TL431: Setup for 12V to 24V in, 2.5V out

Hi Christopher,

This is an interesting schematic. The main purpose of stacking '431 is to control the thermals between the TL431 like devices. I am unsure on how much current they need require TL431 stacking. When you are stacking 5x 'TL431 like devices you can just a wide Vin switcher for a similar cost.

Do you know how much current the supercap will need?

Hi Marcoo,

Thanks for your quick response! Ok, I believe I get the idea then.

The function for this circuit is to provide an UPS function for something like a Raspberry Pi or similar of course - that normally have a flash memory like an SD card that gets sad when power is just cut off. So, to allow a safe shutdown the UPS circuit is used to provide up to 30-60 seconds of additional power when main power is cut. For example the RPi 3B could need as much as 2.5A at 5V when using all its horses - in a shutdown sequence I believe the current needed is much smaller, but I haven't actually measured this yet.

Tests have been made with using a similar circuit on the 5V side, but it doesn't seem to be able to hold as long as using this circuit I'm inquiring about which is used on the 12V side and then fed into a buck converter to get the 5V out.

Anyhow, I do not know how much current the supercap will need. The original schematic use 3F for a much easier 5V load. The second test, which was for RPi, used 10F capacitors.

Can it help with the datasheets of the supercaps?
10F https://www.we-online.com/catalog/datasheet/850617021004.pdf
50F https://www.we-online.com/catalog/datasheet/850617022002.pdf

The idea is to test with 10F up to 50F, to see how long "UPS" function we can achieve. They all have a different rated discharge current and of course, price needs to be as low as possible - so with a good margin we can pick the best caps for our needs.

Thank you once again!

Hi Christopher,

I would say that the key requirement is how long do you need to charge the supercap?

Is there a requirement that the supercap has to be charged in 1 minute or does it not matter?

Is it ok to charge it at as lower current?

I would highly recommend to use something like the TL431 in a discrete Linear regulator. This can be seen in figure 34 of the datasheet.

This will be a much better and lower Iq solution instead of stacking 5 TL431s. With this you can set the voltage output for a low cost solution.

Hello Marcoo,

Sorry for late reply! Of course, a fast charge is always good to be prepared. However in my setup I have a 230V-12V UPS that buffers the whole system. But when that is out of juice the Raspberry Pi still dies abruptly as it has no idea when that UPS fails - there is no output available from it without some serious modifications to someone elses hardware, so that doesn't work unfortunately.

So, this UPS based on supercaps and the TL431 could then take over the last 30-60 seconds allowing the RPi to gracefully shutdown.

Power failure in Sweden is certainly not very common, it happens of course but very rarely. So I'm going for that the supercaps will have time to recharge over a longer period of time +30 min or so. Of course, the faster they are ready the better, but I don't need them ready directly after power comes back.

A lower current charging is certainly fine.

So your proposed solution is to exchange 5 of the TL431s with just one like Figure 34. Vi(BATT) will be the 5 super caps which could still be in series I guess? When there is no external power the super caps will output its voltage through the transistor to Vo.

However, how do they get charged back up? Is it through the same Vo but then instead through the TL431, is that correct?

This would then require a transistor capable of at least 3A and the circuit would need to be charged with 5V.

The original schematic takes 12V and I wanted it to take up to 24V, preferably automatically or manually with a voltage selection jumper. The output of the capacitors, in series, would give me 12,5V - 13.5V.

So this gives me both the external 12-24V and the supercaps 12,5-13,5V which goes into a buck regulator that steps the voltage down to 5V 3A, buck regulator is not shown in original schematic - only the output of roughly 12V where the schematic is cut off on the right side.

Figure 34 gives me a 5V side circuit, if viable for 5V 3A, while the original schematic is for ~12V. So, I guess its two different solutions, correct?

Can you help clear up some confusion.

What is the input and output of your desired circuit?

Are the supercaps on the output or the input side of the 12-24?

Are you trying to input 12V-24V and get 12V output or is it 12.5V/13.5V to 24V?