voltage to DC and probably use some capacitors to prevent the PWM load from affecting my inverter. My Magnum inverter would <probably> be ok with a rapidly switching load, but I don't want to risk it and it would be a lot more complicated.
I did solve another problem in my head though. The issue is that the Classic outputs a 12v signal and an Arduino accepts a 5v signal. I read online that an optocoupler would be a good solution to isolate that circuit and drop the voltage, but I realized a simpler solution would be to switch the classic from 12v to dry-contact mode. That way I just treat the Classic as a simple switch using the 5v from the Arduino and it's completely isolated. I ran a voltage drop calculator because I have a long run of wire that the signal would go through and it came to only a 2.4% drop, so I think it will work well.
So now the entire list of parts I would need are...
- Arduino Nano(?)
- 2x 10a SSR, 5v switching voltage, with heat sinks
- Between 10-500 hours of my time :D
Quote from: hansenjw on March 11, 2022, 10:29:54 AM
Yeah, an Arduino could do PWM. I suppose that would give me infinitely variable control of the heat output, but I would need to rectify the AC voltage to DC and probably use some capacitors to prevent the PWM load from affecting my inverter. My Magnum inverter would <probably> be ok with a rapidly switching load, but I don't want to risk it and it would be a lot more complicated.
I did solve another problem in my head though. The issue is that the Classic outputs a 12v signal and an Arduino accepts a 5v signal. I read online that an optocoupler would be a good solution to isolate that circuit and drop the voltage, but I realized a simpler solution would be to switch the classic from 12v to dry-contact mode. That way I just treat the Classic as a simple switch using the 5v from the Arduino and it's completely isolated. I ran a voltage drop calculator because I have a long run of wire that the signal would go through and it came to only a 2.4% drop, so I think it will work well.
So now the entire list of parts I would need are...
- Arduino Nano(?)
- 2x 10a SSR, 5v switching voltage, with heat sinks
- Between 10-500 hours of my time :D
Yeah I forgot about switching AC like that - and I should know because I have been doing it via AUX 2 and it does effect the inverter output - it makes my LED lights flicker when it is active. Other things like my computer monitor don't notice it . Probably not good for any of the power loads. Though I keep using it because it works so well at utilizing the extra power automatically.
These are the best deal on Crydom zero crossing point relay . I got one and it works fine. The only difference on this and other model relay is that hex terminal but wasn't a big deal as it took a ring terminal okay.
https://www.ebay.com/itm/262852016800
It doesn't have an led to show when it is on - which is a nice feature that is on other SSR that I have.
Lots of cheap ( fake brand names ) SSR's out there on the market , though some of them work okay.
If you don't mind waiting a month or so to get your relays I have bought from this aliexpress company - they have higher voltage DC rated relays too which I have successfully used ( until I went over the 220v dc rating and it quit working !)
https://www.aliexpress.com/store/5105062?spm=a2g0o.order_list.0.0.21ef1802PyIuWx
Larry
I'd just give a shout out to a company here in the States called VB Controls. They make IMO high quality, competitively priced relays. Their support is also good. After a couple of failures with some of the inexpensive internet "deals", I've moved to these and had no problems. We do load diversion, but using DC elements, and relays, on Waste-Not.
We also have a couple of the Teledyne relays, and they've worked well.
Did you get something up and working for you? Best, ~ks