I'm using a MT3608 module to rising up the voltage to 20V, this way caps store a lot of energy allowing to still run the load with not input (which occurs frequently) for 30 long seconds.
All works "fine" but what I notice is that as soon as I connect the supply, The booster charge a little bit the caps, lets say 1V, then the input voltage drops and booster shutdown, then the cycle repeats again while caps get 2V, voltage drops again and so on...
The power supply is a very weird generator that delivers 2 positive pulses of about 13V at 30hz, the average power I estimate this coil offers is around 500mW *216mW
I measured 3A Spikes on the output while caps are in charging. The caps takes 45s to reach 20V and at that moment all works flawless, no more spikes or ripple. Putting any capacitor on the input only makes things worse, it seems that there is not enough power to feed the capacitor and the booster at the same time and it just never starts.
Any idea how I could solve this? Ideally with some IC or new booster rather than creating a circuit myself. I don't have any great knowledge of electronics.
Note: There may be other better approaches such as putting in a battery but I have deliberately chosen this way because of certain restrictions
*Edit: Not sure if the best approach. In order to obtain a more accurate result of available power I attached directly after diode a 1.8v regulator in conjunction with various capacitor capacitances until I reached maximum power while keeping the regulator alive. I got 119mA so ~215mW
Regulator has 89% efficiency (Vin 8v Vout 1.8v @ 120mA)
Maybe because the capacitors when empty are like a short to GND. The current will be too high and the boost converter shuts down to protect itself. The capacitors have some small resistance, and also the boost, but probably not enough. And the caps are in parallel, reducing more their resistance.
Try adding a resistor before the caps to keep the maximum current below 2A, at least to test if this is the reason. 20V / 1A = 20Ω
As commented below, better a bigger resistor with the needed power ratting.
You are right. Maybe better a bigger resistor, 1KΩ or more. I think that the caps will charge fast anyway and will reduce the power dissipation quickly as they get charged.
Yes, exactly, but I think booster shuts down just because there is not enough energy (which is
true) in the input and Vin drops below 2V rather that protecting.
As soon I use any kind of resistor at the output all stop working, seems like the small energy available is wasted on hot rather that charging the caps.
Try without the boost, charge the caps to 13V, with a resistor also.
If it works, probably is easier to add more caps to get the same buffer, than dealing with the boost. In this case at least.
The first aproach was without a booster, which also works fine. The problem is Caps only can be charged to 8V (13V - caps - 40mA load)
To keep 40mA alive for 30 seconds with 8V charged caps (until 3.5v) I need 200mF, 10x more that now, which is a huge bunch of them and a lot of extra space.
That sounds like another part of the problem that ideally you would tell us a little more about if you want to get some useful suggestions.
Another problem there.
My initial response would be "use a better power supply and be done with it", but I guess that's not an option. Or maybe you believe it's not an option although it really is; hard to tell. If you explain a bit more about those restrictions/requirements and explain the application a bit more, it may be possible come up with some more good solutions.
Probably that's the problem. You mention that the source provides about 500mW. And the 13V, peak to peak? will be less, it means about 60mA or less. But boosting it to 20V the current will be half, so when the caps are empty and the differential is about 20V there is no enough power for the load and charge the caps at the same time.
Probably the on/off ripple also disconnects the load for a moment and little by little the caps voltage increases, and needs less current to finalize. And it will happen again when the caps go down, I suppose.
So, you need to charge the caps with a very small current, slowly, to let the rest for the system work. And even then I don't know if it will be enough.
Maybe one resistor and a diode as suggested in post #10. So, charge very slowly, at 10mA... and get current back through the diode.
I don't know, something like that, if you can't address the power supply issue.
assuming 20V output and a limit of 10mA, that should be a res of about 2K right? Does this mean that 200mW (20V * 0.01A) will be thrown into heat? 200mW may sound ridiculous but it's about half of what I have available.
If the schematic in post #1 is correct, then you are powering the boost with a 30 Hz half wave rectified sine wave. How do you expect it to operate correctly from that kind of waveform?
The MT3608 IC has an under voltage lock out that turns the converter off if the input voltage drops below around 2V.
You need to have a smoothing capacitor after the diode to get a DC voltage at the input to the converter, so that the input voltage does not fall below 2V. Thus allowing the converter to run continuously rather than in 30Hz bursts.
Yes, but only when the caps are at zero volts. As they charge the voltage drops progressively and also the current, until they are full. Then there is not voltage drop across the resistor, so no current.
But ok, you can limit more the current. It doesn't matter if it takes a couple of seconds more to charge from zero.
And also as commented, try to rectify the input signal, to use both sides of the wave. And smooth it with a capacitor to avoid that the boost is starting and stooping all the time.
As you have an AC input source, a funny thing would be to try a voltage doubler instead of the boost, with just 2 diodes + 2 capacitors:
https://www.falstad.com/circuit/e-voltdouble.html
In the firts circuit maybe the last cap could be your caps, taking care that you don't get near 25V.
I don't know if this would fit in your project. Time ago I tried something similar, and it surprising works.
Your schematic says the caps are "10mF", that is ten thousand micro Farads, is that correct? How much current can your 13VAC supply source? If it is a transformer, what is it's VA rating?
Wether you want or not, at 30 Hz you need a bridge rectifier and a big cap before the boost. At your small current 1000 microF should do... you cannot use the resistors before the cap as it needs to fully charge in less than 1/60 sec...
