HELP:Solar MPPT->Battery->Inverter

[abicash]

Hello

I have designed a MPPT Buck controller which feeds a battery and the battery is connected to a DC-AC Inverter.

My Panel Voltage is 60Vmp and 67Voc at 30A max power (Isc is 36/37A)

I try to keep the PV voltage at Max power using the P&O algorithm.
For starters i kept the PV voltage at 60V. I limited the current to 20A.
I connected a DC(48V)-AC (230V) inverter at the battery.Loaded the inverter with a 1200W load so that it demands a 25A current.

What i presume : The Inverter Load will demand current from the PV over battery since its
Vo/p= 60V-MOSFET drop- choke drop- protection diode drop = 54V.

What i see: The panel current never crosses 12A although its capable of giving 30A (limited at 20A of course) at noon.

Can someone explain what's wrong?

 

[Tony Stewart]

Panel supplies 30A@ 60V at noon ( with what load ? undercharged battery?)
but Panel only supplies 12A @60V with Inverter load of 1200W ( 1200/60=20A 20/25A =80% efficiency )

Inverter sees 60V? but drops out at 54? or runs at 54V?

is Inverter connected to same battery load as 30A load test? Show differences.

Is that RMS or average or peak current?

 

[mab2]

What i presume : The Inverter Load will demand current from the PV over battery since its...Vo/p= 60V-MOSFET drop- choke drop- protection diode drop = 54V...

What i see: The panel current never crosses 12A although its capable of giving 30A (limited at 20A of course) at noon.

If the controller can sustain a constant battery terminal voltage (54v?) then yes, it should support the load up to it's current limit.
Not sure I understand, but if the above 6v drop represent the minimum Vdrop across the MPPT (at 12A) then that is probably what's limiting your PV current, as the battery will likely start to support the load from 54v until the surface charge has gone from your battery.

Do you know what the actual battery voltage is with the 1200w load? compared with the battery voltage without the load?

What is your target battery charging voltage? (as an aside:- 54v seems a tad low to me - unless you're only using the battery occasionally - I'd be aiming nearer 55.2V -ish for sealed types and higher still for flooded batteries).

My Panel Voltage is 60Vmp and 67Voc at 30A max power (Isc is 36/37A)

Well this is a personal opinion; but I wouldn't bother with an MPPT if the Vmp is 60v as direct connection (via a blocking diode and cables + some sort of dump or PWM regulator) means that direct-connected panels are operating quite close to Vmp anyway, so unless your MPPT is exceptionally efficient it will likely lose more than it gains.

 

[Tony Stewart]

MPPT is normally 80% of Voc or 54V not 60V.or 90%.

 

[abicash]

Hello
Thanks for the replies
My replies in blue below your respective quotes

Panel supplies 30A@ 60V at noon ( with what load ? undercharged battery?)
Panel is supposed to supply 30A@60V panel voltage but it doesn't!
Battery is undercharged (rather has lost charge with the 1200w load, voltage has dropped to 46V)

Inverter sees 60V? but drops out at 54? or runs at 54V?
Inverter doesn't see 60V, but sees 52-53v i think due to the drops mentioned in OP

is Inverter connected to same battery load as 30A load test? Show differences.
Inverter is connected to the battery and the inverter is loaded.

Is that RMS or average or peak current?
all currents were DC currents

If the controller can sustain a constant battery terminal voltage (54v?) then yes, it should support the load up to it's current limit.
Not sure I understand, but if the above 6v drop represent the minimum Vdrop across the MPPT (at 12A) then that is probably what's limiting your PV current, as the battery will likely start to support the load from 54v until the surface charge has gone from your battery
You may be right. But isn't the whole MPPT phenomenon based on the principle that PV voltage should be kept at Max power point?Battery should've taken the max current permissible as per this.
Otherwise the Solar charge controller is just a Buck regulator to keep its o/p at 54v to keep the battery voltage happy.
Isn't it?

Do you know what the actual battery voltage is with the 1200w load? compared with the battery voltage without the load?
Battery voltage drops gradually from 50v to 46v (till the time i took these readings).
When the inverter load is removed, the battery voltage rises with the 12A from PV panel
What is your target battery charging voltage? (as an aside:- 54v seems a tad low to me - unless you're only using the battery occasionally - I'd be aiming nearer 55.2V -ish for sealed types and higher still for flooded batteries).
I haven't actually taken battery charge voltage into consideration.I only keep Panel voltage at 60V constant.
In the case where inverter charges the battery from mains (PFC boost) its 58v though.Rather, charging is terminated here.
And kept to trickle charge.

 

[mab2]

OK, if it's still 12A with the battery down to 46v then it's possibly not the Vdrop of the MPPT.

So the question is: what's limiting the current?
1/ The panels not actually at Vmp:
TS enquired about the normal relative difference between Vmp and Voc for the panels - can you confirm Vmp really is 60V?

2/ The panels are not actually producing 30A:
are the panel really producing 30A? You could try connecting directly to the battery (briefly) and see - as long as the battery voltage doesn't go up too far you should get ~30A.

3/ The sun:
Are you sure you're testing in bright sunlight with no shadows? - not hazy, weak sun?

4/ the MPPT:
when it was outputting 12A did you test the actual panel volts? were they at 60V? or higher (if the MPPT isn't taking all the available power the panel volts will rise above Vmp). I haven't seen your MPPT so I don't know how it is limiting current to 20A - many dc-dc converter controllers limil peak amps through the switch, but if it's limiting at 20A you should still see about 20A out of the MPPT but slightly less going in.

5/ the battery/load:
if the battery volts were down to 46 then we can probably rule this one out (unless Vmp is <60V).

 

[abicash]

Thanks for replying mab2

I will check point 1&2 today.
3 is confirmed.
As for 4.. Yes PV volts is steady at 60v.
I limit current by reducing the duty to the series MOSFET , which is the only control switch here. Same switch controls PV volts at 60v. So if current were to increase beyond 20a ,a step by step reduction in duty is forced thereby reducing current and increasing PV volts.

 

[abicash]

Double post

 

[abicash]

Hello
I checked today.
Tony was right, Vmp was near 54V.
I got max 23A now as against 32A as expected.

I loaded the inverter with the same load and as soon as MPPT starts tracking, it settles near 52V .

 

[mab2]

Good - there had to be a logical explanation.

52v is low for battery charging though, so you may want to consider either direct connection so you don't see a Vdrop under load( that's what I'd do), or reconfiguring your panels (if practical, and your MPPT can take Voc) for a higher Vmp, say, if they're currently pairs of 27v Vmp panels, putting them in 3's for 81V Vmp.

 

[Tony Stewart]

The question again is what test conditions produce at 30A max power?

If the PV can supply 30A at battery voltage of 54V and Inverter pulls voltage below this to 46V

then the inverter must draw more than 30A or the test is incorrect.

Both cannot be correct.

I would bypass the MPPT circuit and go direct as your battery is pretty well matched to Voc *80% ( actually closer to 78% from my experience)

 

[abicash]

Hi Tony

The Panel specifies
Maximum Power: 250Wp (+/-3%)
Voc :37.60v
Vmp : 29.5v
Isc : 9.05A
Imp : 8.43A

I have 2 panels in series and 4 such in parallel.
So theoretically it should produce 8.43 x 4 = 33.7A @ 59v

 

[Tony Stewart]

Check that series drops are matched under load. Theory is not matching results.

Try different loads besides inverter.

PV balancer may be necessary like LiPo balancer.

 

[Tony Stewart]

PV's are essentially leaky current sources with zener limit for Voc.
Series current sources will have differential p.u. voltage according to % p.u. mismatch in solar efficacy

load balancers like LiPo arrays may bee to be done.
Slow half bridge with big dead band to drive midpoint between PV's based on overall V/2 using pulse and integrate adjustments to avoid VI loss in Mosfet.

 

[abicash]

Unfortunately the cost is the limiting factor, and is fixed, so load balancers may be out of question.

I had used 3nos IRFP250 as main switching devices, which contribute to the substantial losses.
I am going to try another MOSFET IRFB4227PBF which seems to have better rdson and current.

 

[Tony Stewart]

You can eliminate cost of MPPT cct and perform task of load balancer with 2 power FETS and L,C + comparator control circuit < $3 + interconnects on PV

These are only 14 mΩ, 56A 100V, $0.65USD (2k)

Only half the voltage Vds needed and may be ganged.

or

These are 16mΩ 60A, 100V 104W $1.04USD (1k) < Rg = 1.15 Ω or an impedance reduction of x 72

The trick is however to use lowest voltage possible to get lowest RdsOn that is lower than Vbat ESR which is determine by quality, ageing, V and Ah rating.

Internal Avalanche protection or TVS may be suitable.

 

[NorthGuy]

I think you pile everything together. Impossible to see, the future is.

First panels. If it's not too hot (I assume you're in some hot place where it's not freezing like here), hazy, and sun shines at 90 to panels, you're likely to get about 28A from your array. It'll be much less if the Sun is at 45 degrees to the panels. Note that it doesn't depend on your controller, battery, inverter, or anything else - only Sun, panel, and panel voltage.

Assuming the angle is not 90 degrees, your 23A seems reasonable. You can try tweaking voltage (without tracking) to see if you get more at a voltage which is sloghtly less (or slightly more). This way you'll know if it tracked your into bad voltages, or your voltages are good.

Efficiency. First your controller. It gets 1350W. Measure current between battery and controller as well as battery voltage. Multiply. Divide by 1350W - that's the efficiency of your controller. If you didn't try hard to keep it high, it might be less than 90%. 1350x90 = 1215. This doesn't depend on what inverter is doing.

Inverter also will have some efficiency, could be anywhere from 85% to 98%. This actually has nothing to do with the controller. If it gives you 1200W and you supply 1215W, batteries are discharging. Battery voltage is going to be 48 to 51V. If not, something's wrong with your measurements.

BTW: What kind of batteries do you use? Lead-acid? Then don't limit current. Discharged battery can take lots of current. Charged battery can boil out from much lesser current. If you want your batteries to last, you need to program some sort of charging profile.

 

[abicash]

Hi Northguy

Thanks for comments.

I got what you said. I will measure on Monday and get back.
I use lead acid batteries.
By a profile ,do you mean charging till full battery voltage (14.5 x 4 = 58v)
Then stop charging till battery drops to maybe 13v. Here restart pwm. I am not sure how to accommodate for absorption voltage,else that can be introduced as well.
But you say do not limit current. I am a bit skeptical.

 

[NorthGuy]

Standard procedure is to limit voltage. You limit battery voltage to 59V (depends on batteries). This is so called absorption voltage. You keep it that way for a certain time (say 4 hours, or until current to the battery goes below certain amount), then you deem that the battery is charged and go to float. During float, you also limit battery voltage, but at much lower float level (53-54V). Float lets you keep the battery charged while your controller supplies current to the loads.

Until absorption voltage is reached there's no reason to limit current unless your battery is hopelessly undersized. After absorption voltage is reached, the voltage limit will cause the current to fall in a nice smooth curve until absorption is done. In float, there will be very little or no current into the battery - everythng you produce (if any) will go to the loads.

 

[abicash]

Yes correct.
But during float ,the voltage is to be reduced to 53v by increasing duty. If there were no load on the charger won't the battery try to over charge or boil?