MC34063 boost configuration. Help needed.

[m12lrpv]

I've just spent the weekend frying some MC34063's and I think it's time I asked for help 

What I'm trying to do is run a led strip from 4 D-Cell batteries. The LED strip draws about 35mA from a 9v battery and can operate up to 12v.

I was initially aiming for about 11v out and up to 200mA and using the downloadable calculator. http://sourceforge.net/projects/mc34063uc/

It worked briefly then burned out.

Then I thought that I would be more realistic and aimed for lower output current for the details aiming for 10v and 50mA.

I forgot to swap the output capacitor to a lower value and burned out the second chip. It worked initially but couldn't keep going and burned out after about 30 seconds

I lowered that cap to the right value and tried again and burned out another one after about 1 minute.

It was after that that I went looking at different calculators and Dave's video to try and get an idea of what I was doing wrong.

------------------------
I get the impression that the internal darlington is what's burning out.

The calculator I was using had the Drive Collector resistor at very low values (50 Ohm and 100 Ohm) which I realised was unnecessary after I had watched Dave's video on it. I'm wondering if this was an issue.

I rolled my own inductor through a large bead using suitable current rated wire and i'm wondering if the build quality of the inductor could be so bad that it could cause my burn outs.

I'm worried about the behaviour of the IC as the battery supply voltage drops. Will it shut off nicely as the voltage drops below 3 volts or will it burn out the chip?

I'm surprised by the variations in the value of the Ipk sense resistor as I change values in the calculators. Am I better designing it for a higher current and a smaller sense resistor and only run it at low current or should I aim for lower output current and larger sense resistor using a design that's closer to my expected current usage?

Is a higher frequency safer than a lower frequency?

What areas of the circuit do I need to pay attention to ensure that the circuit is "robust"?

[c4757p]

Schematic? How exactly have you connected this? Are you sure it's exactly like the datasheet says?

It'd be so much easier to help you if you'd just draw up a diagram. Perhaps even include a photo of the actual setup.

If you do everything right it shouldn't blow up. That's part of what that Ipk resistor is for - safety cutout. I suspect something is connected wrong.

[m12lrpv]

Schematic? How exactly have you connected this? Are you sure it's exactly like the datasheet says?

It'd be so much easier to help you if you'd just draw up a diagram. Perhaps even include a photo of the actual setup.

If you need to see a schematic for this then I don't think you can help. Thanks anyway.

[c4757p]

Schematic? How exactly have you connected this? Are you sure it's exactly like the datasheet says?

It'd be so much easier to help you if you'd just draw up a diagram. Perhaps even include a photo of the actual setup.

If you need to see a schematic for this then I don't think you can help. Thanks anyway.

I had a much longer, more helpful response ready to post, but if that's how you're going to be I don't need to post it.

[m12lrpv]

I had a much longer, more helpful response ready to post, but if that's how you're going to be I don't need to post it.

Then why did you post your first post?

[c4757p]

Because we all make simple mistakes and I wanted to lend a second set of eyes to help check for any. I figured you could prepare a diagram while I prepared a more complete response. But don't come asking for help and then bitch when you don't like the help you get.

[Rick Law]

I had a much longer, more helpful response ready to post, but if that's how you're going to be I don't need to post it.

Then why did you post your first post?

Probably because of c4757p's helpful nature.  He helped me out a few times and I am appreciative of his input and impressed by his knowledge.  I for one always look at what c4757 had to say because chances are, something useful was said.

Rick

[m12lrpv]

Because we all make simple mistakes and I wanted to lend a second set of eyes to help check for any. I figured you could prepare a diagram while I prepared a more complete response. But don't come asking for help and then bitch when you don't like the help you get.

A photo would cover that but you also asked for a schematic which indicates that you don't have the familiarity with the IC in question to answer my question.

I'm grateful that you wanted to help and I don't want to put you off helping others but my questions were more specific than "have I wired this correctly".

Thanks.

[c4757p]

MC34063 is one of my "default" ICs for quick switchers. I'm quite familiar with it. But not every application of the chip exactly follows the datasheet; I know mine don't. It's much more versatile than the datasheet lets on. I've done DAC-programmable versions, SEPIC versions, high voltage versions, high current versions, blah blah blah blah blah. It's quite useful when you're okay with hysteretic control and don't want to fart with the loop stability etc.

People are also prone to taking shitty photos, so I wanted backup to cover for things I couldn't see in the photo.

Hopefully someone else answers your questions because I'm really put off by your repeated insistence that I've no basis on which to answer them.

And frankly, from the nature of your questions, I'd say you have no basis on which to decide that.

[free_electron]

coil going into saturation -> dead ic.

99% chance he's using the wrong type of coil .. but we can't tell as he is too lazy to post a picture or schematic..
And don't give us any excuses...  and don't you dare come back with 'i breadboarded it with 5meter long wires ...

Oh, and please people, can we bury that IC together with the dinosaur bones ? This is by far one of the shittiest switching regulator chips out there. it's got no protection whatsoever.

coil saturates -> smoke
too high an output current -> smoke
oscillator doesn't start ->smoke
wednesdays with northwest wind wil grandpa takes a nap -> smoke
forgot to stand on one leg while plugging it in ->smoke

these things are old and crappy.  grab a simpleswitcher from national semi (now TI)  or a switcher from maxim or linear technologies.

[c4757p]

Oh, and please people, can we bury that IC together with the dinosaur bones ? This is by far one of the shittiest switching regulator chips out there. it's got no protection whatsoever.

coil saturates -> smoke
too high an output current -> smoke
oscillator doesn't start ->smoke
wednesdays with northwest wind wil grandpa takes a nap -> smoke
forgot to stand on one leg while plugging it in ->smoke

I think you just have a blue-grey thumb! I've yet to let the smoke out of one of these! 

grab a simpleswitcher from national semi (now TI)  or a switcher from maxim or linear technologies.

I'll agree here - the Simple Switchers are very nice, and especially some of the older ones are quite reasonably priced now.

[BravoV]

Probably because of c4757p's helpful nature.  He helped me out a few times and I am appreciative of his input and impressed by his knowledge.  I for one always look at what c4757 had to say because chances are, something useful was said.

Rick

+1

Too bad the OP came with such an arrogant attitude while asking for help, a bit of humble won't degrade or hurt self pride. 

Many MC34063 related noob questions/problems in this forum were asisted and helped by c4757p smoothly, heck, I even see him too patience in helping those noobs, its your lost. 

[c4757p]

Many MC34063 related noob questions/problems in this forum were asisted and helped by c4757p smoothly, heck, I even see him too patience in helping those noobs, its your lost.

What can I say, I can smell an '063 question a mile away! So can f_e, I think.  I had him specifically in mind the last time I asked a question about it and preemptively fended off the "that-IC's-shit brigade" 

I was just getting back into electronics after a long and forgetful hiatus a while ago and a number of forum members were very helpful and patient with me. I fully credit this forum with getting me back on track, and I try to return the favor as much as possible. It's been a long time since I've absorbed anything this fast.

[Bored@Work]

This is by far one of the shittiest switching regulator chips out there. it's got no protection whatsoever.

It has a switching current limiter. Sort of ... On a good day ...

[Rick Law]

This is by far one of the shittiest switching regulator chips out there. it's got no protection whatsoever.

It has a switching current limiter. Sort of ... On a good day ...

After c4757p help me understood a board I had, I decided to experiment with the parts I have at hand and I happen to have a few  of those MC34063 from old cell phone car-adaptors.

I made some boost and some buck out of them as experimentation.  What I found strange with them was on some it works very well but on another I could not get it stable.  Depending on the voltage reference resistors and the Rs, rather counter-intuitively, the output voltage increases at certain current load and then drop back down at other current load.

Their voltage reference is suppose to be "precise" and the spec sheets said within 2%.  I had way above 10% swing.  I don't mind it dropping, but going 10% or even 20% over-volt is something I don't want.  So I put away the boost and the buck I made and forgot about them.

I actually have a need to boost 5 to 9 (or 12v) now.  So in a couple of weeks when I get the other parts, I will probably again play with the boost I made and put away.

Rick

[c4757p]

I made some boost and some buck out of them as experimentation.  What I found strange with them was on some it works very well but on another I could not get it stable.  Depending on the voltage reference resistors and the Rs, rather counter-intuitively, the output voltage increases at certain current load and then drop back down at other current load.

Interesting. What kind of resistor were you using? I imagine if it's too inductive (wirewound) it may be unfriendly, but I haven't had this issue before.

It has a switching current limiter. Sort of ... On a good day ...

Yeah, the Ipk sense can help prevent the chip from blowing up, but I wouldn't consider it a proper current limiter. It's definitely meant to protect the 34063, not the output load. At low currents, if I can afford the bit of loss, I'll often stick a current sense resistor in the output (before voltage sense, of course), then a PNP transistor across that to dump some of the output into the feedback if the current goes too high. It's much more precise and reliable. Obviously at higher output currents the 650mV loss can be significant, so I'll use a different chip or a more sophisticated current sensor.

[Rick Law]

I made some boost and some buck out of them as experimentation.  What I found strange with them was on some it works very well but on another I could not get it stable.  Depending on the voltage reference resistors and the Rs, rather counter-intuitively, the output voltage increases at certain current load and then drop back down at other current load.

Interesting. What kind of resistor were you using? I imagine if it's too inductive (wirewound) it may be unfriendly, but I haven't had this issue before.

Yes it was wire wound, and yes those inductors were savaged and unmarked and later measured to be higher than the spec sheet recommends. So you are right on both count!  Thanks for the hint.  If I decided to use the MC34063 (which is a bit under-powered for intended growth), I will unwound part of it and try again.

At the time, not sure of my commitment to do more EE stuff.  Since then, I brought a USB scope.  Using the scope, I was able measure the inductors to be approx 300uH while the spec sheet had a 180uH for it.

By the way, Tayda has 180uH inductors with color codes that looks like a resistor.  Those inductors doesn't look like they are meant to carry power and after a good bit of looking around, I didn't even find any power rating.  Would they work for this purpose?
http://www.taydaelectronics.com/catalogsearch/result/?q=180uH

Rick

[c4757p]

I meant the Rsc resistor was inductive - you shouldn't have problems with a slightly large inductor unless it's positively huge, anyway. IIRC they even call it Lmin in the table.

I wouldn't use those little axial inductors for more than 100mA or so peak. They're more for filters and such. You can test them for saturation with an improvised curve tracer, though - you might get lucky. I have used them in switching converters for tiny output currents (10mA or so) with success.

[Rick Law]

I meant the Rsc resistor was inductive - you shouldn't have problems with a slightly large inductor unless it's positively huge, anyway. IIRC they even call it Lmin in the table.

I wouldn't use those little axial inductors for more than 100mA or so peak. They're more for filters and such. You can test them for saturation with an improvised curve tracer, though - you might get lucky. I have used them in switching converters for tiny output currents (10mA or so) with success.

Weeee... 10ma!

And I need about 700ma.   I suppose I can make a 700ma booster with coffee warmer all in one for the few seconds it would run, or give up on the idea of using those little inductors.

If I take the booster out of the box (I put it away as "my first project upon return to doing EE"), I think I have to look at what resistors I used for Rsc.  I am not even sure did I put away the stable one or the unstable one.

[JUGmobile]

Hi guys,
I found this topic, so I didn't want to open a new one. I do have a big problem with my power supply based on 34063. It is cheap DC-DC 34063 step down converter and I think made by some noname chinese company. In my case I have already burned 3 34063 chips. Last one was working for a little more than 2 months, previous one was burned after few days. I burned one accidentaly (connecting it to the 37V - the capacitor on input is only 25V). Here are some pictures:

This is meant to be 5V power supply (step down from 24VDC) for attiny2313 microcontroller and 7-segment display. The output current does not exceed 50mA. No fast current drop (max 25mA). From the burned chip I can tell you that burned area is on the pins 3 (timing capacitor) and 4 (GND). On the picture above you can see that the whole GND pin of 34063 is missing. When I connect this board to the 24VDC, the 34063 chip is cold. I have tested it on desk for few days and it still was cold. Is it possible that 34063 chips are faulty?

I do not know what have I done wrong with the schematics, or the 24VDC input voltage for converter is faulty, but I have already changed that power supply with no success. Here is the schematics:

VDD is 24VDC input, VCC is 5V output for uC and display. Also I do have an diode connected serial before VDD, so there is no option to connecting the circuit with wrong polarity (+24V on ground and ground on +24V). Have I done any mistake with the choice of values for resistors, inductor or capacitors? I made this scheme with help from this "online calculator": http://dics.voicecontrol.ro/tutorials/mc34063/

Thank you for your help,
Jakob

[JUGmobile]

Input is stable 24V and output is 5V. Is there any easy solution for my circuit, cause I already have 200 of these boards.

[peter.mitchell]

Input is stable 24V and output is 5V. Is there any easy solution for my circuit, cause I already have 200 of these boards.

Maybe if you answer my original questions. What is the output current at 5V?

He already stated, does not exceed 50ma, nor does it change rapidly.

[JoeO]

Right off you can see that it is a very poor design.  Putting a 25V cap on a 24V power rail is wrong.

As AcHmed99 said, use a 7805, or for higher efficiency you can use a little switcher, a Murata 7805SR.

http://www.murata-ps.com/data/meters/dms-78xxsr.pdf

[JUGmobile]

Yes I totaly undestand that capacitor is 25V. Thats my fault, next time I'll check for things like that. I do not want an linear regulator cause I dont wan't radiator on my board. Switching regulator is perfect for my project. I do have around 200 similar boards working out there for about 6-7 years and the do work perfectly with 34063 regulator. When I run out of boards, I made a new schematics and board, since I lost original one. I'm young, I'm human, we all make mistakes. I made a mistake, trusting that web calculator, so here I am.

So AcHmed99: can I just replace 0.1E Rsc to 3.3E resistor and this should work without problems (without burning 34063), or do I need to change inductor and C3 capacitor as well? In that case, can I use 220uH inductor? I didn't found 820uH inductor that fits to my board (again I already have 200 boards soldered). And that tiny capacitor (SMD 603 case) is hard to desolder and solder a new one on board too. It would be nice to just change Rsc, but I dont know.

Can you give me an advice how to design this kind of step down converters with 34063. I will probably do some other project with 34063 and I do not want to make same mistake again. Never.

Thank you for your help!
Jakob

[JUGmobile]

Man you are helpful! Your calculations are great. I have looked a video #110 on youtube and with your help and datasheet I have calculated similar values than yours. I'll stick with 470pF Ct but I will change inductor and Rsc. Inductor I will use is 220uH and Rsc calculated with 220uH inductor in mind is 0,35E. I'll try this configuration next week.

Thank you very much AcHmed99!
Regards,
Jakob