rin67630
Solar Enthusiast
What are currently your two power supplies?
Insulating 100-250V input and 12V output?
Insulating 100-250V input and 12V output?
Simple home made analog MPPT contoller
- Thread starter Warpspeed
- Start date
Warpspeed
Solar Wizard
That was explained on the first page of this thread, post #14.
rin67630
Solar Enthusiast
Yes thank you. At your level of panel voltages, one can just use plain switching power supplies like those intended for mains usage, isn't it?
Warpspeed
Solar Wizard
Those small Chinese boards at only a couple of dollars each are hard to beat.
I use them for just about everything these days, both solar and mains powered.
Perfect for a 48v solar controller.
For the 12v and 24v guys running only one or two panels not really suitable, as these need typically more than 35v to 40v.
Solarstefan
New Member
Hi Warpspeed
Just joined the forum, thanks for your superb contribution. I am going to veroboard build your TL494 version.
Should have most bits from scraped electronics.
Just joined the forum, thanks for your superb contribution. I am going to veroboard build your TL494 version.
Should have most bits from scraped electronics.
Warpspeed
Solar Wizard
Great stuff Stefan, and welcome to the Forum 
Solarstefan
New Member
Thank you
Question, could i substitute the fet with an IGBT devices, I guess Rds on is key compared to the sat Vce on the IGBT.
I should work ok just less efficient? I just have millions of IGBT's from wrecked inverters
Question, could i substitute the fet with an IGBT devices, I guess Rds on is key compared to the sat Vce on the IGBT.
I should work ok just less efficient? I just have millions of IGBT's from wrecked inverters
Warpspeed
Solar Wizard
It depends on the particular IGBT and your switching frequency.
IGBTs have come a very long way in recent times. The older ones can be a bit slow, if its fairly recent, should be o/k at 20Khz.
Definitely well worth a try.
IGBTs have come a very long way in recent times. The older ones can be a bit slow, if its fairly recent, should be o/k at 20Khz.
Definitely well worth a try.
Solarstefan
New Member
Sofar so good the solar switcher is born and is working. Initial test with a power supply and a 14V bulb with output current of 5Amps and an input voltage of 40V at 1.9Amp seems to net 92% efficiency, no noticeable heat. Tried a variety of inductors and they perform much the same at that current. Overcast in Brisbane atm, proper solar test has to wait.
Has anyone experimented with different switching frequencies?
Has anyone experimented with different switching frequencies?
Warpspeed
Solar Wizard
That is open to experiment.
Higher frequencies mean in theory you can use a smaller choke with less inductance, but switching losses in your IGBT will probably be higher.
Suggest you get it all working first, at reasonably high power, and then measure some temperature rises at different frequencies. I doubt if frequency will make very much practical difference. No harm in trying.
Solarstefan
New Member
cheers understood
Solarstefan
New Member
Well, solar regulator tested on some panels and it works as expected, sun has been a bit erratic.
I am amazed how easy it is to find the MPPT by adjusting the input volts. Evidently the efficiency is 89%,
while the commercially build MPPT netted 94.5%.
I should be able to improve on that a little, with a Schottky diode or even trying to convert it to a synchronous switcher.
I am amazed how easy it is to find the MPPT by adjusting the input volts. Evidently the efficiency is 89%,
while the commercially build MPPT netted 94.5%.
I should be able to improve on that a little, with a Schottky diode or even trying to convert it to a synchronous switcher.
Warpspeed
Solar Wizard
Great progress Stefan
Try finding the solar panel peak power in a clear blue sky, then again when its all totally grey and horrible.
There will be a difference, but not very much.
Its more of a broad flat power hump than a peak, the actual panel voltage adjustment its not as critical as many people expect.
What are your nominal input and output voltages ?
See what gets warm, that should give you a pretty good indication of where the losses are.
Greatest improvement in efficiency will be by using better lower resistance choke and switching device.
You will very quickly reach the point of diminishing returns with both.
More can be gained by adding one extra panel than fretting over the last 1% in efficiency.
Try finding the solar panel peak power in a clear blue sky, then again when its all totally grey and horrible.
There will be a difference, but not very much.
Its more of a broad flat power hump than a peak, the actual panel voltage adjustment its not as critical as many people expect.
What are your nominal input and output voltages ?
See what gets warm, that should give you a pretty good indication of where the losses are.
Greatest improvement in efficiency will be by using better lower resistance choke and switching device.
You will very quickly reach the point of diminishing returns with both.
More can be gained by adding one extra panel than fretting over the last 1% in efficiency.
Last edited:
Solarstefan
New Member
Yes of course the efficiency is not that important (except for the heat that has to be removed from the components.)
Since nowadays solar panels are cheaper then chips. Just add another panel.
Its also a matter of optimisation and life expectancy, When a switcher is not happy, it will go up in smoke!
I have been stuffing around with a variety of inductors and osc frequency, as i found that i was entering in to choke saturation at higher input voltage. I wound some and also found an ultra high current 100Amp litzwire inductor.
With only 20uH its a bit low, but increasing the freq... did the trick, Next will be to increase the load. At this point it is very stable with up to 50V input and 14.4V out. But the aim is to have 100V input with 7A current. I know that will take a little more trickery.
Since nowadays solar panels are cheaper then chips. Just add another panel.
Its also a matter of optimisation and life expectancy, When a switcher is not happy, it will go up in smoke!
I have been stuffing around with a variety of inductors and osc frequency, as i found that i was entering in to choke saturation at higher input voltage. I wound some and also found an ultra high current 100Amp litzwire inductor.
With only 20uH its a bit low, but increasing the freq... did the trick, Next will be to increase the load. At this point it is very stable with up to 50V input and 14.4V out. But the aim is to have 100V input with 7A current. I know that will take a little more trickery.
Last edited:
Warpspeed
Solar Wizard
I have been using commercial dimmer chokes. These are at least 200uH and up to 1mH.
The much higher inductance reduces peak to peak ac ripple current, and its only the ac portion of the current that causes skin effect.
As you go up in voltage, core loss starts to really increase. Another reason to keep the volts per turn down by increasing both turns and inductance.
You are making real progress with this Stefan, and its a great thing to see
The much higher inductance reduces peak to peak ac ripple current, and its only the ac portion of the current that causes skin effect.
As you go up in voltage, core loss starts to really increase. Another reason to keep the volts per turn down by increasing both turns and inductance.
You are making real progress with this Stefan, and its a great thing to see
Solarstefan
New Member
I am winding another inductor as i type this, I hope it will get me close to 200uh with 2mm wire. I have a 2.3mH large inductor, that worked well too but the increase Rdc started to accumulate losses. Since most gridpanels are around 30 to 40V MPP, it is a much better proposition to work on that, rather try to use 100VDC solar....
Warpspeed
Solar Wizard
It becomes increasingly difficult to achieve higher power conversion efficiency with a buck converter, if the voltage step down ratio is greater than about 3:1
The choke has to be designed to both have enough inductance, and a low enough core loss to withstand the high dc input voltage, and use thick enough wire to carry the higher dc current at the output.
It can certainly be done, but it all requires a lot more effort to get it right.
What are you using for a core ?
The choke has to be designed to both have enough inductance, and a low enough core loss to withstand the high dc input voltage, and use thick enough wire to carry the higher dc current at the output.
It can certainly be done, but it all requires a lot more effort to get it right.
What are you using for a core ?
Solarstefan
New Member
micrometal is a T200-8 or -2 not sure atm red clear or red yellow (full wound single layer L=was only 80uh)
the other is MS-184060-2 in progress.
There is no reason for a single layers only, or using twin stack cores?
Yes I figured that ratio Vin/Vou great then 3 gets tricky
the other is MS-184060-2 in progress.
There is no reason for a single layers only, or using twin stack cores?
Yes I figured that ratio Vin/Vou great then 3 gets tricky
Warpspeed
Solar Wizard
I have some ancient Micrometals design software here for those cores, and I will plug in some numbers and see what comes out.
Double stacking helps with some things, it doubles the inductance, but core loss also doubles, and it does not help with dc saturation.
Double stacking helps with some things, it doubles the inductance, but core loss also doubles, and it does not help with dc saturation.
Warpspeed
Solar Wizard
The Micrometals software I have is for the newer molypermalloy range of cores, not the older powdered iron types unfortunately.
My powdered iron design software is on and old floppy disc, which is a problem with the computers I have here at the moment.
Anyhow working out a few things with pencil and paper, as I have the Micrometals data book here.
My powdered iron design software is on and old floppy disc, which is a problem with the computers I have here at the moment.
Anyhow working out a few things with pencil and paper, as I have the Micrometals data book here.
