Water Heating Controlled by a Classic 150

[ChrisOlson]

I made yet another improvement to my water heating system today.  The way the system works can be found in other threads.  Prior to this I was using a mechanical relay with the Aux1 output of the Classic on Waste Not Hi mode.

The biggest problem I had with the mechanical relay is that the load was too large for the early stages of Absorb.  So the relay would cycle on and off, and the controller would cycle back and forth between Absorb and Bulk MPPT (when the voltage dropped due to the load), until the bank's charging requirement dropped to where there was enough extra power to keep the 2 kW water heater elements going full time.

So what I needed was an infinitely variable load up to the point where there was enough power to keep that element lit.  I ordered a SSR to try that on Aux2 with PWM.  Works MUCH better.

Some photos below with description under each photo for other folks that might want to set up something like this

[ Attachment Invalid Or Does Not Exist ]
This is the SSR I ordered.  The water heater is 240 volt and this is just a single pole SSR.  I'm running one leg right to the heater and switching the other leg with the SSR.  Yeah, I know, I'm supposed to use a red and black wire for that.  But the water heater has another ground running to the Main so it's wired up with 10/3 Type SEOW with a twist lock plug at the water heater (in case I ever wanted to plug in something else for a load).

The SSR has a little LED on it that glows brighter and brighter at higher PWM frequency (and voltage).  The input side is 3-32VDC, the output side is 24-380VAC, 25 amp maximum.  The SSR is mounted on a heat sink.

[ Attachment Invalid Or Does Not Exist ]
This is the mode used on Aux2 in the Classic.

[ Attachment Invalid Or Does Not Exist ]
This is the settings for Aux2 - at 0.2 volts below charge stage the Aux port begins "tickling" the water heating load.  If the voltage falls more than 0.2 below charge stage, the heater element is off.  Today, as the bank went from Bulk MPPT to Absorb it started flashing the LED.  I checked the voltage going to the heater element and it was only about 5-6 volts.  As the bank charging amps dropped from 90 down to around 50 the voltage going to the water heater element (and the associate load in watts) gradually increased up to where the water heater was drawing the full 2 kW from the inverters with about a half hour remained on the Absorb stage.  The water heater stayed on all afternoon once the controller dropped into float until the sun went down in the sky low enough to where there wasn't enough power to keep the element alive anymore.  And then it gradually dropped the voltage going to the element until it finally shut it off.

So it worked as I planned that it should - we dumped about 10 kWh into the water heaters this afternoon that would've been "wasted" on a good day because the bank couldn't use it.  10 kWh is 34,000 BTU, or enough energy with electric water heating to heat 55 gallons of water from well temp to 130 degrees F.
--
Chris

[gww]

Chris
As always I apreteate when you post pictures.  I have almost the exact ssr and am hoping it is going to work as well with the auxilary funtion on my mx 60 charge controller.  I am still a year off from installation but you give me high hope of success.
Thanks for all your ideals and your testing prior to my having to bite the bullet.
Thanks
gww

[ChrisOlson]

I forgot to post where I got the SSR from:
http://www.amazon.com/DC-AC-Solid-State-Relay-Heatsink/dp/B005K2IXHU/ref=sr_1_2?ie=UTF8&qid=1348973046&sr=8-2&keywords=ssr

This SSR is supposed to be 24-380VAC output.  But its response time is <10ms and with an analog voltmeter hooked up I saw the PWM varying the output voltage from 6-10 VAC to 120 VAC (measured line to neutral).

I'm using 1.0V width for the PWM and that seems to work ok.  However, I may try increasing the pulse width to 2 or 3 volts to see what happens.  The specs for the SSR say the min ON/OFF voltage is >2.4V and <1.0V.  Increasing the PWM width also lowers the PWM frequency.  So I'll have to try that and see what it does.
--
Chris

[gww]

Chris
Good link,  I paid $16 with out a heet sink.  You are above my knowladge level on the pulse width descussion but at least I will be able to refer back to this post when I actually get Into the installation.

Is most 240 volt wired from the electric panel, one wire each  alternate leg with a common neutral?  Three wire.  as I have run 240 before but am not good with electrical terms,  Is your relay using the neutral on the single pole to power up both legs of the heating element?

Do all electric hot water heaters work with the top element heating first then switching to the bottom element?  Are all electric hot water heaters able to adjust the temps on individual heeting elements or do they have to be modified?

I couldn't let this subject pass without getting as much as I can from you as it is in my future to do this.

Thanks
gww

[ChrisOlson]

240V split phase power is 240 line to line and 120 line to neutral.  In the photo the neutral is used for ground (green), one leg is white the other is black.  I'm switching just one leg of the split phase power source.

All DWH's that I've seen heat the water with the top element first, then switch to the bottom.  The elements are individually adjustable for temp, however, whatever you set the bottom one at will determine the temperature of the water in the whole heater.  There's a thermostat controlled relay on the top element, that when the stat set temp is reached, shuts off the top element and sends the power to the bottom one.
--
Chris

[ChrisOlson]

Is most 240 volt wired from the electric panel, one wire each  alternate leg with a common neutral?  Three wire.  as I have run 240 before but am not good with electrical terms

gww,

Just a little note on split phase power - technically it is 4-wire - L1, L2, N and ground.  However, most appliances like water heaters, ranges, etc., are three wire - L1, L2 and neutral.

Where the 4-wire comes in is with main service panels, sub panels and generators.  The neutral is kept separate from ground for those applications so 120V (line to neutral) can be used.  Power does not flow in the ground except for a ground fault situation, but power does flow in the neutral when one leg of the split phase source is used for 120V power.  For 240V  split phase loads power never flows in the neutral since the load is leg to leg (L1 to L2).  The amplitude of the sine wave on each leg is only 120 volts.  But the sine waves on each leg are 180 degrees out of phase with one another so the total difference leg to leg is 240 volts.

Now, in my wife's native country of Sweden 240 volt power is not split phase, so it is three wire like 120 volt power here in the US.  I think the European way is actually better because it uses a lot less copper.
--
Chris

[gww]

Chris
I have run 240 from panels before to dryers stoves, welder.  I can do it by looking at it but can't picture it in my head as it has been sometime since I have did this.  I still don't quite picture what you are saying in my head.  I am sure it will become clearer when I get to doing it.  I was mostly trying to picture how to use a one pole ssr and not letting one leg 110 volt get through to the element untill both legs are powered.

Now if I understand what you are saying.  Putting the ssr in one leg (one wire) of the 240 will keep the element from getting power as the neutral has no power it is the leg to leg that makes the circle?
Thanks
gww

[ChrisOlson]

Now if I understand what you are saying.  Putting the ssr in one leg (one wire) of the 240 will keep the element from getting power as the neutral has no power it is the leg to leg that makes the circle?

Yessir, that is correct.  You still have to use a two-pole breaker to power the whole thing because with just one leg shut off you still have 120V power from the live leg to the water heater case, or ground.  But to turn the element on and off you only have to switch one leg - in fact, that's all a 240 volt water heater thermostat does is break one leg to turn the elements on and off.
--
Chris

[gww]

Chris
OK got it.  Now the only things I don't understand is the pmw pulse width and trickling of power to the water heater.
I am sure This will become more clear as I do it and it is nice to have this post to refer back to.
Thanks again
gww

[ChrisOlson]

The PWM turns the water heater element on and off really fast (several times per second).  The more "off" time there is as compared to "on" time, the lower the voltage will be at the element.  It happens so fast that it has the effect of regulating the voltage at the element (and how much power it draws).

The PWM pulse width is the width (in volts) from full on to full off.  A narrower pulse width means faster control and higher PWM frequency.  But the pulse width has to be adjusted to match the capability of your SSR or load.  AC current in the US, as an example, is 60 Hz frequency.  If you are controlling 60 Hz current, and the PWM frequency is higher than 60 Hz you may not get the desired result because the PWM is turning the current on and off faster than the alternating current can make one complete cycle.

I don't know all the real technical details of PWM pulse width, but I'm seeing today that using a little wider pulse width (2.0 volts) seems to work better with my water heater elements and SSR than the 1.0 volt width did yesterday.  I can see the voltage varying more at the heater element than it did with a 1.0 volt width - jumping from about 38-60 volts AC, and varying constantly, as I type this.  But it seems to be doing a better job of keeping the solar panels at max power as the batteries are absorbing.  I'm only 40 minutes into absorb right now - I'll know more after it gets done with a complete absorb stage to see how the 2.0 volt width works overall.
--
Chris

[gww]

Chris
Are most charge controllers able to adjust the pmw pulse width. MX 60 of course?

I can't wait to hear how it works. 

As a side note I got 1400 more watts solar ordered today.  I am definitely done buying solar untill I install something.  This puts me at about 4800 watts.  I may have to wire some of the solar strait to the batteries along with the turbines as I now have too much for the mx60 cc.

As you can see I will definitely need my hot water to work or I am going to end up blowing up the batteries.
Thanks
gww

PS If the pulse is adjustible would it still be wise to change the elements from 4500 watts to 2000 watts?  Does my added solar play into a decision here?

[ChrisOlson]

4.8 kW.  That's a nice array.  Mine is only 3.5 kW.  But I also got 7.5 kW of turbine power so that helps out too.

I made a little video of playing with this this afternoon.  I played with the offset and the pulse width to see what it does in real time, and that's in the video.

I've concluded that the SSR method on Aux2 on the Classic is definitely smoother than using a mechanical relay on Aux 1.  The controller maintained perfect absorb voltage for 2 hours 30 minutes today, even with varying loads on and off in the house, and wildly varying turbine output from a few hundred watts up to 2.5 kW a couple times.

I don't know if I could put one of them Kill-A-Watt meters inline to the water heater to measure how much goes to the heater every day.  It's 240 volt and I think them things only work on 120 volt.  But it would be nice to know that.  I'd have to get one with a battery in it so it don't reset when the controller turns the element off.  But I don't know how to get one to work on a single leg of a 240V split phase load.
--
Chris

[gww]

Chris
Thanks for the vidio.  I really like you showing the effects of the adjustments. 

Does it boil the batteries exsessivly when the controller doesn't  tell loads from charge making it absorb longer then maby necessary?

Thanks
gww

[ChrisOlson]

Does it boil the batteries exsessivly when the controller doesn't  tell loads from charge making it absorb longer then maby necessary?

Yes, it will cause the batteries to use excessive water if they absorb too long.  It doesn't hurt anything otherwise, and actually eliminates the need for periodic equalization charges.  I got the minimum time at 2 hours 30 minutes and the max time at 3 hours 30 minutes, with absorb ending amps at 48.  So it will drop out of absorb anyway if the max timer is reached.

Normally the wind turbines are putting out enough power so the thing gets "fooled" into the minimum time instead of the maximum time.  So it evens out over the long run.
--
Chris

[boB]

Chris, if  you want to play with the wider width adjustment, try lowering the offset some when width is higher than 1.0 to
bring the dumping power back to near where it was with -0.2V offset and 1.0 V width.  The PWM frequency
will decrease and will operate over a wider voltage range but it might be interesting to see if it works better, worse
or about the range.

It may turn out though that 1.0V width is just what is needed after all.

boB

[ChrisOlson]

boB, I think I gained a better understanding of how the PWM works.  With a zero offset it sent very little power to the heater.  With more offset it sends more power to the heater.  I kept increasing the offset until the panels put out max power, but the proper charging voltage was still maintained.  I ended up at -0.7 seeming to be ideal.  If I increased it to -0.8 the panel output didn't increase, but the charge stage voltage dropped by 0.1 volts (when there was still not enough power available to run the heater at full load).

When I get a chance I will play with that some more to see how your suggestion works.

I'm assuming that the PWM tries to attain a voltage that is lower than charge stage, while the charging part tries to maintain charge stage voltage.  The two find a "balance" where the panels continue putting out full power while the bank doesn't need it all.  I suspect that with different sized solar array, it would probably take a different offset, i.e. "one size does not fit all".  My solar array is larger than the controller's output capacity on a good day, so maybe that's why it takes more offset to really push the array for all it's worth.  With a smaller solar array maybe it would take a smaller offset because there wouldn't be as much "extra" power available.  I don't know.  Can't find much info on it on the internet from anybody that used it in a setup like this.

But regardless, it works awesome (and MUCH smoother) compared to a mechanical relay on Aux 1.
--
Chris

[d34]

Very interesting infomation here.  I like this method, keep up the good work.  One thing I wanted to point out to you is that if you lower the voltage to 1/2, the power rating of the element is lowered to 1/4.  So if you have a 240v 4500w heating element and drop the voltage to 120v the element is only puttin out 1125w.  Just thought I would let u know that just because the voltage is 1/2 doesnt mean the power rating is 1/2.  I would also check the voltage at the element or you can check it in the ssr box by hooking the meter to the output post of the ssr and the other lead tied into the white wire.  That would give you the actual voltage at the element.  Hope this helps.

[ChrisOlson]

Ah yes, that is correct!  Amps = volts/ohms.  Thanks for pointing that out.

It was interesting to see it work today - the solar put out good power this morning, but then it clouded over.  But the wind was blowing good.  It still worked perfectly normal - better than just on solar power, in fact, because we can get 2-3x more incoming power from wind on a good day than we can from solar.  The wind turbines produced 38.7 kWh today since midnight (and they're still running), and a good chunk of that went to the water heater elements.
--
Chris

[OperaHouse]

You may not be getting what you think if you choose the wrongtype of SSR.  A common 25A SSR should be a back to back SCR.  When the SCR turs on it will stay on till the AC cycle  is finished and the current goes to zero.  I looked at the SSR listing and this appears to be a mosfet because it is also rated for DC  I couldn't find a spec sheet listed. With a AC SSR any time it gets a pulse it will turn on and stay on.  You can not control it off.  Two 120V lamps in series as a monitor will what is actually going to the water heater.   Many years ago I built light shows for bands using this same principle, triggering a SCR at a frequency different from the line frequency.  It was a quick and cheap light show.   The LED you look at only shows the drive.  I dont believe you could have anything approaching proportionality with a PWM frequency greater than 5 hz with an AC type SSR.  I did a quick look at the controllers manual the other day.  I was pressed for time as I am now.  Just want to make people aware that PWM into a AC SSR can have some very strange results.

[ChrisOlson]

All I have to do is put the ammeter on the line and I can see it varying the amps constantly.  It says in the specs for the SSR that it's zero cross, whatever that is.

It has been working perfectly, and does what I intended it to do - which is regulate the voltage more precisely than the mechanical relay could during absorb stage, while at the same time keep the solar array at maximum possible output as the internal resistance in the batteries comes up during absorb.
--
Chris

[OperaHouse]

I think there is a possibility is not doing what you think it is doing.  It may only acting in a proportional mode at the very early stages.  Since Midnight says the PWM operates at hundreds of hz through much of the range the relay is just on or off.  This may be sufficient to do what you want but it is not using the full range of the Classic.  I found it odd that Midnight did not go into any detail about what kind of relay to use.  It might be that it is designed to operate with that type of SSR and that the tech writer has no idea what he is talking about.

[ChrisOlson]

I have no idea about any of that.  But I assure you it works perfectly (so far).  As I stated, the reason I went with the SSR is because the slam bang on/off mechanical relay put too much load on the system during the initial part of the absorb stage and the bank wouldn't maintain absorb voltage without cycling the relay on and off.

The SSR seems to vary the load pretty much infinitely, and it maintains proper charge stage voltage no matter what - and at the same time uses any extra available power for water heating - just what I wanted it to do.

So despite the fact that it may be doing something I don't know about - it works.
--
Chris

[toothy]

Hello Chris

I'd just like to say thanks alot!

I don't get how most any of this stuff works but like it when it does and when you help me out with pictures and video's it sinks in a bit better.

Thanks
Wade

[Larsmartinxt]

It says in the specs for the SSR that it's zero cross, whatever that is.

Zero cross means that the SSR can only turn off when there is no current trough it. Basically that means that you can only turn it off if you use it on a AC load (no current when it passes 0V on the sine wave). This is because the SSR contains a thyristor. (You can also buy mosfet based SSR's which works great on DC loads)

Lars

[kitestrings]

we dumped about 10 kWh into the water heaters this afternoon that would've been "wasted" on a good day because the bank couldn't use it.

Isn't that just the best feeling, Chris?  Free fuel.  We started utilizing "opportunity load" in simialr fashion a few years ago.  I think it a is a really effective way to use more of the power you have available.  It is much simplier than a dhw solar install.  And, it may well be a cheaper means of storing more energy - when you consider the cost of battery storage.

We used used DC elements, but otherwise the approach is the same.  I only wish our tank was larger...

Thanks for sharing.

~ks

[OperaHouse]

I was thinking about multiple water tanks.   These tanks in series could be daisy chained electrically by replacing the lower temperature control with the same type thermostat used for the upper.  That way the top element ofthe second tank would come on once the lower element of the first taqnk turned off.  Next year I will be using three tanks.  Two 10 gallon for household water and a 30 gallon for the wash machine.  The diversion to each tank will be controlled by a processor. 

Generally a AC SSR will be a TRIAC under 25A and over that it will be bac to back SCR.  Each of these devices will turn off only when the current goes to zero.  A zero crossing relay will only turn on when the voltage is near zero and will alway give a complete half cycle.  This allows one second of line power to be divided into 120 segments at 60 hz and generally produce much less electrical noise.  Non zero crossing relays allow turn on at any portion of the cycle.  If the controller driving them is synced to the power line, the power can be divided into infinite sections.  I was looking around and this is a good discussion of the process.     http://openenergymonitor.org/emon/node/685

A 120V Kill a Watt could be used on 240V if the power is balanced to the neutral.  It would be placed ahead of the SSR and connected to hot and neutral and the reading would be half of the actual power.  I haven't opened up one of these but it is likely that the current sense is located in the neutral line.   If it doesn't record power then it should be connected "upside down" with the normal neutral connected on the hot side.  This would place the current sensor on the hot leg.

[ChrisOlson]

Isn't that just the best feeling, Chris?  Free fuel.

It is.  That's why we went with all electric appliances in our home, even though many told us you can't run that stuff off-grid.  With the electric range, clothes dryer, etc., we have a "dual fuel" system.  95% of the time RE power can run them, and the 5% or so that it can't we have to use generator power (gasoline).  If we would've went with propane appliances, our choices for fuel would be limited to propane, and you're still stuck burning a fossil fuel to cook, etc..
--
Chris

[ChrisOlson]

I was thinking about multiple water tanks.   These tanks in series could be daisy chained electrically by replacing the lower temperature control with the same type thermostat used for the upper.

OperaHouse - that's the way our system works.  We have two 55 gallon water heaters in series.  The top element in the primary is always on, controlled by the thermostat set at 125°F.  The bottom in the primary, and both the top and bottom in the secondary are daisy chained with those relay stats set as high as the stats will go (~160-165 degrees).  If only the top 25 gallons of the primary is left it takes about 75,000 BTU to heat both up to 160.  So we can dump up to 30 kWh into those heaters in a day, and on really windy days with three turbines going we've done it many times.  We don't have anywhere close to enough solar power (3.5 kW installed capacity) to heat that much water in one day.  But the wind turbines can do it with no problem and we can hit 60 kWh with the wind on good days.

It has worked very well for us.  If both heaters are full of 160 degree water, it lasts us for up to a week.  I put one of those mixing valves on the primary recently so we don't get burned by the hot water.  The mixing valve mixes it to 125 degrees, which is a much safer temp.
--
Chris

[Eric L]

Chris,

Thank you for this very cool idea. I'm new to this forum but have tried your suggestion after reading your post (which I found when boB linked to it from the Midnite website). I bought the same SSR as you suggested, and wired it to to the bottom element (240 v, 3500 watt) on my water heater. The top element is still on grid power but on a low temperature setting. I have a 4.1 KW solar array which is grid supported; i.e., it switches to grid power when the battery voltage drops below a certain threshold.

You're idea is working pretty well for me so far; I tend to have a lot of unused pv capacity in the afternoons. I have however encountered one problem, which is voltage drops and surges caused by the element cycling on and off. My inverter (a Magnum MS4448 PAE) is showing an AC voltage drop of about 17 volts (down to 103 VAC at the lowest) when the element kicks on, before bouncing back up. When I noticed this I disconnected some of my other loads, such as the fridge, to keep them from being damaged by the voltage swings.

Anyway, I'm wondering if you have encountered this and if you or anyone else here has any ideas to smooth the voltage. One thing I may try is to install a smaller element, which would probably help some.

Even if I can't eliminate the voltage swings, this is a really cool idea that has cost me almost nothing to try. If I can't manage to reduce the swings, I'll nonetheless likely put the relatively small household loads on the grid during sunny days and heat water instead, since this would almost certainly use more of the solar array's potential. So thanks again for sharing it.

[ChrisOlson]

Eric, I'm using 2,000 watt elements with no problems.  But could the problem with the voltage drop be that you don't have enough amp capacity to feed the inverter when the heavy load comes on?  Either the bank, or the wiring to the inverter can't supply its surge load requirement?

The Magnum MS-PAE is a good inverter, but I don't think it has the legendary surge capacity of the Trace/Xantrex SW's.  But one thing I found out very early about the SW's is that it takes a pretty heft battery bank to feed them if you want power.  So I think that's where I'd investigate first - try some other loads on the inverter and see how it maintains voltage with other heavy loads.  If it can't, then ask Magnum Energy why it can't, and hopefully get that fixed.

But with the SSR on the element it SHOULD modulate the power fairly well too and not bring it slam-bang on when using the Aux2 with PWM to drive it.

Edit:
Eric - I noticed that my pictures showing the settings and stuff for the SSR disappeared off this thread for some reason.  If you need those to tune yours, let me know and I'll see if I can re-post them.

Another thing I want to note about all the above is that your Classic might be "seeing" one voltage (bank plus resistance in the wiring from the bank to the Classic), while the inverter might be "seeing" a lower voltage (due to bank under-capacity or resistance in the wiring from the bank to the inverter).  This can cause some undesirable results in all this.  When we put in our new battery bank and inverters I ended up going with way overkill on wiring to the inverters, as well as quite unconventional bank wiring to get the amps to them to run big loads, without voltage drop at the DC studs on the inverters.

Most people wire their inverters up for max rated load and forget about surge loads.  Those inverters can't make power out of thin air.  They need DC amps - and lots of 'em - to deliver full surge power without getting a voltage drop.  I rather suspect that's your problem - you don't have big enough wiring to the inverter to support the big load during surge.  One way to verify that is to put a analog DC volt meter on the inverter studs and load it to full rated surge capacity for 5-10 seconds and see how the voltage compares at the studs to what you measure at the bank.  Don't use a digital because all you're going to see is rapidly flashing numbers.  An analog meter will show you the massive dip when you slam-bang a 8 kW load on that inverter.  But that test will tell you how well your wiring (and bank) holds up to big loads.
--
Chris

[gww]

Chris
I just skimed through this thread again and might have missed it but what is your voltage set up to kick on and kick off the ssr?
Thanks
gww

PS Why the missing pictures?

[ChrisOlson]

Don't know what happened to the photos.  They were there and now they're gone.

With the Classic the voltage that the SSR comes on at depends on the charge stage.  I have my absorb voltage at 31.0 volts and the offset relative to stage voltage is -0.7 volts.  So at 30.3 volts it starts "tickling" the SSR, but if there's not enough power available to keep it lit, it don't turn on the elements very much.  As the bank charges up and there's more extra power available, it modulates the SSR faster so more power goes to the heater.

Same thing happens in float - there's lots of extra power so the SSR is on all of the time.  I have float set at 26.4 volts.  So when the incoming power drops to where the bank voltage drops out of float into Float MPPT, it keeps modulating the SSR until the voltage drops to 25.7 and then it shuts it off completely.

I don't know if other controllers can do this.  I think the others just set a voltage for where it comes on, and it's not really relevant to the charge stage of the bank.
--
Chris

[gww]

Chris
Looks like I will be reading the outback manual some more.  Thank you very much for responding.
gww