The switch side of the inductor is a flat line at 12 V, and the base of the external transistor is stuck at 0 V. There is no switching going on. I also probed the timing capacitor, and attached its waveform below. I have also attached PDFs of my schematic and board layout, in case the problem lies there.
Why is my regulator not starting, and what can I do to fix it?
I took some more measurements. I have another (working) MC34063 in a different circuit and took a screenshot of its timing capacitor waveform. It look exactly the same. The screenshot is attached.
I think my transistor might have become damaged during soldering. I took some new measurements and found about 8.3 V (!) at the base of the transistor relative to the emitter. It's constant too, I can't see any switching. When I use the diode test mode on my meter and measure across the base/emitter junction, I get about 0.7. When I swap the leads (red on emitter, black on base), I actually get a smaller reading, of about 0.56. When I measure across a 1k ohm resistor as a test, I get about 0.7. When I measure across the collector and base I get 0.287. When I swap the leads I get the same reading. When I measure between the collector (black lead) and emitter (red lead) I get 0.597. Swap the leads and I read 0.77.
Does this sound like a broken transistor, or are these diode readings worthless in-circuit? The 8.3V on the base should mean something.
Sounds like a stuffed transistor, but the issue will be what broke it.
No switching sounds like the main switching transistor Q1 in the MC34063 is also stuffed.
Whats the saturation current of the series inductor?
The minimum load when I plug it in is about 60-70 mA. It's driving an ATmega328P, some op amps and digipots, an LCD screen, and a TPA3122 Audio Amplifier.
I designed the maximum load to be about 1 amp, though by my measurements it's not reaching an amp except possibly in brief moments during the punchier parts of the sound. When I really crank the music up, it puts out about 4 W average.
I did some Google searching and found that the basd-emitter resistor should be a "low value". Trouble is I don't know what low value means. Is the 1 k that I'm using too large? The internal basd-emitter resistor appears the be 100 ohms should I make my external one smaller? Would that affect the startup?
See if you can measure the resistance between pin 1 and pin 2 of the MC34063.
You will have to disconnect one of the pins from the rest of the circuit to do this.
This is to determine whether the switcher transistor , Q1 in the chip is fried.
You should get very high resistance.
Ok, all I can suggest is you try another switcher chip.
Normally if the triangle wave generator is running, but no switching is happening
then either Q1 or Q2 in the switcher chip is fried.
Its usually Q1 that goes.
Id try just using another MC34063 just by itself without an external pass transistor
and Id also suggest you increase the Inductor value to at least 150 uH with the same saturation
current as the existing inductor.
33 uH is very marginal at startup.
I resoldered the IC back onto the board and even reflowed it with my hot air gun to make sure the solder joints were good. My oscilloscope is now showing a weird wavefom on the base. Picture is attached.
I'm not sure how the base voltage can be so high. Pin 8 isn't showing any switching, the collector of the switch isn't showing any switching, there's just the timing capacitor waveform, which I've already posted, and this weird base waveform.
I have another version of this circuit, a through-hole version that I made first, that works. That version doesn't have an external switch though. I used an external switch for the SMD version because I thought the currents would be a little high for the smaller chip to handle. And now it isn't working.
Unfortunately I don't have another SOIC 34063 on hand. I think I'll try changing my 1k base-emitter resistor to 180 before I order a new chip or two.
Replacing the base-emitter resistor with a 180 ohm instead of a 1 k definately changed the base waveform. pic_17_2 shows the new waveform. Channel 2 on that pic is the timing capacitor. Current consumption at idle also increased to about 100 mA.
Also, I AC coupled the signal on Pin 8, and it's showing some very minor switching. Second pic is attached. It's not much, but some current is flowing through the drive collector. The collector of the external switch also showed a very small switching drop at about 80 kHz, but it's down in the noise; I didn't bother taking a screenshot.
I use this design tool to calculate the optimum values for MC34063 switcher circuits. MC34063A design tool
It gives you the optimum values, for any config , Buck,boost or inverting.
The important value is the Inductor peak current , as it determines what the inductor size
needs to be to avoid saturating.
I found that tool, and it doesn't seem to be giving me sensible values. I get an Lmin of 0 uH and a Ct of 0 pF!
I have sized my inductor properly for peak current. I made a mistake early saying I designed this convertor for an output power of 1 A, it's actually designed for 500 mA ( ~13 W). The peak inductor current for that is about 2.5 A, which is within the 3.1 A that my inductor is rated for.
Boost converters draw more current when they start, than when they are running, as when the converter starts
its looking into charging the output capacitor which looks like a short circuit initially.
If you dont get anywhere, I suggest you download LTspice4 , Design Tools & Calculators | Design Center | Analog Devices
and simulate your circuit.
You can them measure the peak inductor current at start time and run time .
Unfortunately I can't figure out how to add a 34063 model to LTspice. I've tried finding explanations online but none of them made much sense. Is there a model I can add to PSpice somewhere? I know how to use that.
I ordered 5 new SOIC 34063s, along with some 0805 0 ohm jumper resistors. If I can't get a new switching chip to work, I'll short the base & SWE pin to GND so it'll operate like a regular 34063 circuit and bypass the external switch. We'll see what happens then.
Thanks Crossroads, there had to be a way.
To install the files in Ltspice, copy the mc34063.lib file to the lib directory.
Copy the mc34063.asy file to the sym directory, which is in the lib directory.
The mc34063_buck.asc is a good example which can be played with.
Just mod the circuit to suit your boost circuit.
I removed the transistor and replaced R30 with a 0 ohm short resistor so that the switching regulator is in the regular, non-external switch configuration Wouldn't you know, it starts right up and works.
So it must have been something wrong with my transistor configuration, even though I checked the transistor's pinout a dozen times and compared it to the board.