MC34063 in Voltage Boost Configuration Voltage Drop

[sonar_abhi]

Hello,

I am using MC34063 in a voltage boost configuration to boost 3.5V to 5V.
Vin: 3.2V
Vout: 5V
Iout: 375mA
Component Values:
R: 180 ohms
Rsc: 0.22 ohms
Cin: 10uF
Cout: 10uF
L: 22uH
R3/R4(for boosting to 5V): 3.3K/10K

The problem is when I am testing the output with a DMM, it shows proper 5.1 V output, but when connected to even a small load, say like a microcontroller, the Vout drops to 1.5 or 2 V.

I have tested the cuircuit on a breadboard as well as on a protopcb, the same problem persists.

Can anybody please help?

Thanks

[ali6x944]

try increasing the inductance but before doing that could u please give us all the calculations just to make sure we understand what we are is your situation.

[PeterL]

First of all: Please clean up your schematic. The connections might all be there from an electrical point of view, but for a human to read it, it's a mess.
Second: Just selecting an inductor on it's inductance is not enough. You need to make sure it can handle the peak current in your configuration. It also must be usable at the frequency you use.
Third: Your Ct looks very low to me. 22pf fits for an operating frequency of about 1MHz, the MC34063 is designed to operate at max 100kHz.

Finally: There are some online calculators for the mc34063, like this one:
http://gmsystems.com/switching-reg-calculator-for-mc-34063-or-mc33063.html

And Dave has made a video about this, #110.

Good luck

[alsetalokin4017]

FWIW, here's my attempt at a re-draw of sonar_abhi's original schematic. I have just redrawn the original in the normal "between-rails" style, I haven't checked the original for correctness or functionality. I've "?" the timing capacitor value since the original 22 pF is probably too small.

[Phoenix]

Can anybody please help?

A few points/questions
- Why are you using a zener as the diode? The diode needs to have fast reverse recovery, in low power supplies like this that often means using a schottky (or fast recovery diode) i.e. the datasheet shows a 1N5819 schottky and a UF4004 is a jelly been fast diode. I would consider this a major problem. The fast recovery diode will most likely be faster than the schottky but have a higher forward voltage.
- Boost converters aren't good to run without any load as it leads to uncontrollably high voltages.

Things to do
- Change the diode for a schottky (e.g. 1N5819) or fast revcovery diode (e.g. UF4004).
- Do you have an oscilloscope?
   - Check the oscillator frequency is what you expect
   - Provide a waveform of the collector of Q1. This switching node is very important in SMPSs and can tell you a lot about its performance.

[mariush]

I would also like to point out that MC34063 is kind of "risky" for this, as most manufacturers specify the minimum voltage as 3v - 3.2v is fairly close to 3v.  And yes, you would have to use fast recovery diodes, and ideally with small as possible voltage drop.

My suggestion would be to ditch the old MC34063 and go with more modern boost (step-up) regulators, here's just some examples:

MICROCHIP  MIC2288YD5-TR Boost (Step Up) Regulator, Adjustable, 2.5V-10 Vin, 3.6V-34 Vout, 1A out 1.2Mhz: http://uk.farnell.com/microchip/mic2288yd5-tr/voltage-reg-boost-1a-1-2mhz-tsot23/dp/2509929



R1 and R2 can be tweaked to more sensible (common) values , you just choose them to have Vout = Vref  x ( r1/r2+1) where Vref = 1.24v and R2 must be below 5 kOhm .. the datasheet is very thorough
 
Since it runs at such high frequency, you can use very small (and cheap) inductors and ceramic capacitors and resistors and even though it's surface mount, it can still be soldered onto a prototyping board (there's enough spacing between pins to solder them on board) but such high frequency regulators like wide and short traces, every component as close as possible to the chip, so proto boards are not really recommended.

Other similar chips:

MICROCHIP  MIC2619YD6-TR  DC-DC Switching Boost (Step Up) Regulator, Adjustable, 2.8V-6.5Vin, 2.8V-35Vout, 350mAout, SOT-23-6 : http://uk.farnell.com/microchip/mic2619yd6-tr/ic-boost-1-2mhz-ovp-6tsot/dp/2509963  (very similar to the one above)

MICROCHIP  MCP1640CT-I/CHY  DC-DC Switching Boost Step Up Regulator, Adjustable, 650mV-5.5Vin, 2V-5.5Vout, 350mAout, SOT-23-6 : http://uk.farnell.com/microchip/mcp1640ct-i-chy/boost-reg-synchr-500khz-6sot23/dp/1800208



This chip will also give you the capability of running your projects from two AA batteries, since it can run with as little as 0.65v (two AA batteries would be as little as 2x1.2v = 2.4v) . And, it also has a very long and detailed and easy to understand datasheet, which explains how to choose the components (diodes and inductor, what specification matters when choosing them etc) and for this chip, the datasheet even has a picture showing how they recommend the traces to be layed out around the chip, to keep efficiency and performance high.

[sonar_abhi]

Hello Guys,

Thanks for your responses. I should look into the MIC2288YD5 TR chip suggested. It seems to be a better alternative than what I have designed(based on MC34063)

About the diode, my bad, I forgot to clarify...actually I have used 1N5817 in my project. But I could not get the simulation model for 1N5817 in Proteus hence used what was available in Proteus.

From my understanding, I have come to the conclusion that the inductor that I am using is not rated for that current. I will get a 1A/2A rated inductor and work out again.

This is the inductor I am working on

[Phoenix]

This is the inductor I am working on

That looks a lot like a 22ohm resistor...

[SeanB]

No, it is an inductor, probably with a 100ma max current rating, and it likely is saturating at something around 150mA and doing a pretty good imitation of a resistor.

[sonar_abhi]

Hello Guys,

I made several changes in the circuit and tested in for the following changes:
1] Inductor: 22uH/2A
    Shorted the pins 1 and 8 (to use it in a darlington configuration)
    Timing capacitor: 390pF
    Output capacitor: 100uF/16V
    Removed the external transistor and checked it with transistor also

2] Inductor: 42uF/2A
    Shorted the pins 1 and 8 (to use it in a darlington configuration)
    Timing capacitor: 470pF
    Output capacitor: 1uF/16V
    Removed the external transistor and checked it with transistor also

Still the same problem persists. The output voltage at no load is 5V proper but under load, it drops to 3.2V.

The reason I persisted with MC34063 was its low cost (it costs Rs. 10/piece in India compared to Rs. 60/ piece for Micrel MIC2288) and that there are people who have posted that they have used the MC34063 chip with the same Vin and Vout at 300/350mA. But I guess for me, there is no other option than to use the Micrel chip.

Thank you all for your help.

[mariush]

390pF for Ct is a good value.

You need larger input capacitor, recommended at least 100uF. Wouldn't hurt to have higher output capacitor on the output, to reduce ripple (ex for 0.1v ripple it's recommended about 270uF). These capacitors should be low esr - if you don't have new, recycle some from an old motherboard or a bad power supply.
Rsc is pretty important , at around 0.24 ohm - I'd suggest paralleling 4 x 1 ohm resistors if you don't have exact value.

Note that the peak current on the inductor can't go over 1.5A, which will happen with more than about 0.35A on the output - it could be that you're actually going over 0.35A on the output and that's why your circuit fails.

I'd suggest playing around with the MC34063 calculator from sourceforge, it's a decent tool :



But keep in mind those cheap mc34063 chips you're buying may be no-name or Chinese clones which can't operate properly at such high switching frequencies (I've used 55kHz above, some clones may not work well above 40kHz or so) and again, keep in mind that the inductor you use matters. The MIC2288 chip may cost 6 times more but you normally save money in the inductor and the cost of the pcb surface used (surface mount tiny inductor 1210 size or thereabouts compared to through hole large inductor) and you also don't need electrolytic capacitors which use a lot of space and you can use instead cheap ceramic capacitors which you can buy for dollars on a roll, and the MIC chip has more protections built in and it's simply a more efficient design (about 85% and more efficiency, with mc34063 unlikely to pass 65% efficiency without carefully picked parts and provided your device always consumes some current close to the value you configure it for).

In real world, the cheaper cost can bite you in the long term and it's not really that much cheaper. It's still used because like you said, it's cheap and chinese manual labor is cheap (people can insert through hole resistors and inductors and diodes on pcb all day for a couple of dollars), time on a surface mount chip placement machine can still cost more than manual labour.