Overheating mc34063 as boost converter

I've been using the mc34063 in boost converter manner to build an external battery circuit for 3.7v->5v DC-DC conversion, but within few minutes of operation the chip and the inductor begin to overheat beyond touchable (light smell of burning components too), occurring with both circuits that I had. I've tested using the following devices: ipod touch 2G, ipod touch 5G, kross g6 tablet, all of which had the circuit overheat. The power source is 4 cylindrical cells in parallel @ 3.8v ~9A with protection circuit. The batteries do not overheat and I'm guessing the protection circuit regulates the current to an extent. I've tested the current draw of the ipod 5G to 100mA while the circuit is wired for 500mA. This is what the schematic looks like: http://www.eeworld.com.cn/uploadfile/dygl//201210/20121025102243965.png
I would like some help in troubleshooting the overheating of the circuit, Thank you.

Picture of Overheating mc34063 as boost converter
Photo 2014-07-18, 9 33 43 PM.jpg
sort by: active | newest | oldest
Nighter3D2 years ago

be sure to use a well rated inductor that can handle the current and to double-check that you are using a schottky diode. beyond that be sure everything is within the Regulator's rating.

If looking to go beyond the chips ratings may i suggest to look at MC34063 based Nixie supply circuits? They give an
idea at how one can use a external Mosfet to go beyond the MC34063's
ratings. It may not be the solution what you need, But its very
interesting and worth remembering if working with MC34063s often.

Oh and in the future i would probably Test with a Proper Load resistor rather than a actual live device. its a bit risky to test such circuits on a expensive device.

Ive put the numbers in my Design tools (very handy, be sure to look those up). According to that to create 500ma would require Ipk (aka switch current) of...1680mA. That is Over the general limitation of the MC34063 which can handle 1500ma at most. To overcome this you will need to use a External Switch (Look up nixie-supplies for examples on how to do this with say a mosfet and get a high current rated mosfet) and a Inductor rated for atleast 2A...preferably higher.

Ahyku (author)  Nighter3D2 years ago

I've added an external switch to the circuit using the D4126L transistor which should handle 3A collector emitter current. The diode is indeed a 1n5819 Schottky diode which handles 1A. The circuit would operate normally for the first few minutes and charging the device about 5%, but after, over a 10 min period, the inductor and the transistor begin to overheat without any charging at this point. Would this likely be the problem of the inductor?

Nighter3D Ahyku2 years ago

Hmmm. well there is also a different possibility. May i ask what value for Rsc you are using? The tiny resistor at the input. a common beginners error to leave it out or forget about it.

Also a suggestion for future troubleshooting. it would help to give a proper schematic, rather then the standard included reference. preferably with the values and ratings of the used components. so far you haven't exactly told us what rating the inductor has for example. Give all the resistor values and such.

Ahyku (author)  Nighter3D2 years ago

if you mean the resistor between pin 6(vcc) and 7 then that's a 0.22 ohm resistor. The battery is directly connected to vcc of the chip. The inductor I used in the previous picture I personally didn't know the value of. I changed it with a different one that I wound using a toroid calculator where my intended amount if Henry's was 3.54uH(what was shown in the calculator for the chip), the only thing I'm not sure of is if the gauge of the wire is good for this purpose. Another testing with this inductor showed that the only thing to overheat was the transistor, which made a squealing noise in operation. The transistor is suppose to handle up to 4MHz so I'm now wondering if that might be the fault of the value of the capacitor at pin 3

Nighter3D Ahyku2 years ago

3.54uH sounds a little Meager actually. It will have a tendency to "saturate" as stevastrouk mentioned and cause overheating within the transistor. There is nothing wrong with adding some turns...heck its even recommended as the lower the power-draw is the bigger the inductor is advised to use (else it will saturate when not enough is consumed). So you may want to wind it up towards the 100-200u range (most use 120uH as "thumb-rule"). i see nothing wrong with your gauge (if it can handle the current its perfectly fine, my 2.5A coils have a similiar wire).

As for the transistor and oscillator. you will want to aim for the 10-30khz as oscillation (high enough for easy decoupling and good efficiency). So the ceramic capacitor should be around the 600-1000pf(as-long as its above 600).

Ahyku (author)  Nighter3D2 years ago

I've taken your advice and made it ~120uH and increased the capacitance to 1000pf. Testing this with my 2nd gen ipod, it seems to charge fine. The 5th gen however becomes choppy, where it says it's charging then disconnects, and repeats. This may have to do with the chip in the newer apple cables and will probably be as far as I will get with this circuit. Thanks for all the help anyway, much appreciated.

Nighter3D Ahyku2 years ago

No problem, glad to hear its atleast no longer overheating, which is nice.

As for the choppy loader, well the MC34063 isn't that perfect for charging(tends to get unstable with higher current draw and often needs extra output filtering in the form a capacitor-inductor-capacitor filter). For future revisions may i suggest looking at the LT1302? Its very neat little switching regulator... kinda should have suggested it before as its more suitable for what you want, but i forgot about it. you may know it as the "Mintyboost" regulator.

Whats probably happening is the inductor is saturating, and then, when its inducance falls to zero, the chip overheats.

1.) Find an inductor which can handle the design current.

2.) AND use an external switch.

-max-2 years ago

Also, this does not appear to be able to deliver 1.5A out. Rather, that is the peak current of the inductor. When the current is abruptly cut, a voltage transient (spike) is created and that goes through a diode into a storage capacitor to be used. The diode prevents backflow. The RMS (average) current output in the boost mode is 100mA. That will not be able to charge up anything like a smartphone, which can draw as much as 2A.

It's funny, this appears to be exactly the same chip I used, I found it in an old NiCd battery charger. You may first want to try to wire in the step down configuration, and use an external transistor for higher current capabilities.

-max-2 years ago

Try using a higher value inductor, and a high speed schottky diode. Make sure the capacitors are high ESR 105*C type, as well as from a reputable company (Rubycon, vishay, etc.) It seems you have good short connections, so just make sure the circuit is made right.

Also note that lithium ion / polymer batteries fully charged will be at 4.2V. The protection circuit drops about 100mV, so the output will be closer to 4.1V. The difference between this and the 5V output may not be great enough for the booster to regulate properly. See the datasheet on this chip for details.

If you plan to redo this, I would wire the 3-4 batteries in series and use a 12V-5V buck converter. These are extremely common and are often found inside car cigarette lighter phone chargers.

The advantages of the buck converter is that is it slightly more efficient, simpler (in my opinion,) and can deliver more current to the device.

At most the chip can handle a power draw of 1.5A. But even with a 500mA draw it will generate a lot of heat. But the chip can handle about 100 C without failing. May be a good idea keep it well ventalated or put a heat sync on it.