Help me understand this buck converter

[mberg2007]

Hello all,

I'm trying to power a device directly from a solar panel (no battery) and having some trouble determining just how to do that.

My problem is that solar panels generally fit well with battery voltages, but not so much with fine electronics that run at 5V. That's nowhere near the Vmp for most panels so I am not getting a lot of efficiency from my panels.

I have tried various configurations with buck converters and so far have not had a lot of success. Problem is that even with an input cap the buck will crash the panel when the electronics begin to consume more than the panel produces.

I have found this unit on eBay: http://www.ebay.com/itm/191927867733

It seems that it can do some MPPT magic, and I'm guessing that it will somehow allow me to dial in the Vmp, but just how it is supposed to work without a battery on the output side I don't know.

Do any of you smart guys have any experience with this unit, or similar units, that have these "MPPT" settings? What exactly does that setting do, and are these converters superior to regular buck converters?

On a secondary note I have ordered one of tthese: http://www.ebay.com/itm/122054336107

It is a tiny unit but apparently capable of 3A output, which is enough for my needs. The major thing about that unit is that it has an UVLO on the input side, which should allow me to recover better after a panel crash. But this thing doesn't have a magic "MPPT" dial though. I'm not sure if that means I'm once again going to be missing out on panel efficiency.

Again - NO BATTERY. The electronics I want to power can go off and on according to lighting conditions, and I'm not very interested in adding a charge controller and a bulky battery to this setup. I might add a cap or a supercap to handle power spikes on the output side, but that's it as far as energy storage goes.

Regards,
Michael

[bcroe]

I'm trying to power a device directly from a solar panel (no battery) and having some trouble determining just how to do that.

My problem is that solar panels generally fit well with battery voltages, but not so much with fine electronics that run at 5V. That's nowhere near the Vmp for most panels so I am not getting a lot of efficiency from my panels.

I have tried various configurations with buck converters and so far have not had a lot of success. Problem is that even with an input cap the buck will crash the panel when the electronics begin to consume more than the panel produces.

It seems that it can do some MPPT magic, and I'm guessing that it will somehow allow me to dial in the Vmp, but just how it is supposed to work without a battery on the output side I don't know.

Do any of you smart guys have any experience with this unit, or similar units, that have these "MPPT" settings? What exactly does that setting do, and are these converters superior to regular buck converters?

It is a tiny unit but apparently capable of 3A output, which is enough for my needs. The major thing about that unit is that it has an UVLO on the input side, which should allow me to recover better after a panel crash. But this thing doesn't have a magic "MPPT" dial though. I'm not sure if that means I'm once again going to be missing out on panel efficiency.

Again - NO BATTERY. The electronics I want to power can go off and on according to lighting conditions Regards, Michael

You are up against physics. Your MPPT output power is always is limited to a bit LESS than
the available INPUT power.

Most MPPTs use a buck converter, except the feedback is from the input side instead of
the output side. It finds the panel best operating point, and tries to dump all available
energy into a large sink at the output. Trouble is your loads are fixed, not capable of
absorbing whatever energy is available. A battery does that.

A regulated output buck draws just what it needs to keep up the output load voltage. As
load increases or sun decreases, it will just keep pulling more of the panel available current
until the panel V slides below MPPT voltage. At that point the buck converter will not get
enough power, and will collapse the panel voltage trying to.

A problem with these is getting started. Often the buck converter just turns on (shorts) and
waits for the panel current to equal the req'd load current before it starts switching (regulating
output voltage). If the panel can't supply that current, the whole thing stalls. Bruce Roe

[jflorey2]

I have tried various configurations with buck converters and so far have not had a lot of success. Problem is that even with an input cap the buck will crash the panel when the electronics begin to consume more than the panel produces.

Good for you! You have discovered a problem that some people here still don't understand (or don't believe in.)

I have found this unit on eBay: http://www.ebay.com/itm/191927867733

It seems that it can do some MPPT magic, and I'm guessing that it will somehow allow me to dial in the Vmp, but just how it is supposed to work without a battery on the output side I don't know.

No worries; it is just a DC/DC with an adjustable brownout preventer. You set the brownout preventer to some voltage, and the DC/DC will not bring the input voltage any lower than that.

Do any of you smart guys have any experience with this unit, or similar units, that have these "MPPT" settings? What exactly does that setting do, and are these converters superior to regular buck converters?

I have no experience with this particular converter, but have used similar ones. They are not MPPT since they don't track anything; you have to constantly adjust them with a screwdriver if you want to optimize power. But they will work better than off the shelf DC/DC converters.

Again - NO BATTERY. The electronics I want to power can go off and on according to lighting conditions, and I'm not very interested in adding a charge controller and a bulky battery to this setup. I might add a cap or a supercap to handle power spikes on the output side, but that's it as far as energy storage goes.

If you have a 'traditional' charge controller you will also need to charge up the supercap so the charge controller will turn on.

[mberg2007]

No worries; it is just a DC/DC with an adjustable brownout preventer. You set the brownout preventer to some voltage, and the DC/DC will not bring the input voltage any lower than that.

So what does it do when the voltage drops below the set point? Does it work like an UVLO and just kind of "throttle" the output power?

I have no experience with this particular converter, but have used similar ones. They are not MPPT since they don't track anything; you have to constantly adjust them with a screwdriver if you want to optimize power. But they will work better than off the shelf DC/DC converters.

Yeah I figured the MPPT was a pretty ambitious term to describe what it does

So how do you set the minimum input voltage? Do you present it with a voltage, then measure the voltage drop as you turn the little dial, until it sits at whatever voltage you believe is the Vmp for your panel?

If you have a 'traditional' charge controller you will also need to charge up the supercap so the charge controller will turn on.

Yes. I took that as a bit of a challenge in a different project, where I ended up with an MPPT charge controller, a supercap and an tiny Arduino Nano with a latching relay that would simply switch the panels directly onto the cap until it measured a voltage higher than whatever the conttroller needed, then switched the panels back to the controller. From there on the controller could handle everything on its own. Worked pretty well (the Nano ran on its own tiny little solar panel)

Frankly I'm surprised that charge controllers don't have this type of bootstrapping mechanism built in. It was dead easy to do with an Arduino Nano and could probably have been done with discreete electronics in an even simpler way.

-Michael

[Mike90250]

I'm not sure what the nonsense is about super caps in Charge controllers is, and needing to charge THEM FIRST ? None of the major brands have super caps that need pre-charging to make the controller work.
Most controllers since the 80's have circuitry in them to operate the controller, That circuit is powered by the battery and should be connected first so the controller can boot up, see what voltage it should Auto-Set to, and then you can apply the PV power.
Sometimes in controllers and inverters, there is a ordinary electrolytic cap that holds a charge and can spark the final connection making or breaking it. Even car stereos have input filter caps. But don't mistake them for super caps.

[bcroe]

I guess he's saying, that type buck wouldn't start until a battery type voltage was present at the output, but
with no battery a cap could be charged by the panels, then switched over to the output to get started. Bruce Roe

[Sunking]

Well I think what is throwing him is he does not understand solar panels are current source and not voltage sources. My guess is whatever he is trying to design\build based on a voltage source rather than a current source. He is trying to reinvent the wheel or does not know he can buy a Current Booster.

To the OP, you have to take a completely different approach.

First thing is to know your motor FLA (full load current). Let's say 10 amps on a 12 volt Series DC Motor.
Second step is to select a panel wattage. The minimum wattage in this example is motor power or 12 volts x 10 amps = 120 watts. Bu tit would be foolish to select a 120 watt panel, because your motor would only run at full power for just a few minutes around Solar Noon. You want at least twice so let's say 150 watt panel.

The design is simple and straightforward. You design a 10-amp Current Source with an operating input voltage of 1 to Vmp of you panel.

The circuit will start pumping 10 amps into the motor after sun rise and the panel can produce at least 10 to 15 watts. Initially the motor voltage will be very low around 1-volt 10 amps (10 watts). with motors torque is current. You have to develop torque to get the shaft to rotate. Rotation speed or RPM are voltage. So as the Sun grows stronger the voltage will rise and the RPM's will increase, thus more power. When your panel is capable of generating 120 or more watts, the motor runs at full power.

The second design goal is to not allow the current booster voltage exceed the motors voltage limit. Going back where I said use a larger panel wattage than the motor rating can bite you if you do not limit the output voltage. Once your panels outpower exceeds motor max power, the voltage would continue to climb higher and higher. At 250 watts with 10 amps around 24 to 25 volts or twice the voltage. Your motor would burn up of fly apart from excessive RPM. So you have to limit the voltage to limit RPM.

FWIW no Super Caps needed.

[jflorey2]

So what does it do when the voltage drops below the set point?

The output voltage drops until the output power drops to match the solar input.

So how do you set the minimum input voltage? Do you present it with a voltage, then measure the voltage drop as you turn the little dial, until it sits at whatever voltage you believe is the Vmp for your panel?

Basically yes. Then the panel warms up and you have to re-adjust it. (Or get a panel much larger than you need and set the voltage to a lower voltage than you expect to see.)

Frankly I'm surprised that charge controllers don't have this type of bootstrapping mechanism built in.l

It's just plain not needed on most installations.

[mberg2007]

I'm not sure what the nonsense is about super caps in Charge controllers is, and needing to charge THEM FIRST ? None of the major brands have super caps that need pre-charging to make the controller work.
Most controllers since the 80's have circuitry in them to operate the controller, That circuit is powered by the battery and should be connected first so the controller can boot up, see what voltage it should Auto-Set to, and then you can apply the PV power.
Sometimes in controllers and inverters, there is a ordinary electrolytic cap that holds a charge and can spark the final connection making or breaking it. Even car stereos have input filter caps. But don't mistake them for super caps.

Hello Mike,

I'm not sure if your reply is for me or not, frankly it makes no sense to me in context of my original question. But I'll be happy to comment anyway.

Super Caps have all sorts of advantages over batteries - much less complicated to maintain, infinite lifetime, very quick to charge, no messy chemistry to be aware of, easy to dispose of. Etc. So for many it makes sense to try to use them in stead of batteries. But most Super Caps have a tendency to leak current much faster than batteries, so after a good long night you often find that the cap voltage has dropped below the point where it can power the charge controller. The result is a completely dead installation that will not start again until the voltage on the cap goes up above whatever the controller needs - 9V for example.

To overcome that one can temporarily connect the panels directly to the cap, wait for the voltage to go up to 9-10V, then connect the panels to the controller and you're all set for the rest of the day. And in my case I got an Arduino Nano with its own little panel and a relay to do that work for me.

-Michael

[mberg2007]

The output voltage drops until the output power drops to match the solar input.

The output voltage drops? So even if I have set it to output for example 5V/2A, then if the input voltage drops below the set MP voltage, the unit will lower the output voltage below 5V?

Seems like it could just kill the output entirely since whatever it is driving probably won't work below the configured output voltage. But I guess it could work anyway.

Basically yes. Then the panel warms up and you have to re-adjust it. (Or get a panel much larger than you need and set the voltage to a lower voltage than you expect to see.)

Yes I know the Vmp will float around a bit but as long as I'm kind of in the ballpark, it's still much better than just wiring the panel to a supercap and from there on to a buck converter. Because that doesn't recover once the panel has collapsed.

-Michael

[mberg2007]

Well I think what is throwing him is he does not understand solar panels are current source and not voltage sources. My guess is whatever he is trying to design\build based on a voltage source rather than a current source. He is trying to reinvent the wheel or does not know he can buy a Current Booster.

What brought you to the conclusion that I don't understand how solar panels work? If you read my initial post, you'll find that I'm posting here precisely because I do understand that using a 60V panel with a Vmp at around 44V to drive a buck converter that outputs at 5V isn't a good idea.

I completely get that panels are current sources below Vmp. I know what an IV curve looks like.

First thing is to know your motor FLA (full load current). Let's say 10 amps on a 12 volt Series DC Motor.

Motor? Who said anything about a motor? I'm not powering a motor, I'm powering a fairly light load consisting of an Arduino and a Raspberry pi. Plus a few shields and a sensor.

Second step is to select a panel wattage. The minimum wattage in this example is motor power or 12 volts x 10 amps = 120 watts. Bu tit would be foolish to select a 120 watt panel, because your motor would only run at full power for just a few minutes around Solar Noon. You want at least twice so let's say 150 watt panel.

Tell me how you connect the panel to the motor then. Assume the panel puts out 20V rather than 12V, and that the Vmp is 18V. You want to connect that to your 12V motor? You'll need much more than a 120 watt panel then.

The design is simple and straightforward. You design a 10-amp Current Source with an operating input voltage of 1 to Vmp of you panel.

Design a Current Source? Are you talking about a buck converter here? What current source (other than the solar panel) is involved in your design?

The circuit will start pumping 10 amps into the motor after sun rise

If the motor requires 12V then you have unrealistic expectations. The panel will not be producing anywhere near 120 watts at sunrise.

I think you must have somehow misunderstood my question. A motor is a very convenient way to burn sunpower because it can operate at various voltages. But a Raspberry computer cannot do that. It needs 5V all the time, which is why you place something like a buck converter between the panel and the computer. The problem with that setup is that the buck computer will collapse the panel as soon as it cannot deliver the power that the raspberry requires, for reasons that are evident once you understand how a buck converter works. It will not recover from this situation easily.

The conventional solution is to use an UVLO (Undervoltage lockout) mechanism in the buck converter. This essentially stops the buck converter from trying to pull more power from the panel once the panel voltage drops too low. The panel is then able to recover and the buck will resume conversion. Sure, the Raspberry will die during that non-conversion period but this is acceptable for my project. What is not acceptable is that a cloud will crash the whole system until the next day.

If you have any good suggestions for how to solve that, and specifically if you can shed some light on the two buck converters I linked to in my initial post, then I'd be glad to hear about that. But talk about motors doesn't help.

-Michael

[Mike90250]

Super Caps are NOT batteries.
I'll repeat again, they are NOT batteries.

They will, for a short, limited time, behave like batteries, but only for the upper 5% or 10% of their power power curve that intersects battery voltage.
Once below standard battery voltage LVD (11v, 22v, 45v) the gear powered by them will shut down or go bonkers because of low voltage, Still gobs of
un-useable power in the cap, but it cannot work as well as a battery,

Yes, some have modded a car battery case to hold super caps, but you get 1 chance to start the engine, and if you have left a light on, you have no juice in 30 minutes,

Any time you try to use a super cap as a battery, you WILL have an unhappy surprise when you expect it to perform like a battery. Like the overnight drain of a charge controller will bleed a super cap down.

I shout, because I don't want neophytes to be sucked into thinking that caps will solve all the problems, like diodes.

rant mode off..

[Sunking]

Motor? Who said anything about a motor? I'm not powering a motor, I'm powering a fairly light load consisting of an Arduino and a Raspberry pi. Plus a few shields and a sensor.

My bad, some how I got the idea you were trying to run a pump.



Tell me how you connect the panel to the motor then. Assume the panel puts out 20V rather than 12V, and that the Vmp is 18V. You want to connect that to your 12V motor? You'll need much more than a 120 watt panel then.

Design a Current Source? Are you talking about a buck converter here? What current source (other than the solar panel) is involved in your design?



If the motor requires 12V then you have unrealistic expectations. The panel will not be producing anywhere near 120 watts at sunrise.

Exactly like any Linear Current Booster they sell to run 12 volt motors. No need to go into that as it does not fit your application.

All you need is a very simple Voltage Regulator. It can be either a Linear Series VR or a switching VR that can operate from any input voltage from Voc down to about 1.5 above output voltage. So if you use say a 12 volt battery panel and the load is 5 volts, 6.5 volts to 22 volts. Piece of cake, it can be with a simple 3-terminal 5-volt VR like a 7805.

However once the panel output power is less than what the load demands, it will crash no doubt about it. That is what is suppose to happen. That is why batteries are used. If you insist on not using a battery, OK there is a way. For example if you load is say 5 watts at 5 volts, just use a large panel like 100, 200, or a 1000 watts with a LM7805. That way just after sun rise the regulator will come to life and will last to almost sun set when the panel voltage dips below the 6.5 volt threshold. If you used say a 10 watt panel, you would only get an hour or two of operation around solar noon, and if any clouds or shade occur during that time, it will crash. But if you used a much larger panel, even moderate shade or clouds can still generate 5 watts.

Easy peasy.

[jflorey2]

The output voltage drops? So even if I have set it to output for example 5V/2A, then if the input voltage drops below the set MP voltage, the unit will lower the output voltage below 5V?

Correct. And it will continue to drop until the input goes back into regulation.

Seems like it could just kill the output entirely since whatever it is driving probably won't work below the configured output voltage.

That's what most DC/DC converters do; it's a function called UVLO. (undervoltage lockout.)

Since many loads (lights, fans etc) will reduce their current draw as voltage drops, reducing the output voltage, rather than cutting it completely, can work to reduce total power.

[jflorey2]

Super Caps are NOT batteries.
I'll repeat again, they are NOT batteries.

They will, for a short, limited time, behave like batteries, but only for the upper 5% or 10% of their power power curve that intersects battery voltage.
Once below standard battery voltage LVD (11v, 22v, 45v) the gear powered by them will shut down or go bonkers because of low voltage, Still gobs of
un-useable power in the cap, but it cannot work as well as a battery,

Everything you have said is correct except the part about "5 to 10%."

Energy is 1/2CV^2. So let's say you want to use a Maxwell BMOD0058 ($150) to replace your boring gel cell battery. Since the BMOD is good to 16 volts you set your charge controller to 16 volts. Since your inverter works down to 10.5 volts before giving up that becomes your lower limit. That means when fully charged your cap stores 7424 joules of energy - and you can discharge it down to 3197 joules. So you have been able to use 4227 joules (watt-seconds) or 58% of the energy in the capacitor. You can actually use a greater percentage than you should be drawing from a lead acid.

However, here's where it starts not being such a good deal:

The smallest 12V I could find out there was a 1.2 amp-hour BP1.2-12-T1 for $12. That will give you approximately 12 volts at 1.2 amps for about 36 minutes if you discharge to 50%. That's 31,000 watt-seconds or 31,000 joules. Per unit energy, the gel cell is 91 times more cost effective. So unless someone has a truly bizarre requirement (i.e. maintenance at 2 years is unacceptable but maintenance at 10 years is OK, or required discharges to 100%) the cheaper battery is the way to go.