Author Topic: Is Youtuber "Simple Electronics" here? Re: MT3608 modules catching fire.  (Read 4416 times)

0 Members and 1 Guest are viewing this topic.

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 1954
  • Country: us
I believe I know why your MT3608 boost converter modules caught fire, and why yours may not be the only ones.



I bought this module from Banggood:

https://www.banggood.com/DC-Boost-Converter-2A-Power-Supply-Module-2V-24V-To-5V-28V-Adjustable-Regulator-Board-p-1100757.html

which I believe is the same as yours.  Shown below is a schematic that appears to be the same as our modules.  You will note that the potentiometer is shown as 100k.  The formula for the output voltage is:

0.6v x (1 + R1/R2)

So if the wiper is at the bottom of the pot, the maximum output voltage is 27.87V, which is the spec.

The pot on my module is marked very clearly P-104, which is a 100K pot.  So far so good.  But in fact, the pot actually measures 35K.  And that means, using the formula, that the maximum output voltage is - ta-dah - 10V.

So, without knowing it, you were actually trying to use the module as a step-down converter, which it isn't designed to be, and that's why it melted.

If all of your modules from a year ago, and the one I received, have the same problem, I'll bet there are many out there like that.

Fortunately, I only need around 9V from a lipo 3.7V input, so I'm good.  But I'm going to notify Banggood about the problem (for all the good that will do - yeah, right).

Others be warned - measure the pot.

 

Offline kjr18

  • Regular Contributor
  • *
  • Posts: 197
  • Country: pl
Re: Is Youtuber "Simple Electronics" here? Re: MT3608 modules catching fire.
« Reply #1 on: February 14, 2018, 03:02:33 pm »
I have a few of them, so I replaced that pot and second resistor with two resistors to select output voltage around 9V. After connecting it to power supply current draw hit a limit. If it was a battery, some magic smoke would escape. So it's not necessarily a pot is a problem here, as I had no pot to begin with. After reflowing all of the joints with lead based solder module works fine. I think that problem might be with those solder joints, might be a good idea to touch them a bit before powering up your circuit.
 

Offline soubitos

  • Frequent Contributor
  • **
  • Posts: 352
  • Country: gr
    • I sell on Tindie
Re: Is Youtuber "Simple Electronics" here? Re: MT3608 modules catching fire.
« Reply #2 on: February 14, 2018, 05:15:35 pm »
The formula for the output voltage is: 0.6v x (1 + R1/R2)

In my early attempts to use MT3608 i used a schematic quite like the one above, a pot and a single resistor.
With that setup, i dont know why but i did fry several chips.
Also, the capacitors i use are 25V versions so i have to limit output voltage to that limit too...
Now, i use a 47k resistor in the upper branch of the voltage divider a 2.4k on the lower branch and a 50k pot in the middle. interestingly, the pot is actually affecting voltage only at about 1/3 of its turns... so with the pot in its original position full left, i get Vout~Vin After several turns it starts boosting.
Now, i read somewhere that in the modules with only a pot on the upper voltage divider branch, you need to bridge the loose end of the pot to the center pin and then it works all the way....
Another well known variant is modules with 68k resistor in series with a 30k pot and a 2.2k resistor in the lower branch.

The most "weak" parts in these modules though are the caps...
Also, adding a 10uF tantalum cap in the output will help with "noise"

What is the code on the chips in your modules?
 

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 1954
  • Country: us
Re: Is Youtuber "Simple Electronics" here? Re: MT3608 modules catching fire.
« Reply #3 on: February 14, 2018, 06:36:56 pm »
It turns out I was dead wrong about the pot.  I removed the pot, and it was fine - 99.4K.  I thought the module board followed the schematic, but it didn't.  On the module, as soubitos suggests, the wiper is also connected to the bottom of the pot and the top of the 2.2K fixed resistor.  While I thought the ends of the pot would always measure 100K per the schematic, the reading actually depended on where the wiper was.  So I'm back to being puzzled about exactly why those modules in the video lost smoke.

It turns out that to get a nice near-linear output voltage range, the module did it right, and the schematic does it wrong.  The output voltage is:

Vout = .6(1 + R1/R2)

So if the wiper is in the middle of the pot, at the 50K point, the module has it like this:

Vout = .6(1 + 50/2.2) = 14.23V

which is roughly half of the maximum it will do.  But the schematic would give you:

Vout = .6(1 + 50/52.2) = 1.17V

which is not even a legal value - because it's a boost converter.

Also, I think soubitos' additional resistor above the pot is a good idea.  Properly chosen for your input voltage, that would prevent you from selecting an output voltage that's too low, which I think could release smoke.

I ended up with fixed resistors instead of the pot.  I've been a bit surprized that my 9V output voltage from an input lipo cell isn't noisier than it is.  So far it doesn't seem any worse than the 9V wall wart I was using.

 

Offline soubitos

  • Frequent Contributor
  • **
  • Posts: 352
  • Country: gr
    • I sell on Tindie
Re: Is Youtuber "Simple Electronics" here? Re: MT3608 modules catching fire.
« Reply #4 on: February 14, 2018, 10:00:01 pm »
Actually, no matter what resistors/pot you select MT3608 will NEVER give you an output lower than VIN-0.3V aprox.
If in disable mode, it gives exactly that as the current from the input passes through the inductor and schottky diode (hence the voltage drop)... The problem, and the point i fried several MT3608 chips is when you force it to an output voltage higher than 28V.. Again, checking the output capacitors or even replacing them with a tantalum one, or at least a very good quality X7R perhaps with correct voltage rating is essential both for the proper function of the IC but also for better ripple in the output
 

Offline soubitos

  • Frequent Contributor
  • **
  • Posts: 352
  • Country: gr
    • I sell on Tindie
I use MT3608 as the output stage in my TP4056FlexAdv modules.
It does have a lot of limitations but its cheap and makes a simple solution for non-demanding tasks.
Here is how it behaves in various voltage outputs.... it claims 2A but in real life you can consider it an average 500mA capable chip.
Here are my results

 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6899
  • Country: ca
I think fake/bad MT3608 IC's that are the problem.
I bought some MT3608 boost modules off Aliexpress store Module Sky and they are all duds, for $0.50 each.
No boost and no smoke, a big nothing burger.
 

Offline soubitos

  • Frequent Contributor
  • **
  • Posts: 352
  • Country: gr
    • I sell on Tindie
This is indeed a possibility..
I buy mine from LCSC.com and i can say i trust them to deliver genuine parts.
I also buy IC's not modules (not since i make and sell mine that is)...
What i have come up with is that the chips themselves (if genuine) are more than ok...
What limits them (including my builds so far i admit) is proper pcb layout and escorting parts.. coils and capacitors like in all switching converters are sometimes critical and limiting the chip itself.
I expect parts and pcbs to check several changes to my current designs, i think it is promising to say the least but i dont really know if it is worth the extra cost involved to design properly for a cheaper converter to get its max potential or simply build cheaply around a little more expensive chip which will deliver the same result....
I think like most things it is a constant swirl of compromises, acceptation of the capabilities and being cost effective.
I am sure the chinese modules worth 0.50$ with swapped parts, better coils/caps will perform much better... yet they do their job for what they are...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf