What would the best possible charge controller be like??

[Ben]

Seems msb controllers are having "problems"
From every thing i know of msb is the only "cheap" controllers
But are they truly cheap if all info is considered?

Could a higher voltage controller help pay for its self by saving copper wire cost?
Is there a better way than MPPT?
A better than MPPT or not as good but simple way of turning high voltage to low voltage?
Nothing is 100% efficient but how close can a controller be?
To the dreamers if you could have it do anything explain what that would look like?
What features are must haves?

From the way its looking msb may not be a good the best way to go

So most people like me dont have a clue what to even look for to even know what is a good deal!

I hope this post will make it easy to make a better choice!

 I hope everyone who reads this will add there TWO CENTS from what they use to what would be nice to have!
Even if some one has already said what you would say please say you second what has already been said! That will let the reader !
I hope the pros from all around or under the world chime in along with reg people like myself and definitely the dreamers!

Dreamers are the start of ALL the good stuff even though the pros would like to kick them in the head to shut them up! LOL LOL
I will start this by saying wire can add a very big cost to off grid solar.   

How much money can be saved by high voltage panel setup?

[Sid Genetry Solar]

3 hours ago, Ben said:

Could a higher voltage controller help pay for its self by saving copper wire cost?

Very subjective...depends on how much more the "higher voltage controller" costs!

Worth noting that what's most fun to calculate is the percentage of power loss across your solar wires (wire resistance per 1k feet against the solar panel "short circuit amps" for a worst-case calculation).  The longer the wires between the panels and MPPT, the higher your losses here will be.  Bigger wire is obviously considerably more expensive--and at some point, you will break even with a higher voltage MPPT.

 

I find the MSB's voltage rating to be very limiting especially with the newer higher wattage panels, which run significantly higher voltages than the older ones.  You simply can't put more than 2 of the bigger panels in series with an MSB--which puts you barely above battery voltage (if 48v).

 

3 hours ago, Ben said:

Is there a better way than MPPT?

No, not for solar panels anyway.  In short, solar panel current remains very stable (based on the amount of light), while the voltage can be taken from no load to a short circuit--meaning that the "maximum power point" is going to be at the highest voltage.

The reason "solar panel voltage" is not a defined constant, is because it is highly variable based on the solar panel temperature--so a "sweeping maximum power point tracking" method is the quickest and simplest way to try to get the most power out of a panel.

 

3 hours ago, Ben said:

A better than MPPT or not as good but simple way of turning high voltage to low voltage?
Nothing is 100% efficient but how close can a controller be?

So "MPPT" is the term used for the regulating methodology.

The actual conversion topology is generally a non-isolated synchronous buck converter.  (Some random smaller Chinese units sometimes are designed around boost topology--but that's highly specific to a particular small usecase.  And of course, most uber-cheap Chinese "MPPT" units are just outright fakes, with no buck conversion circuitry whatsoever.  Basic PWM chargers.)

Most "industry standard" MPPT units run >99% efficiency, often reaching or exceeding 99.6%.  For example, my Morningstar Tristar TS-MPPT-60 will run a happy 3,200W at full load--with no fans, and a heatsink that's warm (but not scorching hot) to the touch.  If 99.6% efficient, that's a paltry 12W of heat.

Worth noting that the enemy of MPPT is amperage--which is why they tend to limit out at 60/80/100A.  The more current, the more heat generated--which is why an MPPT that can do 3,200W at 48v can only do 800W at 12v.  Same 60A.

 

3 hours ago, Ben said:

From the way its looking msb may not be a good the best way to go

They are a cheap way to get started, for sure.  I can't say that their MPPT function is the world's best (or anything close to it)--but a lot of people seem to like them.  I personally have never used one (I have Morningstar Tristar and Epever Tracers), as I prefer MPPTs without fans if I can!

[Ben]

with msb one and one half batt voltage is what your shooting for on panel input to controller   

why so low why not 4 or 500 volts on the solar input to the controller is it just a safety issue   

 could a controller work like a low freq inverter but stepping voltage down not up with a steel core tranny 

seems like the high amp stuff could be done by a tranny

a tranny can be very efficient i know the cost of low freq tranny would be crazy high just wandering 

[Ben]

If no shading would high voltage  on controller input charge batt in lower light or is there not much benefit?   

As always i am full of questions but i think others would be wondering also so not a waste if answered where all can freely read 

[dickson]

I find the MSB's voltage rating to be very limiting especially with the newer higher wattage panels, which run significantly higher voltages than the older ones.  You simply can't put more than 2 of the bigger panels in series with an MSB--which puts you barely above battery voltage (if 48v).

I  try connecting  4  solar panets at 60 VOC  in series and parallel   and the MSB  blow up  and melted the terminal  it got so hot in the summer .     My  other  6  MSB  been working good  with  just 2 solar panels connected  and  producing  under 500 watts per hour for each  MSB  for 5 hours of sunlight  now .       One MSB been working for 3 years .    

[Sid Genetry Solar]

12 hours ago, Ben said:

with msb one and one half batt voltage is what your shooting for on panel input to controller   

why so low why not 4 or 500 volts on the solar input to the controller is it just a safety issue 

Firstly, RdsOn (on resistance) in FETs significantly increases with higher-voltage parts.  Higher on resistance = more heat at the same current.  And the more heat generated, the lower the system efficiency.

IGBTs are generally used for high voltage switching--as their "saturation voltage" can often result in lower losses than a comparable MOSFET.  Worth noting that SiC-based FETs (silicone carbide) have significantly narrowed the gap between MOSFETs and IGBTs.

 

12 hours ago, Ben said:

 could a controller work like a low freq inverter but stepping voltage down not up with a steel core tranny 

seems like the high amp stuff could be done by a tranny

a tranny can be very efficient i know the cost of low freq tranny would be crazy high just wandering 

Cost, weight and efficiency demands for an MPPT pretty much rule out low-frequency iron-core transformers!  Worth noting that the one thing an MPPT cannot provide (because of the solar panels!) is any sort of surge power.  This makes them ideal for HF-style designs.

I've looked into the possibility of making a GS MPPT--but unfortunately due to the PJ fiasco, I've got CONSIDERABLY more important things on my plate right now!

In short, my plans specced a dual-input 200-400v panel voltage range, utilizing a forward/flyback (isolated) topology, and providing 4 isolated "12v" outputs that could be series/parallelled for full output at 12v, 24v, or 48v if desired.  (No other MPPT on the market offers this.)

Due to the topology, it wouldn't work with low voltage panel systems--it just won't turn on.  And yes, it would be based around an HF core design to significantly reduce wire losses (which in my experience account for more than 90% of the losses of an LF transformer!)

Just think about it: if you need 200 turns of wire around a large LF core, that's a pretty significant resistance.  But if you can take that same thickness of wire...you'll only need 20 turns on a suitable HF core--meaning that you now have 1/10th the resistance.  And that means that you have 1/10th the losses--well, apart from the black magic of high-speed switching.  Which among other things includes the "skin effect" of wires starting to become hugely important at higher frequencies!

 

But like I've said, I've got MUCH bigger things that need attention right now!

[Warpspeed]

As stated above, high voltage mosfets are miserable devices with much higher Rds On.

Suppose you have a mythical buck converter that transforms 500v down to 50v at 50 amps.

Current at the high voltage input might then be only 5 amps (assuming an impossible 100% efficiency).  The problem is, the mosfet has to supply 50 amps and also switch at 500 volts.  It gets hit with the worst extremes of both voltage and current.  There are two choices, either build something that is going to get very hot and be inefficient, or build something very expensive, with a rather large number of high voltage mosfets. 

Neither approach has much appeal, either technically, or commercially which is why its not done that way.

 

[Sid Genetry Solar]

Totally agreed, for the typical non-isolated synchronous buck converters.

My thought was to utilize a forward/flyback topology with synchronous rectification on the output instead.  This topology is commonly used in SMPS units easily exceeding 2kw per unit (think server rack PSUs).  Most often, they internally run from a 400v (PFC) DC bus...using a HF transformer to do the "dirty work" dropping 400v down to the 12v @ 220A continuous output.

Yes, efficiency on those units generally is just under 90%--which is a considerable efficiency loss when normal "low voltage" MPPTs tend to near 99% efficiency levels.  (Pretty simple if you think about it I guess--the less "work" being done, the more efficient it likely is to be!)

 

It is definitely possible to attain reasonable efficiency at said levels...as Morningstar Corp offers a 600v MPPT with a "peak" efficiency of 97.9%.  (Of course, "peak" is by no means indicative of "average" or even "full load"...gotta love marketing!)

[Sid Genetry Solar]

Hey @Ben here you have proof that I'm not @Warpspeed.......😉

[Warpspeed]

I post on several different Forums as Warpspeed. 

I am me and nobody else.

[Warpspeed]

56 minutes ago, Sid Genetry Solar said:

Totally agreed, for the typical non-isolated synchronous buck converters.

My thought was to utilize a forward/flyback topology with synchronous rectification on the output instead.  This topology is commonly used in SMPS units easily exceeding 2kw per unit (think server rack PSUs).  Most often, they internally run from a 400v (PFC) DC bus...using a HF transformer to do the "dirty work" dropping 400v down to the 12v @ 220A continuous output.

Yes, efficiency on those units generally is just under 90%--which is a considerable efficiency loss when normal "low voltage" MPPTs tend to near 99% efficiency levels.  (Pretty simple if you think about it I guess--the less "work" being done, the more efficient it likely is to be!)

 

It is definitely possible to attain reasonable efficiency at said levels...as Morningstar Corp offers a 600v MPPT with a "peak" efficiency of 97.9%.  (Of course, "peak" is by no means indicative of "average" or even "full load"...gotta love marketing!)

I have not given any of this much thought, but one way to do this would be to build a "voltage reduction front end" to drive a conventional MPPT controller of your choice.

Step down the very high solar voltage to something much lower, say 600v down to 150v or something, whatever is required.

An efficient way to do that would be with a fixed duty cycle forward converter and a 4:1 transformer.  No mppt, no control of any kind required.  Just a 4:1 (in this example) voltage reduction. The forward converter always operates at high voltage and low current, and the transformer always operates at full duty cycle, so it could all be made to operate at reasonably high efficiency.

The output current is increased x4 (in the above example), and that is done via a transformer. All the clever mppt stuff could be done after that.

Not too difficult or complex.  Any voltage ratio would possible. Maybe 800v down to 18v ?

[Sid Genetry Solar]

31 minutes ago, Warpspeed said:

I post on several different Forums as Warpspeed. 

I am me and nobody else.

I was never questioning your identity...just noting that Ben has multiple times asked me if I'm not somehow Warpspeed.  I keep assuring him that I'm not--so there's nothing like a thread with Ben, theRealWarpspeed and myself in it for proof....

[Sid Genetry Solar]

16 minutes ago, Warpspeed said:

I have not given any of this much thought, but one way to do this would be to build a "voltage reduction front end" to drive a conventional MPPT controller of your choice.

Step down the very high solar voltage to something much lower, say 600v down to 150v or something, whatever is required.

An efficient way to do that would be with a fixed duty cycle forward converter and a 4:1 transformer.  No mppt, no control of any kind required.  Just a 4:1 (in this example) voltage reduction. The forward converter always operates at high voltage and low current, and the transformer always operates at full duty cycle, so it could all be made to operate at reasonably high efficiency.

The output current is increased x4 (in the above example), and that is done via a transformer. All the clever mppt stuff could be done after that.

Not too difficult or complex.  Any voltage ratio would possible. Maybe 800v down to 18v ?

Definitely a feasible idea, one that I hadn't considered either.

One thing to keep in mind is that the more "conversion levels", the lower the total system efficiency--as a "voltage reduction front end" would also still require switching -> filtered DC rectification on the output circuit.  Use of a standard rectifier diode is very inefficient at any significant current levels (at least by MPPT standards!), requiring huge heatsinks, etc.  Synchronous rectification is a fun little trick to implement--but drastically more efficient!  (Thinking <40mV drop instead of >1,200mV at 100A!)

And if an efficiency of a "voltage reduction front end" of >95% could be achieved, there's no reason it couldn't be implemented as an MPPT.  Coding an MPPT can't be that hard...

[Ben]

i never thought you was warpspeed   

warpspeed is a truly awesome person old school cool to the max!! 

i thought you may have been poida 

 poida and you sid are truly gifted and have a way of making me understand unlike anyone else i have ever known   

 warpspeed would lead me to the water over and over and over and over but he would not drink it for me 

i think he wanted me to do things for myself so i could better understand

i truly appreciate warpspeed them guys down under are hard to beat  

you have a man with a  life time of wisdom chiming in on here 

 it makes me very happy to see warpspeed on here

Hey warp ALL my esg002 inverters are running strong you i have had zerro problems!

One of them is my buddys only power running day and night 24/7 

[Ben]

Thanks for you in put dickson i have to argree with you     

We now have the worlds best old school and the worlds best new school thinking things through

WOW WOW WOW!! 

  Some thing great will have to happen!!

[Warpspeed]

I too really admire Poida (Peter Burtles) from The Back Shed forum.

We are both in Melbourne but have never met.

[Ben]

i did truly think Sid was poida for a very long time 

 

[Ben]

If any one likes the charge controllers they use i and others would like to hear about it 

There is not any "cheap" mppt charge controllers besides msb that i have seen 

I guess even some of the ones that cost the most can be even cheaper than msb

That is if they are built to last and have great customer service for parts or repair  incase you have problems 

I have read horror stories about the best name brand products nowadays 

I dont like to gamble if some thing cost more than around 100 bucks but to be fair what can 100 bucks by now days?

I guess 5 or 600 bucks for a charge controller is still kind of cheap if it lasts 10 or more years?

There is people with way more wisdom on this subject than me that could school us all and maybe already been trying lol       

[dickson]

There is not any "cheap" mppt charge controllers besides msb that i have seen

YES     I  only connect 2  solar panels per MSB   and no problem for 3 years .     Aliexpress  sell MSB now .       

[Sid Genetry Solar]

7 hours ago, Ben said:

If any one likes the charge controllers they use i and others would like to hear about it 

There is not any "cheap" mppt charge controllers besides msb that i have seen 

I guess even some of the ones that cost the most can be even cheaper than msb

That is if they are built to last and have great customer service for parts or repair  incase you have problems 

I have read horror stories about the best name brand products nowadays 

I dont like to gamble if some thing cost more than around 100 bucks but to be fair what can 100 bucks by now days?

I guess 5 or 600 bucks for a charge controller is still kind of cheap if it lasts 10 or more years?

There is people with way more wisdom on this subject than me that could school us all and maybe already been trying lol       

The MPPTs I'm currently recommending--with some caveats--are the Epever Tracer xxxx-AN series.

The pluses:

• Fanless design.
• Nicely designed, with comm interface support built right in.
• true MPPT tracking

The only minus:

• very slow regulation: you'll need to set the "high voltage cutoff" to prevent it from overvoltaging the batteries.
• requires the MT-50 meter (or computer connection) to set it up.  This is not unusual.

[Warpspeed]

Just to kick start this thread back into life..........

The perturb and observe MPPT algorithm is generally accepted as the best approach to tracking solar panel loading over the full operating envelope of our solar panels.

It is not without its problems, such as locking onto false peaks caused by partial shading, and slow response to massive changes produced by passing fluffy white clouds in an otherwise clear blue sky.  It does have the advantage of being reasonably idiot proof, as the MPPT software will try to find a peak with unknown or miss matched panels.

I was intrigued by all this, and a few years back decided to try a method of power tracking that was totally different and a lot simpler. To cut a long story short, what I did was build a simple buck converter that I could control the duty cycle manually, with a potentiometer from 0% to 100% duty cycle.  This was hooked up to one 30v panel and a 12v car battery, all monitored by a Turnigy power meter. The power meter was on the solar side of the buck converter and displayed volts, amps and watts.

The original aim was to test and compare various panels I had here at the time.

I was rather surprised to find that I could adjust the solar panel voltage over a much wider range than I had expected, and the measured power remained much more constant than I had expected.  Sure, there was a definite point of maximum power, but it was not the peak I was expecting, but a very broad almost flat hump, and the power maximum hardly changed over a very wide operating voltage range.  The power maximum was always pretty close to the max power voltage printed on the rating plate of the panel, over pretty much everything from twilight to mid day full sun.

Power varied hugely over a full day, but the trick is to optimally load the solar panels just enough to hold the solar panel voltage constant.

If the sun gets blocked by a cloud, reduce panel loading, in full sun increase panel loading, its as simple as that.

So what I did was build a buck converter with feedback that adjusted the duty cycle to hold the solar voltage constant. In effect a simple straightforward PWM shunt voltage regulator.  The same regulator had a second control loop that regulated the battery voltage to a maximum charging voltage.

In effect, it would bulk charge up to the set battery voltage holding the solar voltage constant. It would then taper the charging current down to zero in the usual way, allowing the solar voltage to rise as the duty cycle throttled back to zero.

This worked so very well, I decided do some back to back testing with an MSB solar controller, using two Turnigy power meters, two identical halves of my main solar array feeding into the same battery.  To my utter amazement there was no measurable difference in performance during bulk charging under everything from full sun in a clear sky to total evil grey cloud cover.  I did swap solar panels and power meters around, but the results were always within a very few percent of each other. Maybe two or three percent.

My biggest problem was that it was impossible to set both controllers to identical output voltages. Even just a few millivolts difference would cause one controller to stop charging while the other continued to trickle charge at very low current for several hours, which skewed the total amp hour readings.  Its only a fair test when bulk charging.

Anyhow, my conclusions were that the biggest difference in efficiency was more due to to the power components than the charging algorithm. I could make my buck converter either more or less efficient than the MSB controller by changing mosfets and choke, and that was reflected by heating of the controllers.

The MSB perturb and observe was definitely superior in extremely poor solar conditions.  The difference was something like 15 watts from the MSB controller and 10 watts from my controller from 1Kw worth of panels in twilight conditions.  I don't think that really matters in the great scheme of things.

On the other hand, in rapidly changing solar conditions, such as passing clouds, my controller was much faster.  Milliseconds versus several seconds to adjust.

It might be argued that a proper perturb and observe software algorithm with a two dollar microcontroller is still better and costs no more.

My circuit has the advantage of extreme simplicity, its easy to home build, easy to understand, and easy to diagnose and repair. Its just a standard analog PWM control chip and a few resistors and capacitors. The power components would be the same in either case.

If you need a 100 amp controller, its just as easy to build ten 10 amp controllers and run one per series string of panels. Also, you can build it with 600 volt (or higher) voltage mosfets if you want, plenty of choices on how to build something like this.

This may be of interest to some of you here, but my experience on the Forums is that the majority would rather just buy cheap Chinese and grumble about all the problems and disadvantages.

Below is a picture of the very first prototype board during construction. It was built to be very easy to fix if it blows up. Mosfets  can be changed with a screwdriver, and the whole control board just plugs into the main wiring and power components.

I am still running my MSB controllers  right now because they are still working.  If they die or blow up, they will definitely be replaced by my own constant voltage controller.

Too many other projects on the go here right now, this is now on the back burner, but I will definitely come back to this.

 

 

The MSB

[dickson]

Look like an easy way to connect mosfets  without soldering  the mosfets  .  

[Sid Genetry Solar]

7 hours ago, Warpspeed said:

I was rather surprised to find that I could adjust the solar panel voltage over a much wider range than I had expected, and the measured power remained much more constant than I had expected.  Sure, there was a definite point of maximum power, but it was not the peak I was expecting, but a very broad almost flat hump, and the power maximum hardly changed over a very wide operating voltage range.  The power maximum was always pretty close to the max power voltage printed on the rating plate of the panel, over pretty much everything from twilight to mid day full sun.

Power varied hugely over a full day, but the trick is to optimally load the solar panels just enough to hold the solar panel voltage constant.

If the sun gets blocked by a cloud, reduce panel loading, in full sun increase panel loading, its as simple as that.

So what I did was build a buck converter with feedback that adjusted the duty cycle to hold the solar voltage constant. In effect a simple straightforward PWM shunt voltage regulator.  The same regulator had a second control loop that regulated the battery voltage to a maximum charging voltage.

In effect, it would bulk charge up to the set battery voltage holding the solar voltage constant. It would then taper the charging current down to zero in the usual way, allowing the solar voltage to rise as the duty cycle throttled back to zero.

This worked so very well, I decided do some back to back testing with an MSB solar controller, using two Turnigy power meters, two identical halves of my main solar array feeding into the same battery.  To my utter amazement there was no measurable difference in performance during bulk charging under everything from full sun in a clear sky to total evil grey cloud cover.  I did swap solar panels and power meters around, but the results were always within a very few percent of each other. Maybe two or three percent.

I had actually noticed exactly the same thing...not from real life, but from reading solar panel datasheets.  I noticed that the maximum power point voltage was pretty much the same across the entire light range.

This gave me a naughty idea: if the user could enter the solar panel's datasheet max-power-point voltage (and temperature coefficient) into the settings of a GS MPPT, some software trickery could be used to not only handle very quick regulation via CV (constant-voltage) methodology, there was another wrinkle that could be thrown in.  If the code could reasonably tightly pin down the actual max power point of the panels, we could utilize the panels' temperature coefficient to literally determine the panels' surface temperature!  Once the temperature is known, we could use other coefficients to determine how close to max generation power the panels actually are (i.e. calculate solar irradience)....on and on it goes.

Just for clarity in how much I can string math together to generate useless stats, there's a "Transformer Efficiency" stat on GS inverters--the inverter calculates the losses in the transformer based on the measured (and calculated) voltages.  Curious...yes.  Useful...not so much 😉.

 

7 hours ago, Warpspeed said:

If you need a 100 amp controller, its just as easy to build ten 10 amp controllers and run one per series string of panels.

To some extent...though from a manufacturing perspective, ten 10A separate controllers would have a LOT of redundant parts and significantly higher total.

 

7 hours ago, Warpspeed said:

Also, you can build it with 600 volt (or higher) voltage mosfets if you want, plenty of choices on how to build something like this.

Just my personal 2 cents...when dealing with high voltages like that, IGBTs often present a much lower total loss.  MOSFETs are good for lower voltages (<150v or so), whereas IGBTs often have lower losses when dealing with higher voltages (>200v or so). 

It's funny how that works...I've been asked several times "why don't you just use one of those 600A IGBT blocks for a 12v inverter?"  Losses are literally unfathomable when compared to 40v MOSFETs!  Yes, they're used in electric cars to switch insane amounts of power--but they're also liquid-cooled for a reason!  Their series losses are the same if you're switching 12v @ 600A as if you're trying to switch 800v @ 600A.  One of those two is significantly more power--but the DC loss will be the same!  (Switching losses obviously will be significantly higher at the 800v mark, but that's besides the point at hand.)

 

8 hours ago, Warpspeed said:

This may be of interest to some of you here, but my experience on the Forums is that the majority would rather just buy cheap Chinese and grumble about all the problems and disadvantages.

Hear 'ya there 😉.  I've started on theory for a GS MPPT--but due to everything else significantly more important, that's way down on the project priority list at the current time.

It'd be a 100A size--and the challenge of course is how to do that without generating a heap of heat.  And being unique enough to stand out in a market saturated with Chinese products.

 

8 hours ago, Warpspeed said:

Milliseconds versus several seconds to adjust.

I have a feeling that the reason a lot of the cheap Chinese controllers are SO SLOW is because they're using long-term averages as a Band-Aid over sloppy regulation code, to avoid regulation oscillations.

 

8 hours ago, Warpspeed said:

Anyhow, my conclusions were that the biggest difference in efficiency was more due to to the power components than the charging algorithm. I could make my buck converter either more or less efficient than the MSB controller by changing mosfets and choke, and that was reflected by heating of the controllers.

I would think that a synchronous buck converter--though much easier to blow out--would be far more efficient at higher currents.  At lower currents (i.e. 10A), a diode does function pretty well (though dissipating an easy ~10W of heat alone)...but at higher currents, you kinda HAVE to go synchronous just to keep the losses down!

Playin' with the choke...that's a "no experience" area for me, though it literally is the heart of the MPPT!  Weird stuff like switching frequencies, skin effect, ferrite composition....

[Warpspeed]

Agree with everything you say Sid !

This constant solar voltage idea has bugged me for years, so I just had to try it.  My situation is different to most, I am running a 96 volt system, so currents are much lower for me.  The fact I design and build all my own equipment, its no more difficult to do it for 96v as for any other voltage.

If combined max power on the rating plates work out to say 130 volts for example, and you swing the solar voltage regulating adjusting potentiometer between 110v and 150 volts, the power might fall off only by about five percent below the peak at 130 volts.  Correcting that 130 volts for temperature would certainly be possible, but hardly worth the trouble. The non criticality of the actual regulated solar voltage is absolutely astounding.

What DOES matter is loading the panels correctly. Even a small change in insolation has a very large effect on panel voltage at a given duty cycle.  That is the change that must be tracked.

Setting the regulator voltage to 120v, 130v, or 140v makes no real difference.  What matters is keeping the voltage within that range as the power potential  of the panel goes from zero just before sunrise, to peak power at solar noon, and back to zero just after sunset.

I run IGBTs in my Warpverter, and at low power, typically a few hundred watts, the voltage drop across my 200 amp rated IGBTs is quite low.  Only at the multi kilowatt output level do IGBT conduction losses start to become even noticiable.  For a typical domestic inverter, the average power will be quite low compared to the peaks throughout the day, which are never very long lasting.

IGBTs have come a very long way in recent years, and would be well worth considering for a higher voltage solar controller.  After all, IGBTs are now petty universal in higher voltage commercial grid tie inverters.

Synchronous buck converters are definitely the way to go for lower voltages and higher current, but caution is required because a synchronous buck converter is bi directional for power flow when charging a battery.  Current can flow backwards through the choke and the "synchronous" diode, turning a buck converter into a boost converter in the reverse direction.  That particular problem is not widely known or discussed in the literature.  There are ways around that of course, but if writing software, the duty cycle must never become so short that current can reverse through the choke.  The very high voltages created on the solar side can be particularly destructive of mosfets and electrolytics.

Magnetics design has been a particularly interesting area for me, and also pretty important in a buck converter.  Its pretty difficult to make something that is small really efficient. Big may not be beautiful, or cheap, or easy to house, but something that runs cool needs to be larger than what is usually found in most Chinese designs.

Its possible to home build something that works much better, but the cost of the parts can easily exceed the cost of buying a ready made cheap Chinese product.

That is particularly true of Lithium cell balancing circuits.  The manufacturers all seem to be striving to undercut each other in cost, rather than trying to sell something that is genuinely superior in performance.  Unfortunately to make something decent would never sell, everyone is looking for a bargain.

Fastest way to go bankrupt in this industry is to make a solid reliable well performing product that nobody is prepared to pay for.  Everyone just wants cheap.

Even giving full technical details of how to build your own SIMPLE and efficient solar electronics, its all just too much trouble, nobody is interested.

 

 

 

[Sid Genetry Solar]

9 hours ago, Warpspeed said:

Correcting that 130 volts for temperature would certainly be possible, but hardly worth the trouble. The non criticality of the actual regulated solar voltage is absolutely astounding.

I wasn't saying that I'd "correct for temperature", as much as a very, very slight "perturb and observe" from the "base MPP voltage" to try to find the actual peak power point--and then when comparing the actual to the room temperature nominal, we could use the temperature coefficient to calculate the actual solar panel temperature.  A geeky stat to display, with absolutely no usable value 😉.

 

hardware

9 hours ago, Warpspeed said:

Current can flow backwards through the choke and the "synchronous" diode, turning a buck converter into a boost converter in the reverse direction.  That particular problem is not widely known or discussed in the literature.

I actually came across an application note that utilized this exact methodology for a battery system.  I wouldn't be surprised if it's what Tesla uses in their Powerwalls--as to my knowledge, though the output voltage is in the 300-400v range, the internal cell voltage is all of 48v.

9 hours ago, Warpspeed said:

There are ways around that of course, but if writing software, the duty cycle must never become so short that current can reverse through the choke. 

An alternate option is to use a powerful enough PWM module that a fixed "high side FET" pulse width can be used--then it doesn't matter what the PWM duty cycle is.

Or hardware synchronous rectification...completely alleviating the MCU requirements!

 

9 hours ago, Warpspeed said:

I run IGBTs in my Warpverter, and at low power, typically a few hundred watts, the voltage drop across my 200 amp rated IGBTs is quite low.  Only at the multi kilowatt output level do IGBT conduction losses start to become even noticiable.  For a typical domestic inverter, the average power will be quite low compared to the peaks throughout the day, which are never very long lasting.

Well, anything at no load shouldn't generate much heat 😉.  But the number of GS customers who run their 6kw inverters at long-term redline might surprise you.  Me personally, I've designed things to be hyper efficient--so my house inverter rarely runs past 3kw continuous anyhow.  But for everyone else...

My reference to IGBT/FET losses is always based on "max rated load", and calculating losses.  For a "12v" example:

• a set of 12v FETs at 1.4mOhm each / 6 per block = DC resistance of 0.233mOhm total @ 600A = 83.88W.  So across a full H-bridge, we'd have 167W of heat.  Not horrendous.
• pick a 600A IGBT module out of the clear blue sky: CM600HU-12F.  The vCEsat (collector-emitter saturation voltage at full on) is 1.6 to 2.2v.  So 1.6v @ 600A = 960W of heat (2.6mOhm equivalent)--oh, and there's a second one in the an H-bridge power path, so 1,920W of heat in the IGBT modules alone!  On the bright side, this is a 100% efficient space heater that generates a tiny bit of AC output power on the side as a bonus 😉.
  • obviously, losing 3.2v out of 12v is practically devastating for headroom!

If I should flip the math to 400vDC, the IGBT module would be the clear winner over a MOSFET of the same voltage rating. 

 

10 hours ago, Warpspeed said:

That is particularly true of Lithium cell balancing circuits.  The manufacturers all seem to be striving to undercut each other in cost, rather than trying to sell something that is genuinely superior in performance.

Daly BMS and other units that provide a pitifully useless 0.03A balance current--and who ever thought that using 22AWG wires to "steer" huge off-grid battery banks would work??  That's like using a spiderweb to rein a horse. 

But oh boy do those toys sell like hotcakes.

 

more later 😉