Capacitor blew up, don't know why

Hi Guys,

I am up to the next part in my induction annealer control circuit project.

I have run into trouble.

I wired up the circuit (Drawing attached. Sorry about the drawing. Free online circuit drawings don’t allow some components, so I have drawn them as text boxes, hope that’s ok).

When I plugged it in, Capacitor C1 blew up within 1 second.

The whole electrical process is meant to go as follows:

1 - 36V DC from the meanwell supply is feed to the relay contacts, to run the induction circuit when called by the relay
2 - The LM2575T Adj drops the voltage from 36V to 12V, so I can run the 12V relay coil
3 - The LM317T Adj drops the 12V power to 5V so I can run the arduino and other control components (Momentary switch, Time setting pot, LED, and MosFet which switches the 12v relay)
4 - The LED lights up and shows when the is power being supplied to the MosFet

The application for this circuit is basically a timer, controlled by arduino (and code).
When the momentary switch is pushed, 5V power is feed to the MosFet gate.
The circuit closes the 12v relay for a period of time.
This period of time is set by the timer set pot (up to 10 seconds, as an input into the arduino).
This allows the induction circuit to be on for that period of time.
Once the timer runs out, the 5V feed is dropped from the MosFet and the relay turns off, turning off the induction circuit.

I am very new at this.
This circuit is kind of a bit of a frankencircuit, as I grabbed all the necessary circuits I needed and put them all together hoping it would work.
It probably has some heaps of parasitic capacitance, or stray inductance or the like.

I just want to know why C1 popped (as it is a requirement for the LM2575 circuit), and how to fix it.

Thanks in advance guys. :confused:

You had it backwards?

The induction heater circuit is pushing back huge voltage spikes?

Yeh, polarity might have been it. I didn't even check that!!

um...the induction circuit will draw roughly 9 amps when running. Does that mean the capacitor C1 will be exposed to that 9 amps? Or with it by-pass the C1, as the LM2575 and everything after it will draw a lot less (maybe 0.5-1amp)?

steltzer: Yeh, polarity might have been it. I didn't even check that!!

um...the induction circuit will draw roughly 9 amps when running. Does that mean the capacitor C1 will be exposed to that 9 amps? Or with it by-pass the C1, as the LM2575 and everything after it will draw a lot less (maybe 0.5-1amp)?

Most likely "reverse polarity". Check the polarity on all Polarized Caps.

Did your relay even turn on in the 1st Second? If not then it cannot be the Induction Heater.

If the Induction heater does generate large kickback voltage then consider adding a reverse bias diode across Induction Heater. and/or an RC snubber.

There might also be a large ripple current - some details of the load would be useful to figure out more.

Electrolytics put in backwards will explode (well, burst, they are deliberately weakened to reduce the danger from shrapnel).

Polarity is my guess too, especially since:-

Yeh, polarity might have been it. I didn’t even check that!!

Always check the polarity of electrolytic caps when fitting them This includes tantalum caps. They make a nice BANG too. (DAMHIKT :smiley: )

More is about to blow up. A slow 1N4004 is bound for silicon heaven in a switching supply.

C2 is on the wrong side of the inductor.

Why R1. It's is very dangerous there. When (not if) the pot looses contact... Should be on the other side of the pot.

Why the second 5volt regulator. The Uno already has one. Just feed the 12volt from the first regulator into the DC socket.

Pin13 should not be used. Leo..

@mrsummitville - I have checked it, placed the replacement cap in the right way.....blew up in 2 seconds. And no the arduino and the induction circuits were not even connected, I was just trying to test and make sure the circuit was going to give me 12V.

@MarkT - yeh, the load was not connected at all. neither was the arduino.

@Wawa - LOL silicone heaven, Love it.

I decided to wire up just the LM2575 part of the circuit, on a breadboard to see if i could get a stable 12V out of it. I replaced the diode for a 1N5819 Schottky Diode.

I did realise that the capacitor was on the wrong side of the inductor in the diagram. I actually had it wired the correct way, I just drew it wrong for you guys. So that was all correct in practice.

Ok, with R1, what's the problem there? Why would the Pot lose contact? and why so definite with a when, not if? And if it were to lose contact, I suppose the would send a huge voltage spike on the output side??? is that the danger?

The second LM317T regulator, is because the Uno is only temporary. I am actually going to use a mini for ease and physical size once it is embedded into the project, and this required a 5V input as far as I am aware. It has not USB plug or anything, so I need to make 5V with a reg.....I might be overthinking it?

Pin 13 should not be used? how come? I wanted to utilise the LED on the Arduino for confirmation that the arduino was sending 5V to pin 13 while running. Not a good idea??

And another question, is there a difference between an RF choke inductor and a wound toroidal inductor? If you had 330uH RF choke inductor and a 330uH wound core toroidal inductor, could they be interchangeable in this circuit?

I really appreciate all your help guys. This is where people can learn! Thanks HEAPS!

steltzer: Ok, with R1, what's the problem there? Why would the Pot lose contact? and why so definite with a when, not if? And if it were to lose contact, I suppose the would send a huge voltage spike on the output side??? is that the danger?

Pin 13 should not be used? how come?

When you turn a pot, it could have contact bounce while adjusting. The adj pin is grounded through R1 during those short bounces, resulting in ~35volt spikes at the output of the regulator.

Pin13 could have Arduino's onboard LED attached. That could be set to flash during bootup, causing the relay to chatter. Leo..

Hi, Can you post a picture of your project so we can get an idea of how you have wired everything.

A couple of pictures from different angles would be great.

Thanks Tom.... :)

steltzer: @mrsummitville - I have checked it, placed the replacement cap in the right way.....blew up in 2 seconds. And no the arduino and the induction circuits were not even connected, I was just trying to test and make sure the circuit was going to give me 12V.

@MarkT - yeh, the load was not connected at all. neither was the arduino.

@Wawa - LOL silicone heaven, Love it.

I decided to wire up just the LM2575 part of the circuit, on a breadboard to see if i could get a stable 12V out of it. I replaced the diode for a 1N5819 Schottky Diode.

I did realise that the capacitor was on the wrong side of the inductor in the diagram. I actually had it wired the correct way, I just drew it wrong for you guys. So that was all correct in practice.

Ok, with R1, what's the problem there? Why would the Pot lose contact? and why so definite with a when, not if? And if it were to lose contact, I suppose the would send a huge voltage spike on the output side??? is that the danger?

The second LM317T regulator, is because the Uno is only temporary. I am actually going to use a mini for ease and physical size once it is embedded into the project, and this required a 5V input as far as I am aware. It has not USB plug or anything, so I need to make 5V with a reg.....I might be overthinking it?

Pin 13 should not be used? how come? I wanted to utilise the LED on the Arduino for confirmation that the arduino was sending 5V to pin 13 while running. Not a good idea??

And another question, is there a difference between an RF choke inductor and a wound toroidal inductor? If you had 330uH RF choke inductor and a 330uH wound core toroidal inductor, could they be interchangeable in this circuit?

I really appreciate all your help guys. This is where people can learn! Thanks HEAPS!

I think, you have the C2 CAP connected to the LM2575 improperly? The Data Sheet show C2 AFTER the Coil. The C2 cap is in the wrong place? Output CAP C2 should be, at least, 150% x 12 Volts = 18 Volts (min)

Data Sheet shows D1 = 1N5819, but you substituted 1N4004 ? Is that OK ?

Let's REMOVE the 12V Relay, the LM2575 and Cap C1 ... Connect the Battery and verify proper polarity at C1 with voltmeter. Does C1 actually have 63 Volts printed on it? Many times the LINE/MARK on the CAP is "Negative" terminal.

1) Shut down, insert C1, now does C1 blow?

2) Shut down, insert LM2575, now does C1 blow?

3) Shut down, insert 12 V Relay, now does C1 blow?

LM2575_DataSheet

Input Capacitor C1 The Input Capacitor should have a Low ESR. For stable operation of the switch mode converter a low ESR (Equivalent Series Resistance) aluminium or solid tantalum bypass capacitor is needed between the input pin and the ground pin to prevent large voltage transients from appearing at the input. It must be located near the regulator and use short leads.

mrsummitville:
The Input Capacitor should have a Low ESR.

I don’t know about the current rating of your 36V DC power supply, but 36V/Low ESR = lot of current at t0. As LM2575T do not seems to supply a lot of current (Relay <0.1A + Arduino <0.05A +LED 0.02A + Resistors and Trim pots <0.05A ?) would it be a good idea to add a limiting resistor between the 36V DC supply and C1 ? I’m not sure that the C1 capacitor is supposed to supply current to the induction heater so what do you think about adding a diode to avoid excessive discharge current ? It’s also a good pratice to put a protecting resistor beside a trim pot to avoid any unwanted overcurrent when the pot is set to rail or to ground. If the capacitor C1 blow again maybe you can consider using 2x 47uF (63V) instead of one to split the current a bit…

I assume that your Arduino pin 9 is an Input Pin otherwise don’t put it to ground please I care about all Arduinos.

Well Guys, it is working. I went back to the drawing board. Got the correct parts as per the spec sheet. Wired up....no smoke!

Only a couple of things. 1 - @Wawa , when you say R1 needs to be on other side of the pot, I am not sure how you mean so the the LM2575 gets its division between the 2 resistances, and yet if the pot bounces, there is no 36v spike. Can you clear that up for me? I might need a super simple drawing, Sorry mate.

2 - My math for selecting the resistances (pot1 and R1) were way off. I got that R1 needed to be 1k and R2 needed to be 9K, but to acheive 12V on the output (pin 2 of LM2575), I had to wind pot1 to 6.6k. Not sure how to work that.

But I got 12V!!

Thanks for all your help guys! You have all been so helpful. If 1 thing I have learned.....I need to learn A LOT more!!

www.ti.com/lit/ds/symlink/lm1575.pdf Figure 26.

R1= 1k1 R2= 10k Should give ~12.4volt.

If R2 is open circuit, the chip gets no feedback and thinks it needs to raise the output voltage. If R1 is open circuit, the chip gets full feedback from the output, and thinks less is needed.

It makes sense to make R1 adjustable, e.g 1k resistor and 200ohm pot in series. Then a crappy/intermittend pot can't spike the output voltage. Leo..

hhmm....@wawa.. . I made R1 1K and R2 10K, and I got 1.233V. That seems a factor of 10 off your prediction of 12.4V

Have I done something wrong? I cannot get 12.4V with that setup.

steltzer: hhmm....@wawa.. . I made R1 1K and R2 10K, and I got 1.233V. That seems a factor of 10 off your prediction of 12.4V

Have I done something wrong? I cannot get 12.4V with that setup.

He said 1K1 and 10K (Not sure if that's the problem - I can't remember exactly which part of the circuit is being referred to. Just an observation.)

Crap I read it wrong…my bad LOL

I don’t have any 1.1K’s or 6.6ish K’s, so I will stick with the POT1 and R1.

I just ordered the right bits so I don’t have to use the POT1.

Attached redone drawing…Should be a bit better.
Cap2 is in the right position now.

Can anyone see any further issues with this circuit.

Should be ok…Next issue is to work out why my Arduino mini’s are not connecting to my iMAc…That will be on another thread.

Thanks Guys!

You can use any number of resistors to compose a 1K1 resistor. A 1K + 100 Ohms would do that, and two 560 Ohms would get you pretty close too.