3
\$\begingroup\$

I have an arc lighter similar to this one:

example arc lighter

The switch failed after a couple days (typical Chinese garbage), so I took the unit apart to replace it. Also, due to its operating frequency (~16Khz), it makes an appalling screeching sound that hurts my ears. So while I had it open, I decided I would attempt to modify it for a higher frequency.

By photographing both sides of the PCB; by using a multimeter; and by shining a light through the board, I managed to reverse-engineer it. Mostly.

My initial intention was to leave the micro in place, but insert a 555 timer between the chip and the MOSFET that drove the output coil. Its reset line would be asserted high by the chip's square wave output (by way of a diode, a resistor and a capacitor), and it would output a (thankfully inaudible) 22KHz signal to the MOSFET.¹ So far, so good.

My initial tests were promising, but in the end a pin broke off the chip and necessitated its replacement. Onto plan B then.

I figure I can just replace the (now defunct) chip with another microcontroller (specifically, a PIC or Atmel one), and write some firmware to perform the original functions, albeit at a higher frequency.

The microcontroller is unmarked, but fortuitously I have other microcontrollers with a compatible pinout. There is a problem, however. I don't understand the purpose of one of the connections on the board!

Here are the photographs I worked from. The high voltage transformer connects at the far right, and the 3.7V lithium battery at the left.² The lower image is flipped for convenience.

photos

From those pictures, I derived the following schematic:

schematic

My understanding of the circuit is as follows:

  • The LTC chip is responsible for charging the battery.
  • Voltage on the USB connector bypasses the power switch by way of Q1 and Q2, activating the microcontroller, in order that the charging lights may illuminate in sequence.
  • The microcontroller monitors the USB voltage on pin 2, and refuses to trigger if a voltage is sensed.
  • Pin 3 is for detecting when the battery is fully charged, in order to display the LEDs statically lit to indicate charging is complete.
  • Pin 5 is the output pin, by which the chip outputs a 60% duty-cycle 15Khz square wave³ to the transformer via a MOSFET.

So, my question is...

  • What is the purpose of the current sense resistor R12 (measured by pin 13)?

I have no way to know the model of microcontroller used in this product, or the internal workings of its firmware. So, what can be deduced from the circuit, in terms of how pin 13 is likely to affect the microcontroller's behaviour?

EDIT: I notice there is no back-EMF clamp diode on the transformer. Should there be?


¹ Unfortunately, I discovered that the 555 drives its output low when reset is low; driving the MOSFET on continuously, except when the square wave was present. There was some smoke, and the MOSFET physically desoldered itself from the board! Fortunately I was paying attention and disconnected it immediately; and I'm pretty sure no permanent harm was done. I added an NPN transistor and the problem went away.

² Yes, those dents happened during disassembly. Oops.

³ I measured this with an oscilloscope.

\$\endgroup\$
12
  • \$\begingroup\$ It's a current limiting resistor when FET is on. \$\endgroup\$ Commented Mar 15, 2021 at 6:23
  • \$\begingroup\$ I think not. First, it measures a tenth of an ohm, which would be a pretty ineffectual limit. And second, there is an ADC pin connected to it. It's a current sense resistor. The question is: why? \$\endgroup\$ Commented Mar 15, 2021 at 6:35
  • \$\begingroup\$ I don't deny the cpu is sensing it. But that is not the purpose of R12. \$\endgroup\$ Commented Mar 15, 2021 at 6:38
  • \$\begingroup\$ 3.7V divided by 0.1Ω gives a current limit across this resistor of 37A. Are you suggesting that the MOSFET would be likely - in the absence of this resistor - to see thirty-seven amps of current from the Lipo? Besides, even if you're correct, I'm more concerned about the purpose of the sense pin, than of the resistor itself. My firmware doesn't need to care what the resistor itself is doing. \$\endgroup\$ Commented Mar 15, 2021 at 6:44
  • \$\begingroup\$ RON of MOSFET too. \$\endgroup\$ Commented Mar 15, 2021 at 6:45

2 Answers 2

1
\$\begingroup\$

I have seen a similar circuit used in my Gas operated water heater it also consist of an electric arc generating circuit and operates on two D type cell battery.

And have tested that circuit and I got to know that this kind of circuit need to change the Ton Time of the MOSFET to generate proper electric arc between their electrodes as fix duty cycle will not work efficiently in such case, as battery voltage will be gradually reduced Ton time should be increased.

So as I think in your case R12 provides a low resistive path for most of the current to flow to the ground and with a high resistor value like 4.7k R11 is used to sense voltage feedback so that it can vary the duty cycle accordingly.

Still, I am not 100% sure about this but you can test it with an oscilloscope and providing a variable power supply to the circuit.

\$\endgroup\$
8
  • \$\begingroup\$ Unfortunately, I can't measure this, because the original chip is now broken. It was the output pin that snapped off :( \$\endgroup\$ Commented Mar 15, 2021 at 16:31
  • \$\begingroup\$ Do you have any idea how I could, by artificially generating the square wave, perform tests (without the original chip) to give me a good idea of the correct behaviour with regards to this pin? \$\endgroup\$ Commented Mar 15, 2021 at 16:33
  • \$\begingroup\$ It's tricky but you can build a simple variable square wave output pulse generating circuit with a 555 timer. or if your oscilloscope supports outputting a variable square wave pulse you can use that also. \$\endgroup\$
    – Dharmik
    Commented Mar 15, 2021 at 17:04
  • \$\begingroup\$ What I don't understand though, is how it's possible to sense correct operation by monitoring pin 13. How can the microcontroller "know" if the arc is happening? \$\endgroup\$ Commented Mar 15, 2021 at 17:34
  • \$\begingroup\$ The microcontroller doesn't know the arc is happening or not but it can sense that the power flowing from the primary side of the coil. and from that, it can manage the duty cycle. Have created a simple simulation that might help to understand tinkercad.com/things/a0UqSvhkEnm vary the input voltage and observe the change at the output voltage from 4.7K resistor. \$\endgroup\$
    – Dharmik
    Commented Mar 15, 2021 at 17:39
0
\$\begingroup\$

It might be a cheap alternative for a fuse. Its bad practice but I have seen it

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.