ST MC34063A DC-DC converter

[boylesg]

I am trying to get one of these pricks of things to give me 5V from 15V it it just wont work.

I have been through the calculations on the datasheets, done the same calculations with online calculators and still this mole of thing wont give me 5V.

The best I can get is about 3.84V.

I have randomly tried different sense resistor values such as 1R, 0.22R and a couple of others.

I have tried larger and smaller output capacitors.

I have tried varying R2 on the voltage divider.

And it all makes no difference - I still get 3.84V.

I am on the verge of chucking these things in the bin because they just aint worth the hassle!

I just don't understand what the hell you have to do with these things to get them to work as reliably and hassle free and the linear voltage regulators.

 

[tpetar]

What inductor you use uH and mA ?
What Schottky diode you use ?
On pin 3 what capacitor value?
What current you need on output ?



Try this parts :

Inductor 1,2A-1,5-2A 34uH
Rsc 0,3R
Co 25uF
Ct 150pF
R1 10K then try 1K
R2 30K then try 3K
Schottky max 0,4V



Or try this configuration :

:wink:

 

[betwixt]

I use lots of them and they all work properly for me. I suppose you are using the correct schematic, it's different if you are boosting the voltage or dropping it.

Take a look at this page: **broken link removed** The diagram will change according to the input and output voltages.

Note that you MUST use a fast diode in the circuit, a normal 1N400x type of diode will not work.

Brian.

 

[boylesg]

These are in an excel spreadsheet with all the formulas.

I have had a heap of trouble getting these formulas correct because the datasheet does not tell you what units each of the calculated values should be in. So I have had to engage in a lot of guess work. But after watching a video tutorial that runs through how to do all the calculations I think I have them right now. But either way I have done these same calculations with the online java calculators on various websites and the answers seem to come out about the same.


VIN (V) 15.00
VOUT (V) 6.00
IOUT (mA) 200.00
VF DIODE (V) 0.75
VSAT (V) 0.50
FMIN (kHz) 100.00
VRIPPLE (V) 0.0010
tON / tOFF 0.96
(tON + tOFF) (ųs) 10.000000
tOFF (ųs) 5.090909
tON (ųs) 4.090909
CT/C3 (pF) 220.91
IPK (mA) 400.00
RSC (Ω) 0.75
CO/C1 (ųF) 500.00
LMIN(ųH) 104.32
RANGE E12
R1 (Ω) 1200
R2 (Ω) 4560.00
R2 STANDARD (Ω) 3900.00
VOUT STANDARD (V) 6.00

- - - Updated - - -

The chip I have is an ST: CHN 063AC / KFQ024

Is this even a standard MC34063A DC-DC converter or do I have something different?

Is this the correct datasheet for this chip?

http://www.datasheetcatalog.org/datasheets/2300/501448_DS.pdf

 

[betwixt]

It is the correct data sheet and probably the correct IC but I'm suspicious of the values from your spreadsheet.

Make sure you are usng the schematic on page 7 of the data sheet and not the one on page 6.

100KHz is rather too fast for the oscillator. As you increase the speed you also increase the losses in the internal switching transistor and the external diode. To some degree the frequency is a compromise between size and efficiency but personally I would target a frequency in the range 20KHz to 30KHz. Note that you MUST use a fast recovery diode or preferably a Schotky type in the circuit, a normal diode will not work.

I would suggest you use the values suggested in the schematic for 5V and use R1 = 2.4K and R2 = 9.1K if you want 6V out. These are easily available values.

Brian.

 

[boylesg]

It is the correct data sheet and probably the correct IC but I'm suspicious of the values from your spreadsheet.

Make sure you are usng the schematic on page 7 of the data sheet and not the one on page 6.

100KHz is rather too fast for the oscillator. As you increase the speed you also increase the losses in the internal switching transistor and the external diode. To some degree the frequency is a compromise between size and efficiency but personally I would target a frequency in the range 20KHz to 30KHz. Note that you MUST use a fast recovery diode or preferably a Schotky type in the circuit, a normal diode will not work.

I would suggest you use the values suggested in the schematic for 5V and use R1 = 2.4K and R2 = 9.1K if you want 6V out. These are easily available values.

Brian.

My *** this so f'ing frustrating!

This buck converter circuit is straight out of the datasheet and implemented in Multisim 12.

It still doesn't f'ing well work.

What the #$%& am I doing wrong?????



- - - Updated - - -

I corrected the 470F to 470pF but it still makes little or no difference to the output.

Clearly I am misinterpreting the schematics some how but I am damned if I can see it.

 

[tpetar]

You try this only in simulator ?

Try this on protoboard, dont waste time specially with 34063 in simulators.

Try circuit from post #2, I want to say make real circuit dont play in simulator, its ok if you have lots of free time.

 

[betwixt]

I agree with tpetar. I rarely use simulations, they work fine in some situations but rely entirely on the mathematical modelling of the IC to work in this kind of application. In turn, the modelling relies on the person who wrote it and where they got the information from themselves. Internal schematics from data sheets are mostly representative rather than the real design and of course the numbers from data sheets vary from one manufacturer to another. A few years back I worked in a small team designing VLSI test strategies and often two of us would model the same device and end up with entrely different code.

Just build it, with only 8 components it quicker to construct it than enter the simulation data!

Brian.

 

[Doctor_E]

From my experience with the MC34063A,

When you talk Power never use simulator or even BreadBoard,
Always work with PCB,

For the Schematic:
1-Change the capacitor from 470pF to 1.5nF
2-Use a different Diode from this general purpose Shottky
( Ultra High Response time around 25 to 35nS )
3-Change the Output Capacitor from 470uF to 220uF25Vdc

Regards

 

[tpetar]

Output voltage is controlled with two resistors (values on your schematic 1,2K and 4,7K), and formula is :

Vout = 1,25 x ( 1 + 4,7K/1,2K ) = 6,145V



When you accumulate strength and try this without simulator in real life, if there is no results :

Change capacitors, maybe you have old, bad and crystalized capacitors with high internal resistance, change them with fresh caps. :smile:

Do not make long connections between parts.

 

[boylesg]

Thanks guys.

As far as using a simulator goes you have to remember that I am not industry experts like you guys.

Rather than figuring out how to get things to work extremely well like you guys, where it matters how accurately various components are simulated, I am merely trying to get circuits that I am not familiar with to work at all.

And using the simulator is more convenient that soldering and re-soldering matrix boards and even setting things up on a bread board and trying to measure currents all over the place with 2 multimeters.

Once I figure out how it works with the simulator THEN I move to the bread board and THEN to a soldered circuit.

Believe me, for a raw beginner at this stuff, the above process is fairly efficient.


I believe I have at least figured out the simulator part of my problem. I was using MC34063ADG that was available in Multisim 12. Either this is not a stock standard MC34063 type chip and is is wired up differently to the manner specified in the datasheet I have. Or else this component, in Multisim 12, has been created incorrectly and either way is not behaving as I was expecting.

I managed to find a SPICE model for an MC34063A on the ON Semiconductor website and successfully created a user component in Multisim 12. When I use this component in place of the above one it works exactly as the datasheet says it should.

Based on this I can only conclude that the ST CHN063AC chips I was sold are not standard MC34063A chips or else they are faulty. I have not found any datasheet for CHN063AC so I think these chips are basically bin worthy. There useless if I can't find the correct datasheet.

- - - Updated - - -

All the web pages on this device seem to say that 1N5819 schottky is fine and that is what I am using on my bread board.

 

[betwixt]

Expertise comes from experience. Basically, we have all gone through the same process of frustration as you experienced and those of us still surviving are here to help others :smile: As for the others - we will never know!

Beware of cheap ICs, especially if they don't come from authorized distributors. There are many fake or rebranded parts on the market, I've even heard of large can electrolytic capacitors being opened up and having a smaller, lower rated one inside, complete with plastic sleeve and inferior rating printed on it.

Brian.

 

[tpetar]

I've even heard of large can electrolytic capacitors being opened up and having a smaller, lower rated one inside, complete with plastic sleeve and inferior rating printed on it.

You mean like this :








With name CHN063AC I cant find datasheet, but I find lots of chinese sites for selling this part. Probably you have something what is not MC34063. :-? I hope you didnt buy few kilos of this. :|

 

[betwixt]

:-D EXACTLY like that - but that isn't the picture I've seen before so it may be widespread.

The other problem is genuine parts that failed when tested during manufacturing. They could be real MC34063 devices but ones that were rejected because they failed the quality control tests. I've heard of waste skips being hi-jacked as they left the factory and their contents sold on the black market. Many will work to a point but may not deliver full current or their output voltages may be incorrect. To be safe, only buy from authorized distributors, they can trace the parts right back to the production line so you can be certain they are the real deal.

Brian.

 

[tpetar]

The other problem is genuine parts that failed when tested during manufacturing. They could be real MC34063 devices but ones that were rejected because they failed the quality control tests. I've heard of waste skips being hi-jacked as they left the factory and their contents sold on the black market. Many will work to a point but may not deliver full current or their output voltages may be incorrect.

I always ask myself what they do with this products which didnt pass quality testing.

 

[boylesg]

Output voltage is controlled with two resistors (values on your schematic 1,2K and 4,7K), and formula is :

Vout = 1,25 x ( 1 + 4,7K/1,2K ) = 6,145V



When you accumulate strength and try this without simulator in real life, if there is no results :

Change capacitors, maybe you have old, bad and crystalized capacitors with high internal resistance, change them with fresh caps. :smile:

Do not make long connections between parts.

Is the voltage output entirely controlled by the voltage divider or do the other parts play some role in getting it close to the level required with the voltage divider fine tuning it?

If the voltage divider entirely controls the voltage output then how is that these devices are efficient? Because I would think that voltage divider would be an efficient means of changing the voltage of a source.

- - - Updated - - -

Oh......if you were to require top efficiency in converting the output of a solar panel to 5V or so required by a hall effect sensor and an arduino microcontroller then would these MC34063A chips be adequate or could you do a lot better?

For the Science Works solar car challenge

 

[betwixt]

The voltage is entirely controlled by the divider as long as the switching circuit is still working within it's capabilities. The other values will dermine it's efficiency and for each input/output condition there will be different optimal LC values. It is very tolerant of LC values though, the feedback mechanism inside the IC will compensate if the values are not best and the efficiency may only suffer a few % if the values are not optimal.

For solar power conversion there is another problem altogether which is a science in itself. The power from PV is the voltage it produces multiplied by the current it is supplying (W = V * I) but the voltage drops as the current increases and also as the light level (insolation) decreases. The point where V*I is highest depends on the light level and load you put on it. They are not linear, there is a peak in power at a particular load level and it changes with each light level. To get most power you have to use a system called MPPT which is quite complicated. MPPT is "Maximum Power Point Tracking" and it uses switch mode regulators controlled by a microprocessor to optimize the output. Usually it 'swings' the load on the PV slightly up and down to hunt for the maximum power point. It's a whole leap ahead of MC34063A technology so I would advise mastering them first!

Brian.

 

[boylesg]

The voltage is entirely controlled by the divider as long as the switching circuit is still working within it's capabilities. The other values will dermine it's efficiency and for each input/output condition there will be different optimal LC values. It is very tolerant of LC values though, the feedback mechanism inside the IC will compensate if the values are not best and the efficiency may only suffer a few % if the values are not optimal.

For solar power conversion there is another problem altogether which is a science in itself. The power from PV is the voltage it produces multiplied by the current it is supplying (W = V * I) but the voltage drops as the current increases and also as the light level (insolation) decreases. The point where V*I is highest depends on the light level and load you put on it. They are not linear, there is a peak in power at a particular load level and it changes with each light level. To get most power you have to use a system called MPPT which is quite complicated. MPPT is "Maximum Power Point Tracking" and it uses switch mode regulators controlled by a microprocessor to optimize the output. Usually it 'swings' the load on the PV slightly up and down to hunt for the maximum power point. It's a whole leap ahead of MC34063A technology so I would advise mastering them first!

Brian.

Well these solar car are just small things that run around on a track so I don't think high school students are going to be able to cope with this level of electronics that is no doubt applied to road based solar vehicles.

As far as I can see the solar panel simply has to supply enough power for the microcontroller and small DC engine.

I don't know that much about solar panels at present but I am presuming that the solar panel that the teacher has obtained will supply enough power for both over a reasonable range of light conditions.

I guess this will be part of the development process - working out what sort of power the solar panel can provide, what sort of current the microcontroller will draw and therefore how best to optimise the lot to power both devices. It will probably take more than just this year to get a good solution and we can build on it as we gain more knowledge.

This will be a new learning process for myself as well as the teacher and the students involved.

I am going to suggest that they all join this forum and take advantage of your collective and formidable expertise.

So look out for new members and more questions along these lines.

- - - Updated - - -

The voltage is entirely controlled by the divider as long as the switching circuit is still working within it's capabilities. The other values will dermine it's efficiency and for each input/output condition there will be different optimal LC values. It is very tolerant of LC values though, the feedback mechanism inside the IC will compensate if the values are not best and the efficiency may only suffer a few % if the values are not optimal.

For solar power conversion there is another problem altogether which is a science in itself. The power from PV is the voltage it produces multiplied by the current it is supplying (W = V * I) but the voltage drops as the current increases and also as the light level (insolation) decreases. The point where V*I is highest depends on the light level and load you put on it. They are not linear, there is a peak in power at a particular load level and it changes with each light level. To get most power you have to use a system called MPPT which is quite complicated. MPPT is "Maximum Power Point Tracking" and it uses switch mode regulators controlled by a microprocessor to optimize the output. Usually it 'swings' the load on the PV slightly up and down to hunt for the maximum power point. It's a whole leap ahead of MC34063A technology so I would advise mastering them first!

Brian.

Well these solar car are just small things that run around on a track so I don't think high school students are going to be able to cope with this level of electronics that is no doubt applied to road based solar vehicles.

As far as I can see the solar panel simply has to supply enough power for the microcontroller and small DC engine.

I don't know that much about solar panels at present but I am presuming that the solar panel that the teacher has obtained will supply enough power for both over a reasonable range of light conditions.

I guess this will be part of the development process - working out what sort of power the solar panel can provide, what sort of current the microcontroller will draw and therefore how best to optimise the lot to power both devices. It will probably take more than just this year to get a good solution and we can build on it as we gain more knowledge.

This will be a new learning process for myself as well as the teacher and the students involved.

I am going to suggest that they all join this forum and take advantage of your collective and formidable expertise.

So look out for new members and more questions along these lines.

 

[boylesg]

When I asked the question is there any more efficient means of transforming voltages than the standard MC34063A, I really meant in terms of the fixed output chips that are based upon the generic MC34063A but that I have been reading can provide even greater efficiency. Are these fixed output type DC to DC converter chips likely to be worth the additional expense for this particular application or not really?

Also......

With the linear voltage regulators the excess voltage is dissipated as heat, as would be the case with a simple voltage divider, and I understand this fairly well.

But how is the MC34063A chip doing exactly if it is the voltage divider that controls the output voltage?

Regulating the amount of current passing through the voltage divider and therefore not wasting as much power in heat dissipation by it?

 

[betwixt]

No, the way they work is quite different. I'll try to explain as simply as I can:

Look at the properties of a (hypothetical) perfect switch - when turned on it has zero resistance between it's contacts so the power loss as current passes through it is (I * I * 0) = zero, when turned off it has infinite resistance so no current flows (0 *0 * infinity) = 0. So a perfect switch would not dissipate any heat at all. Of course in real life the switch would have a small resistance and it would get hot if passing enough current.

In a switch mode regulator (there's a clue in the name!) instead of controlling the output by inserting a resistance in series with the output, it uses the relative time that FULL power or no power is allowed to reach the output. Because it switches almost completely off or completely on, the series pass transistor works like the perfect switch and produces no heat. If you just used the output as it came from the pass transistor you would have a very coarse spikey voltage going from zero to the input voltage which is obviously not what you want. This is where the diode, inductor and output capacitor come into play. You will note that the diode is connected so it normally DOESN'T conduct but there is a reason for it being there. You will also note that the inductor is in series with the output so all the load current has to flow through it.

It works in cycles like this:
1. The oscillator in the 34063 turns the series pass transistor fully on and it shorts the input to the output pin. The diode is reverse biased so it doesn't conduct and the output capacitor starts to charge through the inductor.
2. The oscillator turns the transistor off. The current through the inductor stops and the magnetic field it created collapses. This produces a voltage with reverse polarity to when the transistor was conducting. The diode now conducts and clamps the input end of the inductor close to ground potential. The output end of the inductor is now lifted up by it's stored energy and again charges the output capacitor.

So the transistor is always hard on or hard off, both conditions of low heat loss and the inductor is doing two things, slowing the rising edge of the output voltage when the transitor conducts and filling in the gaps when it's switched off. you can see why the inductor has an optimal value and why it changes with the switching frequency. Basically, the inductor is working as an energy reservoir, taking over from the transistor when the oscillator turns it off. The speed at which the current reverses in the inductor is also the reason why a fast diode is needed. A normal silicon power diode takes quite a long time to stop and start conducting as the polarity across it changes but in this application it has to stop and start conducting at the oscillators speed. Typically the switching is at around 20 - 30 KHz but a normal 1N4001 type of diode can only manage maybe 2 KHz at best.

To actually control the output voltage, the relative time the transistor is turned on compared to it's off time is varied, this limits the energy that can be delivered to the inductor and hence the output voltage. Inside the 34063 the voltage fed back from the divider resistors (which is proportional to the output voltage) is compared to an internal reference voltage and the circuit uses the difference to set the duty cycle of the pass transistor.

I hope that makes sense. I oversimplified it a bit !

Brian.