pv dump load to water heater on midnite solar classic?

Options
We're off grid, plan to have enough pv panels for cloudy winter weeks. And a surplus of power when the sun's out. I assume I can dial back the Midnite Classic's max current to the battery to be an appropriate C/10 so as to not overheat the battery on sunny days. Would like to string the panels up for 200 volts, and PWM any excess power into a 50 gallon electric water heater, perhaps pumping a week's worth of heat in a large insulated cistern in the basement. The Midnite Classic manual reads: "Pv V on High: This mode is PWM based and will PWM Aux 2 above a user set voltage (V High) based on the input voltage to the Classic and stop when it hits a low voltage set point (V Low)." Sounds like it might possibly do exactly what is needed here, though not obvious how the duty cycle of the PWM signal varies with PV voltage. Would run Aux2 into the gate of a large NFET, using that to switch PV power into a standard 220 volt water heater element (if using the Classic 250).
Seems this should work with any MPPT controller if you add a small external PWM controller to replace that Aux2 feature, sends PV power to the water heater whenever the the PV voltage exceeds a (possibly temperature compensated) fixed value for Vmp from the panels, at a duty cycle that is proportional to amount that the panels exceed that estimated Vmp value. The MPPT will find it can increase current (and thus power) without changing the PV voltage significantly from that Vmp estimate, until it sucks enough current to shut the PWM to the water heater off entirely. So the battery has precedence over hot water, as it should. And power to the water heater does not go through the charge controller, battery, or AC inverter, allowing all of those to be kept small and cheap. For use with MPPT controllers with a PV Vmax of 150 volts, there are 120 volt water heater elements available at a reasonable price. But it's big bucks for a large 24v element.
Quick Edit: I see other threads using Aux2 for a dump load to a water heater, though mostly from the battery or from an inverter, not directly from the PV panels. They speak of a PWM pulse width at Aux2 measured in volts. So my guess is the pulse width is however long it takes the water heater to suck enough power from the Classic's caps on the PV input till the voltage falls from VHi to VLo. I was thinking of something more like 30 KHz or so, any inductance in the heater element or wiring to it could be analyzed as the coil in a buck mode switching power supply
«13

Comments

  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    When talking about solar, this type of load would be considered an 'opportunity load' rather than a 'dump load'. Dump loads are used when the wattage must go somewhere, used in generators like wind and water. Solar can just 'not produce' and it's happy.

    Several ways to do this with the Midnight CC. look for 'waste not' here and on the Midnite forum. I suspect there will be better ways as they introduce the battery module, that will measure the current actually going into the battery. This should be released shortly, soon, any day now...
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    OK, an opportunity load. Curious they can't yet monitor current into the battery, makes me wonder if the Classic really does have a good way to set maximum current to the battery when using an oversized PV array as I plan to do. I find that a few inches of 12 gauge copper wire makes a fine current shunt. Thanks for the response!
  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    grid ..

    The Classics have very fine Shunts built into the CC. But to know exactly how much of the current that the CC is delivering into the battery bank, (minus inverter loads etc) there needs to be an additional Shunt in the negative lead of the battery to monitor the actual battery charge current.

    The first test units are just now beginning to ship to the earliest outside testers.

    This MidNite Battery Current Monitor has been promised as a very inexpensive device.

    As Photowhit noted, search for " waste not " here and on the Midnite Solar Forum. Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Other than the New Schneider (XW?) which does this, and the Apollo which does this, but is really just a Trimetric and does nothing with the info(from what I've read) this is a new and needed concept. Accurately measuring what's going into the batteries when a system is running is a very good thing. Personally I think the people at MidNite are about as good as it gets in this industry.

    I too have, actually will have, an over sized array for my battery, 4Kw up and running and 2.7 to go, feeding a 800Ah 24V battery. My heavy loads are in summer and winter will be a breeze in general. I've been switching my water heater on and off manually, either before it's out of 'bulk' or after it's in 'float', so the batteries get topped of during 'absorb' and it has worked during the summer, but I've been laided-off and around a good bit. Things will change and I'm looking forward to my system taking care of this it's self.

    I could have and I think will use the waste not feature but want to be sure the charge controller know the state of the battery. You'll get more responses in time. Likely more people over at the MidNite forum will be familiar with the aux uses. boB visits here at times, and is engineering of the many facets of the MidNite CC, and has many other fires going at the same time.
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • vtmaps
    vtmaps Solar Expert Posts: 3,741 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    Photowhit wrote: »
    Other than the New Schneider (XW?) which does this, and the Apollo which does this, but is really just a Trimetric and does nothing with the info(from what I've read) this is a new and needed concept. Accurately measuring what's going into the batteries when a system is running is a very good thing. Personally I think the people at MidNite are about as good as it gets in this industry.

    When you say "new and need concept", are you referring to the ability to measure (and act on) the current into the battery? If so, I should point out that Outback systems have been doing this for at least a few years with their Flexnet DC battery monitor system.

    --vtMaps
    4 X 235watt Samsung, Midnite ePanel, Outback VFX3524 FM60 & mate, 4 Interstate L16, trimetric, Honda eu2000i
  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    vtmaps wrote: »
    When you say "new and need concept", are you referring to the ability to measure (and act on) the current into the battery? If so, I should point out that Outback systems have been doing this for at least a few years with their Flexnet DC battery monitor system.

    I thought they did some sort of adding and subtracting system that measured at the charge controller and inverter, requiring both be Outback and didn't allow for measurement going into the battery, or account for DC loads? I'm not very familiar with the Outback system, only what I've read here, with at least one discouraging review.
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • vtmaps
    vtmaps Solar Expert Posts: 3,741 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    Photowhit wrote: »
    I thought they did some sort of adding and subtracting system that measured at the charge controller and inverter, requiring both be Outback and didn't allow for measurement going into the battery, or account for DC loads? I'm not very familiar with the Outback system, only what I've read here, with at least one discouraging review.

    The Outback system can read three current shunts. If one of the shunts is a battery shunt, the Outback system can do end amps and other tricks.

    --vtMaps
    4 X 235watt Samsung, Midnite ePanel, Outback VFX3524 FM60 & mate, 4 Interstate L16, trimetric, Honda eu2000i
  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    vtmaps wrote: »
    The Outback system can read three current shunts. If one of the shunts is a battery shunt, the Outback system can do end amps and other tricks.

    So it doesn't require an Outback inverter? This is different than what I have understood.
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    I can see an external battery shunt would be useful, but not central to getting a water heater up as described. A system designed to work through a week of mostly cloudy weather will necessarily have way more current into the battery when the sun's out than the typical load.

    The Classic manual has no mention of "waste not". There's a full description of "Pv V on High" for Aux 1 on page 38, where it turns on at a high voltage set point measured at the battery terminal, and then turns off at a low voltage set point. There are also programmable delays on each, calibrated in whole seconds. The description of "Pv V on High" for Aux 2 on page 40 has a vague mention of PWM, no description of how the duty cycle of that PWM might vary. Can anyone point me to an authoritative description of "Pv V on High" for Aux 2?

    Reading through some of the other threads, appears that powering a hot water heater directly from PV panels using "PV V on High" has been tried, but found not to work. I'd expect it to to turn on the heater for a second or two,which would drag the PV voltage down by tens of volts (or hundreds), and then turn it off till the PV panels charge up those input caps to the Classic enough to reach the on set point again. Which might well happen in tens of microseconds, probably far faster than the Classic is cycling the Aux pin on and off.

    Everybody eventually gives up on this, and either powers the water heater from DC at the battery or through the inverter. (The "capacitance" of that battery bank is many orders of magnitude greater than the capacitance at the PV terminal into the Classic.) The solution might be to PWM with a duty cycle proportional to the voltage above our set point at a very fast rate of perhaps 30 KHz as suggested in my original post. If the Classic firmware cannot do this, then a very simple controller to generate this hot water PWM signal could be built that uses only the PV voltage as an input. Digikey and Mouser have appropriate NFET's for 10 dollars or so for the power switch. I believe this need not disrupt the MPPT algorithm running simultaneously in the charge controller (though some algorithms expecting a typical PV panel response may be fooled), that 30 KHz into a water heater element should work fine in spite of any inductive load encountered if a freewheeling diode is included across that NFET switch, and that radio interference from the 30 KHz chopping of 200 Volts at 20 Amps can be easily controlled. I could be wrong.
  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Not my end of the world, yet.

    Have you gone over to the MidNite forum and asked/searched? Here's a search at MidNite forum. The search here are not as direct, at looking toward the information your seeking.

    I would/intend to use AC water heating, controlled by relay, but worry that drawing over 1Kw with water heater (converted to 120v) and other loads, fridge, A/C, will not allow for proper charging with either timed Absorb or end amps through the Classic, ...yet. My reasons for waiting.

    If you put some breaks in your posts it will make them much easier to read!
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Photowhit:
    Thanks for the advice, I will be poking about on the Midnite site. But I think the external PWM controller could work with any make of MPPT charge controller, so decided to launch the discussion here. Am drawn to the Classic since it can handle 250 VDC, making it a good match for standard hot water heater elements. And a good match for my 100 yard run of wire out to the proposed site for the ground mounted panel farm.
    I would think the Classic firmware would be designed to deal with 1 Kw loads cycling on and off, and still charge the batteries correctly. Though I'm sure there are threads out there suggesting it could be a little bit better.
    On a separate note, if that ~30KHz PWM scheme does not work well for some reason, an alternative would be to have three or four heater elements switched into various series/parallel arrangements with relays to present a varying load to the PV panels. Good enough to efficiently drive the water heater, but probably not enough resolution to play well with the MPPT algorithm in the charge controller. So would probably want to arrange it so you are either charging the batteries or running the water heater, not both simultaneously.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    I previously wrote "a very simple controller to generate this hot water PWM signal could be built that uses only the PV voltage as an input"
    That PWM controller should also have a temp probe on the water heater, so it can stop pumping power to it when an upper temp limit is reached. Set the old AC thermostat on the heater a few deg higher as a failsafe, though it may arc till it burns out on DC if it ever tries to open up.
  • Photowhit
    Photowhit Solar Expert Posts: 6,004 ✭✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    gridloose wrote: »
    ... Am drawn to the Classic since it can handle 250 VDC, making it a good match for standard hot water heater elements.

    The water heater doesn't need a 'match' as it's a pure resistance load. My 3600watt element for 240v is running at about 900watt on 120v. And it would run at 3600watts on 240v dc or at a similar reduction in between. There are couple threads here on water heaters.

    It's a 'great' match for a 100 yard run, but to be honest if I was going that far I might look into the 600v Xantrex charge controller, or placing the batteries and inverter near the array and sending 120 or 240v ac to the home!
    Home system 4000 watt (Evergreen) array standing, with 2 Midnite Classic Lites,  Midnite E-panel, Magnum MS4024, Prosine 1800(now backup) and Exeltech 1100(former backup...lol), 660 ah 24v Forklift battery(now 10 years old). Off grid for 20 years (if I include 8 months on a bicycle).
    - Assorted other systems, pieces and to many panels in the closet to not do more projects.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Does not need to match, but can be considerably more efficient if we keep the PV voltage near Vmp. And does need to match instantaneously if we are trying to simultaneously run the CC's MPPT algorithm. If clouds or sun angle reduce available current from the panels to half that of full sun, the heater PWM should have a duty cycle half what it would be in full sun. If the duty cycle remains fixed at a value appropriate for full sun, the PV voltage in half sun will fall to half (so the panels are only providing half of the full sun current to the heater's fixed resistance, since volts=amps*ohms), and power to the heater winds up being volts*amps = 1/4 of full sun power instead of the 1/2 of full sun power we get if we adjust the PWM duty cycle properly. I'm assuming we have enough capacitance on the PV input to the CC to smooth out the 30 KHz PWM current pulses into the heater such that the PV voltage is pretty much constant.

    Just pulling Vmp off the panel data sheet and keeping the PV voltage near there is a "good enough" algorithm for MPPT, so we effectively have a second MPPT controller for the heater that cedes any and all power to the Classic's MPPT that the batteries might need.
  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    grid ..,

    Think that most any MPPT CC might get a bit upset if something is yanking around on the PV input. And high-ish frequencies will work the input capacitors of most any MPPT CC, which is not so good for them. Then there can be the inductive effects of this many kHz load on the input -- when it switches off, there can be an inductive kick, that needs to be clamped, and so on. Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    "most any MPPT CC might get a bit upset if something is yanking around on the PV input"

    Could be, depends on the algorithm. As an extreme example, if the "MPPT" algorithm simply assumes Vmp is some percentage of Voc, that would fail. But if it actively searches for the maximum power point (as the acronym MPPT suggests) without making assumptions about typical PV curves, it should work fine. As explained in the original post, it would find it could draw more and more current at a nearly constant voltage until the water heater PWM is totally shut down. As current increases from there, it will be seeing the usual PV current/voltage curve.

    The water heater PWM is switching fast enough that the capacitance in front of it buffers the power from the PV panels such that the CC sees a constant voltage and the panels are sending a constant current. So no issue with inductance between panels and PWM. We'd most likely add appropriate caps at the input to the PWM, which would be in parallel with the caps at the PV input of the CC. Yes, 20 Amp surge currents to the heater would work those caps hard. But that's not atypical of large switching power supplies.

    The heater would see pulses of current from the NFET, but at 30 KHz the NFET isn't on long enough to discharge the caps more than a fraction of a volt from Vmp. Any inductance in the heating element would slow the rise and fall times of that pulse of current to the heater, though I'd guess there's not enough inductance to have a major affect at 30 KHz. If it did have a major effect, the topology is exactly that of a buck mode switching power supply, and the heater element current would have a sawtooth waveshape rather than a square wave, which is not a bad thing. Edit on 9/13: triangular, not sawtoothie.

    I'm not familiar with the design of the Classic, but suspect it all comes down to a switching power supply running at 30 KHz or more. Gear shifting across such a wide range of possible input and output voltages is not easy. This PWM circuit would be relatively simple, designed for a specific panel voltage and a particular type of heater load.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    The good MPPT controllers actually do search for power point. Nothing to do with Voc. That point may not be at Vmp for any given condition. Panels being a current source they try to produce Imp at any V from zero up to Vmp (after which the current begins to fall off). The controllers make use of that in their algorithm, varying the "load" on the array as Voltage rises until the current & Voltage are at their maximum.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Here's what I think might be an appropriate FET:
    http://www.digikey.com/product-detail/en/IXFH50N30Q3/IXFH50N30Q3-ND/2773855?enterprise=12
    Comes from this family of parts:
    http://ixapps.ixys.com/family.aspx?i=1
    Note that at 300 volts Vds max, they have parts that can deal with up to 210 Amps. We'll have 20 Amps.

    A couple of these caps in parallel placed very near the FET between Vpv in from the panels and ground would provide 2000 uF, and handle the 20 Amp surge currents.
    http://www.digikey.com/product-detail/en/B43564A4108M000/495-3513-ND/1647735
    With a 30 KHz square wave at 20 Amps into the heater with a 50% duty cycle (our worst case), ripple voltage on the CC's PV input would be (1/(2*30KHz) * 20Amps / 2000uF = 0.166 Volts peak to peak riding on top of that roughly 200 Volts from the panels. The capacitor's 130 mOhm ESR would add around a volt to that (with each cap dissipating a Watt), so may be better off (and cheaper?) with 10 caps of 100 uF in parallel. But this gives an idea of what we're up against.

    Probably want a large external Shottky diode across the FET from source to drain to deal with any heater and wiring inductance (though the FET has an intrinsic diode that could take care of this if not worked too hard).

    A driver for the FET gate is fairly easy since the FET's source lead can be tied to ground. So voltage from the panels goes directly into the heater, the lead out of the heater goes to the FET, and the FET uses the PWM signal at its gate to sometimes connect that lead of the heater to ground. Lots of single chip FET driver solutions out there for $5. Driver should have an enable pin driven from a thermostat of some sort mounted on the hot water tank.

    If the Classic can emit a 30 kHz PWM signal whose duty cycle is proportional to Vpv (the voltage from the PV panels) between Vlow and Vhigh setpoints (always off if below Vlow and on if above Vhigh), that pretty much completes the design. Alternately, (perhaps for use with other MPPT charge controllers) could generate a PWM signal with $10 worth of electronics using a 555 based sawtooth oscillator at 30 kHz, a comparator, and a level shifter to shift Vpv down by Vlow volts so the comparator can work near ground.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Integrating the PWM into the Classic's firmware would have a major win: the MPPT algorithm can inspect Vpv synchronous with our water heater PWM, so the Vpv voltage ripple is no longer an issue until it gets so large as to significantly effect the efficiency of the PV panels. This would allow slower PWM switching, and/or smaller caps across the input to the FET. If we slow PWM switching enough (or perhaps just pause it every 10 ms or so) the DC arcing at old water heater thermostats designed for AC would no longer be an issue. (Could no longer allow a 100% duty cycle on the PWM.)
  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    grid ..,

    Have you had the chance to SEARCH here and on the MidNite Forum, using the search term ' "waste not" ' ? ??

    Much of this is covered ground. Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • NorthGuy
    NorthGuy Solar Expert Posts: 1,913 ✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Sorry, I missed most of the thread.

    It is much better to take power from batteries, than from PVs. This will let Classic to do its work, and at the same time you get "clean" power.

    If you do what you want, you, first of all, permanently add MOSFET I2*Rdson losses to your system.

    Then, Classic is switching too. If your switching frequency has comon harmonics with theirs, you can get oscillations, which will most likely destroy Classic's MPPT algorithm.

    Finally, your heater impedance will change with temperature, so you won't be able to produce a stable load that matches Vmp at all times. If your heater impenance becomes too low, the voltage will drop, and your heater will get only a fraction of power. If your heater impedance becomes too high, the voltage will increase beyond Vmp and you will get even less. As a result, it will be increadibly inefficient.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,501 admin
    Options
    Re: pv dump load to water heater on midnite solar classic?

    I would be very careful with 30 kHZ PWM current across thermostat contacts/circuit breakers... I am not sure what would happen with a 30 kHz "arc".

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?
    NorthGuy wrote: »

    > Sorry, I missed most of the thread.

    > It is much better to take power from batteries, than from PVs. This will let Classic to do its work, and at the same time you get "clean" power.
    > If you do what you want, you, first of all, permanently add MOSFET I2*Rdson losses to your system.

    From the battery is better than all schemes I've seen posted here previously, granted. I believe they suffer from too low a switching frequency.
    Our largest consistent load otherwise would be a 400 Watt fridge. Would prefer not to burden the CC and possibly the inverter with an extra 4KW load.
    The FET (I**2)*R losses only impact power to the water heater, not the CC, and an Rds-on of 0.08 ohms at 20 Amps would burn less than 1% of the 4KW into the water heater.


    > Then, Classic is switching too. If your switching frequency has comon harmonics with theirs, you can get oscillations, which will most likely destroy Classic's MPPT algorithm.

    Good point.
    Either add enough caps to the input of the FET to bring Vpv ripple down to a small fraction of a volt, and/or have the MPPT sample Vpv synchronous to the PWM. The latter may be necessary if the MPPT algorithm is looking at very small voltage fluctuations of perhaps 0.1% of Vpv, as I suspect it may. In that case, perhaps modify the MPPT algorithm to look for 1% changes in Vpv, and monitor that synchronous to the switching of the PWM.

    .

    > Finally, your heater impedance will change with temperature, so you won't be able to produce a stable load that matches Vmp at all times.
    > If your heater impenance becomes too low, the voltage will drop, and your heater will get only a fraction of power. If your heater impedance
    > becomes too high, the voltage will increase beyond Vmp and you will get even less. As a result, it will be increadibly inefficient.

    That's the whole point of the PWM, it sends an average current to the heater such that it does not bog down the panels. Exactly like an MPPT controller, except it assumes a particular value for Vmp instead of tracking it. (Having two algorithms try to track would be trouble. But if all of this is folded into the Classic's firmware, there may be better ways to do this.) Duty cycle of the PWM signal at the FET gate is controlled to maintain a constant Vpv somewhere near Vmp. If the CC sucks enough current to drop it a bit below the assumed Vmp, the PWM shuts down entirely


    #########
    > I would be very careful with 30 kHZ PWM current across thermostat contacts/circuit breakers... I am not sure what would happen with a 30 kHz "arc".
    > -Bill

    Yes, that's why I was suggesting pauses every 10ms or so. Or using the AC thermostat only as a backup failsafe, in which case the worst it could do burn out permanently.

    Edit: This is all one post, with previous posts flagged by the '>' character. I can't get this fool editor to format quotes vs new material properly, or even display them as different.
  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    Vic wrote: »
    grid ..,

    Have you had the chance to SEARCH here and on the MidNite Forum, using the search term ' "waste not" ' ? ??

    Much of this is covered ground. Vic

    Many, many of these issues, and their solutions have been dealt with on the MidNite Forum:
    http://www.midniteforum.com/index.php

    And, on the Wind-Sun Forum (ie, here).

    Try the Search Term, ' "waste not" ', which is what the MidNite Classic implementation calls this approach.

    Have you had the opportunity to SEARCH this term on these two sites. Much has been written. Some of these posters may be tired from all of that typing, and may perhaps be reluctant to type much of it again, ... am just sayin'.

    FWIW, good luck searching! Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    Vic,
    I've looked, though not exhaustively. Will do so as I have time. Have yet to see a convincing argument in the responses here why this won't work. Which is not to say I'm sure it will.
  • NorthGuy
    NorthGuy Solar Expert Posts: 1,913 ✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    gridloose wrote: »
    That's the whole point of the PWM, it sends an average current to the heater such that it does not bog down the panels. Exactly like an MPPT controller, except it assumes a particular value for Vmp instead of tracking it. (Having two algorithms try to track would be trouble. But if all of this is folded into the Classic's firmware, there may be better ways to do this.) Duty cycle of the PWM signal at the FET gate is controlled to maintain a constant Vpv somewhere near Vmp. If the CC sucks enough current to drop it a bit below the assumed Vmp, the PWM shuts down entirely.

    I thought you were going to switch panels between Classic and the heater. Looks like you do not want to shut down Classic, but only switch your heater.

    Classic may not find the Vmp with you. During absorbtion/float, Classic tries to keep the battery voltage at a fixed point. It does this by moving voltage away from Vmp. To do PWM, it needs to know where Vmp is to move Vpv closer to there. For your idea to work, it needs to stop PWM, find Vmp, then start PWM back.

    At 30kHz you will get high switching losses, especially with big MOSFET with Rdson of 8mOhm
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    I previously wrote:
    > Either add enough caps to the input of the FET to bring Vpv ripple down to a small fraction of a volt, and/or have the MPPT sample Vpv
    > synchronous to the PWM. The latter may be necessary if the MPPT algorithm is looking at very small voltage fluctuations of perhaps
    > 0.1% of Vpv, as I suspect it may. In that case, perhaps modify the MPPT algorithm to look for 1% changes in Vpv, and monitor that
    > synchronous to the switching of the PWM.

    One solution that should avoid having the PWM disturb the CC's MPPT algorithm would be to set the PWM's assumed Vmp at 5 or 10 volts above the actual Vmp of the panels. So the PWM will not activate at all until either the batteries are charged or the CC is in current limit. This might work well with most any MPPT controller. But if the water heater PWM feature were integrated into the CC's firmware so it's all coordinated, it's a non issue. There are half measures, such as using a CC Aux output to tell us when to enable the PWM.
  • NorthGuy
    NorthGuy Solar Expert Posts: 1,913 ✭✭
    Options
    Re: pv dump load to water heater on midnite solar classic?
    gridloose wrote: »
    One solution that should avoid having the PWM disturb the CC's MPPT algorithm would be to set the PWM's assumed Vmp at 5 or 10 volts above the actual Vmp of the panels.

    If the array voltage goes 5-10 volts above Vmp, you loose 80-90% of production.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?

    NorthGuy,

    Yes, there could be conflicts with the MPPT algorithm. I don't yet fully understand what algorithm is used on the CC. But with access to the CC's firmware, I'm convinced this can be made to work. I suspect it could be made to work in a generic way with most MPPT controllers, I don't think it needs to shut down MPPT as you suggest. If the CC is tracking, then that means it is hogging all panel power for the batteries. It is only when the batteries are charged or the CC is in current limit that the PWM might turn on, since it only does so when Vpv rises a few volts above the actual Vmp (as determined by the CC's MPPT algorithm) to whatever voltage we set the PWM up for as its assumed Vmp. This discussion assumes two entirely separate entities, the MPPT in the CC and the PWM we hang off the side to drive the heater. If both are integrated into the Classic's firmware, I'm sure there are more elegant solutions.

    Most switching power supplies are switching at 30 KHz or more. Even big ones. The Classic CC is almost certainly one of them.
    As stated previously, the loss due to Rdson of the FET I called out would dissipate less than 1% of the power meant for the water heater.
    That happens even if the switch is constantly on.

    Where 30 kHz switching losses come from is the fact that the FET cannot instantaneously switch on or off due to capacitance from FET gate to source. At the instant for which the FET is halfway on, the FET has as large a voltage drop across it as we have across our 4KW heating element, or 100 volts. And both are seeing the same current. So the element and FET are both dissipating 1KW at that instant (power into the fixed heater resistance is proportional to voltage squared, I'm assuming it does 4KW at 200 Volts). If the gate driver can pump (and sink) several Amps into the gate, this situation is over with in a few millionths of a second.

    These days, even large power supplies of many kW are switching at speeds of several hundred kHz. Though switching losses increase, it allows for smaller caps and coils, resulting in a smaller and lighter package.
  • gridloose
    gridloose Solar Expert Posts: 40
    Options
    Re: pv dump load to water heater on midnite solar classic?
    NorthGuy wrote: »
    If the array voltage goes 5-10 volts above Vmp, you loose 80-90% of production.

    I doubt it. 10 volts out of 200 is 1 volt across each of 10 panels strung in series, where each panel has 20 volts across it.
    A quick look at the curves for a typical PV panel meant for a 12v lead acid battery show a very little loss when running 1 Volt away from a Vmp near 20 Volts.
    And 10 volts should be more than plenty, the PWM ripple could be brought down to under a volt as shown in a previous post.

    Edit: corrected to read 10 panels, not 20