Hi everyone, I'm trying to design a simple PCB for a energy harvester.

The energy harvester potentially can develop up to 40v. The problem I have is that I want to shut off input voltage from the harvester if it goes below 3.5v. I have looked at many circuit designs similar to what I need, like solar panel & battery chargers etc. But I cannot find anything that goes up to 40V.

The reason for this is that we have lots of these energy harvesters and we want to later connect them in series, so at the moment I have them connected to XL6009 buck boost device and have it set to 40v. So that when connected in series all the devices are at the same voltage and we don't get a drop in voltage from a weaker energy harvester. The problem is that when the voltage from the harvester is lower than 3.5v the XL6009 doesn't function but just passes the lower voltage through.

I want to stop this happening so when they are in series all the devices either have a voltage of 40v or nothing at all.

Hope this makes sense.

Thanks for any input on this problem. John

 

Connecting them in series makes no sense. You cannot achieve a constant output this way. You can connect them in parallel for this purpose. Problem is that a weak one will draw power from the stronger ones. You can connect them in parallel with Schottky diodes, but depending on the output current you will lose several tenths of a volt in each diode. At least the weak ones won't drag down th the stronger ones.

I'm going to go out on a limb here and say that you are probably wasting your time in terms of what you hope to accomplish. Energy harvesting is no more than a curiosity. It has few if any practical applications. Solar power generation is another matter entirely.

 

Connecting them in series makes no sense. You cannot achieve a constant output this way. You can connect them in parallel for this purpose. Problem is that a weak one will draw power from the stronger ones. You can connect them in parallel with Schottky diodes, but depending on the output current you will lose several tenths of a volt in each diode. At least the weak ones won't drag down th the stronger ones.

I'm going to go out on a limb here and say that you are probably wasting your time in terms of what you hope to accomplish. Energy harvesting is no more than a curiosity. It has few if any practical applications. Solar power generation is another matter entirely.

Hi Papabravo, thanks for you quick reply.

I'm involved in a project at a College in the UK, which has asked for some kind of power supply that can handle many different power inputs and many different kinds of energy harvesting inputs. (TEGs, Solar Energy, Wind, Water Wheel, kinetic etc...)

We have 4 power management boards each with ten inputs with each that connect to a XL6009 buck boost convertor set to 40v output. The output of these are then connected in series, so if all the XL6009's are working we should have output of about 400v. The 4 power management boards are then connected in parallel and in turn these are connected to a Inverter.

So the problem I want to solve with a nice cheap and simple circuit to put in front of each the XL6009 convertors to either stop or restrict any voltages lower than 3.5v but can also handle any voltage up to about 40v. As you said it will bring down the voltage of the whole circuit to the lowest voltage and the XL6009 convertors only start working at 3.5v.

Hope this makes sense. Thanks again John

 

I second Papabrovo's comment.
Although it is hard without real details. It may not be so, but when I hear "energy harvester" I think of trying to get a few mW from vibration, temp differences, sound, and other very low power sources.
Maybe more details as mentioned above would help.
Volts are pretty irrelevant, but power gives a better idea of what you are after.
It could be you are trying to get your 400V at 1mW, and that is a very different kettle of fish compared to 400V an 10KW.

The XL6009 have an enable pin so if you pull it down for low volts, the inverter switches off. But if you are thinking of running a number in series, I predict smoke!

 

I have the similar problem for fan feeding at heavy-duty RF field. Every kind of 78xx is `going mad` inspite of any groundworks and Faraday cages, and voltage may get up to 50 V in input while output must be stable 12. So, my experience says the best solution is most slowest bjt what are able to find, for example russian P2XX or kT8XX what are so hopelessly slow that RF is not disturbing em at all. So,m Zener diode, bjt and voila!.
In Your case, probably one another parameter is important, that is voltage V(ce-F). Then try to find a germanium bjt what was produced in early 70-ies, their V(F) is ca 0,2...0,5 V instead of silicium 0,5...1,5 V However their V(R) may be too tiny, then 6 Volts was counted be `high voltage`, but exist there options even for 50 V.(ulta-high super duper).
Other may be better option for Your case is to use a igbt of Mosfet with lowest possible V(ce) or V(ds) and sdteer it by another (normal) bjt or jFET.
As an example https://qph.fs.quoracdn.net/main-qimg-8bed3d17a67189b1fd26b7b62fc17e63
https://www.electrical4u.com/images/march16/1463572856.PNG
or any other

 

However, I may wrong understood Your question, as the word regulator have two far different meanings. I wrote about stabilizing meaning, but probably You was about BOOST-UP smps?
Then read of https://en.wikipedia.org/wiki/Boost_converter and come with a questions again.

 

I second Papabrovo's comment.
Although it is hard without real details. It may not be so, but when I hear "energy harvester" I think of trying to get a few mW from vibration, temp differences, sound, and other very low power sources.
Maybe more details as mentioned above would help.
Volts are pretty irrelevant, but power gives a better idea of what you are after.
It could be you are trying to get your 400V at 1mW, and that is a very different kettle of fish compared to 400V an 10KW.

The XL6009 have an enable pin so if you pull it down for low volts, the inverter switches off. But if you are thinking of running a number in series, I predict smoke!

Hi Dendad, I understand your concerns. We have nearly 40 types of energy input most of them are low voltage and low amp's that have been developed by other students, the idea of the project is to prove that you can gather energy from anywhere and almost any source. Some of the energy inputs are more advanced than others, so involving all years of the College students. I the highest voltage energy input is no higher than 40v (with most being much less) and the highest amp energy input is no higher that 3 amps.

You mentioned the enable pin on so the inverter switches off? We are using 10 of the following boards mounted to a main board.
https://www.ebay.com/p/Boost-Buck-A...m=182860486317&_trksid=p2047675.c100011.m1850

Thank again John

 

I'm not at my computer at the moment. It is bed time here. Hail iPads
Still, if you look at the data sheet of the switcher chip, one pin is the enable. The chip will run if you connect it to the input voltage.
And if it is 0V, the chip stops.
So it will need to be lifted off the PCB pad carefully, and
have a pullup resistor added from input to the enable pin.
Then is you use an open collector output comparitor, LM393 I think from memory, wire up to compare a ref volts to the input, it can switch the regulator on and off.

And I think you would be better served to have all these converters driving one load in parallel, diode isolated on the outpus, all charging a big cap that feeds the main 40V to 400V boost supply.
Running these all in series will be dangerous as the inputs will alll have to be floating as each next input source will be 40V higher than the previous one, and you run a real risk of blowing things up with gross overvoltage, and that is not mentioning the other real risk of electroplating a student who may be trying to measure his potato battery that now is a couple of hundred volts above ground.

 

I'm not at my computer at the moment. It is bed time here. Hail iPads
Still, if you look at the data sheet of the switcher chip, one pin is the enable. The chip will run if you connect it to the input voltage.
And if it is 0V, the chip stops.
So it will need to be lifted off the PCB pad carefully, and
have a pullup resistor added from input to the enable pin.
Then is you use an open collector output comparitor, LM393 I think from memory, wire up to compare a ref volts to the input, it can switch the regulator on and off.

Thanks Dendad, I will look in to that. Sorry for keeping you up.

 

I added to the above post too, after you commented.
It should maybe have been a now post but, well....

Hopefully it is of some help. Tomorrow a circuit could be added if someone else has not beat me to it.
Please keep us posted of your progress.

 

Oh, and one final thought before I try to sleep, having each unit at ground reference potential will make it a lot easier to meter the individual performance for comparison.

 

Being at a college in the UK does not fill me with confidence that you are going to end up with anything other than a burnt out pile of boards. Let us know how it works out.

 

Oh, and one final thought before I try to sleep, having each unit at ground reference potential will make it a lot easier to meter the individual performance for comparison.

Please don't reply, I'm just updating you.

The idea you had about putting the buck boost in parallel then having a large cap to boost it to 400v is great, it will eliminate many components needed for each buck boost and in turn will be safer and a simpler design as well as being more efficient by using all available voltages.

I was going to have a 2 pin connector on each input to plug in either a Mini 2-wire Volt Meter Display or to be connected to a sensor board that in turn is connected to a Rasp PI for monitoring.

 

There is an issue with diode-ORing multiple voltage regulators to a common output - the highest voltage wins.

Presumably the objective is for each source to contribute according to its means. If the output of each converter is a current, rather than a voltage, it is easy - currents sum. Otherwise you have a situation where the device with the highest output voltage contributes "first" and that with the lowest output voltage contributes "last" and then only as each device is overloaded, allowing the voltage into the "final" converter to drop, which only happens if the output is asking for more power than what is available. This may be acceptable or it may not.

TEGs are a problem because they require a maximum power point (MPP) converter for highest efficiency (which is horrible, at best) but also require a more-or-less constant load for a given heat input to prevent destruction of the cells. Photovoltaic sources require an MPP converter for highest efficiency, but they don't care about being operated without a load or with gross overload. It is pretty much universally true that there will be a maximum power point on the voltage versus current curve for any source with limited input power. The "knee" of the curve may be very sharp or very rounded.

Any switchmode converter that has less power available at the input than the load would require for the converter to regulate will simply "collapse" the input supply by driving the duty cycle to the maximum. For a buck converter - if the duty cycle went to 100% the output is directly connected to the input through the switch, inductor and diode. For a boost converter, 100% duty cycle (which can't be allowed) will simply short-circuit the input supply through the inductor and switch. This will apply to every converter in the system, including the final one that produces 400 volts from the 40 volt rail.

For the converters that feed the 40 volt rail, this is actually a useful behavior, provided the input voltage is prevented from falling too far. For example, an MPP converter will set a minimum voltage at the input if there isn't sufficient power to regulate the output voltage. Instead of delivering a well-regulated voltage, the converter delivers the maximum current it can for the amount of power available at the input - which is perfect for the requirement of taking from each according to its ability.

Inverting or "flyback" converters have some merit because there is never direct current flow from the input to the output. Energy is stored in an inductor during part of the cycle and delivered to the load during another part of the cycle. Because of this, the absolute value of the output voltage can be either less than or greater than the input voltage. The inductor will discharge its stored energy at 2 volts or 200 volts, which makes the converter well-suited for putting energy into a shared output pool (capacitor). It still has the issue of managing the input power.

 

Hi Epb thanks for your input, didn't think this would be so complicated. I have been looking at circuits with a large cap as mentioned by Papabravo, but then you have got me interested in the fly-back convertor method. Do you have any suggestions for a circuit or even an off the shelf product that will fit my needs.

Thanks John

 

Being at a college in the UK does not fill me with confidence that you are going to end up with anything other than a burnt out pile of boards. Let us know how it works out.

Hi Papabravo, I have been looking at circuits to boost the 40v output to 400v but not sure what would be the best way of doing this. Ebp posted and mentioned a Fly-Back convertor, i'm really not sure which way to go. I started this project thinking it would be a straight forward solution, but I seems to be much more involved than I expected.

Thanks John

 

Hi Papabravo, I have been looking at circuits to boost the 40v output to 400v but not sure what would be the best way of doing this. Ebp posted and mentioned a Fly-Back convertor, i'm really not sure which way to go. I started this project thinking it would be a straight forward solution, but I seems to be much more involved than I expected.

Thanks John

The flyback converter is a workable approach to your problem. I think you need a more basic starting point and it is this:

The power output of ANY device will ALWAYS be LESS than, sometimes much less than, the power input.​

In practical terms you need to do some "back of the envelope" calculations. Let us suppose that you need 400 Volts DC @ 500 mA. Multiply these two numbers together and the result is 200 Watts of output power. I like to use an efficiency estimate of 80% for any device which has not yet been designed to estimate what kind of input power will be required. Dividing 200 Watts by 80% gives 250 Watts of input power that will be consumed by the device. Now suppose the input voltage is 4 Volts DC (the low end of you input voltage range), that means that 62.5 Amperes of current will be required since 4 Volts times 62.5 Amperes is 250 Watts.

I don't know if you are familiar with the kind of challenges you will face designing a circuit to handle 62.5 Amperes, but they are substantial. Mistakes and oversights in design and layout will result in toxic smoke and possibly shrapnel. Components that can handle these power levels are extraordinarily hard to find, fabricate and replicate. This is no country for amateurs. Please tread carefully, and always wear eye protection.