Topics Topics Edit Profile Profile Help/Instructions Help Member List Member List  
Search Last 1|3|7 Days Search Search Tree View Tree View  

DC-DC converter

:: EPE Chat Zone ­:: ­Radio Bygones Message Board :: » EPE Forum Archives 2007-2009 » Archive through 14 August, 2009 » DC-DC converter « Previous Next »

  Thread Last Poster Posts Pages Last Post
  ClosedClosed: New threads not accepted on this page        

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 140
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 12:18 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi
This one is a bit esoteric ( thats OK among consenting adults ) It concerns the popular DC-DC chip MC34063.
To design a step-up/boost converter you have to decide on the Ton and Toff ratio of the oscillator. I'm using a 12V SLA and I want about 100V out. So, I want to go from 10 to 100V approx. According to the datasheet, you require a Ton of 90% of the duty cycle. This sounds like utter nonsense to me. What happens if you want 200V or even 500V?
Can anyone enlighten?

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

alexr
Frequent Contributor
Username: alexr

Post Number: 63
Registered: 02-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 01:43 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Plugging 200V into the formula you get a duty cycle of 95% and for 500V its up to 98%.

I agree that this seems ridiculous but really all that the figures are telling you is that such high levels of voltage step-up are not practical with a simple boost configured converter.

10V to 100V is probably do-able, just, but for higher boost levels you should be thinking about using a flyback transformer circuit arrangement.
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 141
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 02:32 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Alex,
Thanks for that.
My starting point for this is a website by Ronald Dekker. He describes a flyback, but also a simple boost converter which gets 180V from 12V for Nixie tubes. My application is similarly low current, so there really shouldn't be a problem. He also gets his 180V by a 50/50 duty cycle.
He doesnt use a 34063, but a 555, although I cant see the difference. Unfortunately, his site is littered with errors and he doesnt answer emails so I'm stuffed.
And I'm puzzled.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

chippie
Frequent Contributor
Username: chippie

Post Number: 243
Registered: 11-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 02:42 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

What sort of level are we looking at for current at 180v?

How about using a small audio transformer fed by a 555 with a Cockroft Walton multiplier....

I found this on google...

http://www.techlib.com/electronics/highvolts.html



(Message edited by Chippie on 30 June, 2009)

(Message edited by Chippie on 30 June, 2009)
eccentric millionaire financed by 'er indoors
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 142
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 07:12 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi chippie,
Yes, it is possible to make a high V supply with a reversed mains tranny, but I want to get to grips with SMPS theory.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

chippie
Frequent Contributor
Username: chippie

Post Number: 244
Registered: 11-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 30 June, 2009 - 07:22 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Ahhh...I see..

Have you tried the National semiconductors website?
They have an E bench for that sort of thing...
eccentric millionaire financed by 'er indoors
Top of pagePrevious messageNext messageBottom of page Link to this message

alexr
Frequent Contributor
Username: alexr

Post Number: 64
Registered: 02-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 01 July, 2009 - 06:13 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

The duty cycle calculation tells you what the duty cycle needs to be for continuous operation, i.e. no dead time between cycles. You can of course drive the regulator with a lower duty cycle but it will give lower efficiency and more ripple on the output.

The MC34063 chip is not a good choice for your application since its switching transistor is only rated for a Vce of 40 Volt. Also it can't go above 85% duty cycle but this only effects efficiency, it could still work if you use an external switching transistor.

For a good run-down on switching regulators take a look at http://www.maxim-ic.com/appnotes.cfm/an_pk/2031

The Linear site also has some good applications notes at http://www.linear.com/designtools/app_notes.jsp#power
Top of pagePrevious messageNext messageBottom of page Link to this message

stylers
Frequent Contributor
Username: stylers

Post Number: 157
Registered: 04-2005

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 01 July, 2009 - 01:45 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

some of the maxim MAX chips might be better for step up purposes - there is a design for a nixie supply somewhere on the net that uses one - it has very good efficiency but the PCB layout and component choice for that efficiency is rather critical
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 143
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 01 July, 2009 - 05:31 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

OK guys, thanks.
I think I may have grasped the essentials of designing the cct, so I'm currently making said circuit with a frequency, duty cycle and, more importantly, inductance all calculated by me. Only time will tell if I got it right. If it works I'll be very smug; if it doesn't work, I shall just say I never wanted one anyway.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 144
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Friday, 03 July, 2009 - 07:37 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi,
Theres a couple of points that I want to pursue.

I dont know if anyone has had time to view Ronald Dekkers site, but I wish you would.
Its called Flybacks for Dummies or somesuch.
Part of the trouble I have concerns his use of the 555 chip. Configured as an astable, he claims its running at 15KHZ, while I make it nearly 30KHz. This makes a nonsense of his calculations and so makes learning tricky. He doesnt answer emails.
A basic boost converter uses an inductance. I calculate that I need 1mH, but thats not the important point. I have tried to make one using various ferrite torroids I've salvaged. A big surprise to me is that the intrinsic inductance of these torroids varies by a huge amount; two or 3 orders of magnitude. Of course, by using a torroid with a large A(L) you need fewer turns, which makes life easy, but I feel there has to be a catch, and my guess is saturation. Is that right? How would I check for saturation?

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

alec_t
Regular Contributor
Username: alec_t

Post Number: 28
Registered: 03-2009

Rating: N/A
Votes: 0 (Vote!)

Posted on Friday, 03 July, 2009 - 08:00 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

....How would I check for saturation?...

I'm no expert on this, but I would try putting a primary and secondary winding on the toroid, driving the primary with a sinusoidal signal, monitoring the secondary voltage on a scope, and bumping up the primary drive until the onset of clipping of the secondary voltage is seen (which would suggest the onset of saturation).
Top of pagePrevious messageNext messageBottom of page Link to this message

alexr
Frequent Contributor
Username: alexr

Post Number: 65
Registered: 02-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 04 July, 2009 - 05:39 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

The circuit on Dekker's site shows a 2n2F timing capacitor, in his calculations he is using 4n4 as the capacitor value. The circuit values as shown will indeed give a free-running frequency of about 30KHz but since he is not calculating or using the optimum value of inductor its not really important. The output voltage is controlled by the feedback circuit rather than the inductor.

You are correct about toroids, the higher the permeability the low the saturation current. This is why in a switching supply you would generally use a iron powder toriod rather than a ferrite.

The testbench on Dekker's page would be a good way to test the saturation current or else a plot of the converter input current against out output current while increasing the load, the point at which the input current sharply increases will indicate saturation.
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 145
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Sunday, 05 July, 2009 - 09:46 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Thanks Alex.
I reckon he simply got his sums wrong. He almost certainly used a 2n2, as 4n4 doesnt exist.

Confident that I understood the maths, I wound a choke and ran up the circuit. You can imagine how I felt when it didnt perform! I expected 100V, but got 60V maximum ( no FB ).
By inserting a small resistor in series with the choke I was able to determine that the current in the choke rose to 240mA during Ton, whereas it was calculated to rise to 350mA ( at least that explains the 60% of expected. ) It looks like the choke may be the culprit, although why its current rise is muted I dont know. Unless it has something to do with saturation.
I didnt know that torroids could be made of iron powder OR ferrite ( in fact, to be honest, I thought they were the same thing! ) Is there any way of telling which is which just by looking at them? Or holding them up to the light? Or seeing how many times they bounce when you skim them on the sea?
I have in mind to make another choke with a torroid with a lower Specific Inductance ( which I can measure ) and see if it performs closer to that calculated.
If you have any other bright ideas, I'd be only too glad to hear them.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

gordon
Frequent Contributor
Username: gordon

Post Number: 531
Registered: 05-2005

Rating: N/A
Votes: 0 (Vote!)

Posted on Sunday, 05 July, 2009 - 12:19 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

This application sheet for the MC34063 may be useful. There is a high voltage/low current design on page 27, which could be a starting point. With the selection of the inductor core, there are so many types available, it is probably try it and see without any data.

What about using the inductor core from a broken CFL bulb, or maybe something from a SMPS in a broken/unwanted appliance.

http://www.intusoft.com/onsemipdfs/AN920-D.pdf
Top of pagePrevious messageNext messageBottom of page Link to this message

alexr
Frequent Contributor
Username: alexr

Post Number: 67
Registered: 02-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Sunday, 05 July, 2009 - 12:21 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Bruce:

I knocked up the circuit on a breadboard just to see how it would perform and it works fine for me.

I tried it with a 100uH toroid wound on an iron dust core and a 500uH toroid on a ferrite core, both worked fine but I am only driving a neon. No doubt if I try to draw any sort of current the ferrite will saturate.

Ferrite cores are generally plain uncoated dark grey while iron dust cores are epoxy coated with the colour of the coat indicating the core properties. Generally a yellow or blue/green are suitable for SMPS.

One thing to watch for is you must use a fast or ultra fast diode. A normal power diode such as a 1N4007 will not work. In fact when I replaced the fast diode with a 1N4007 the output dropped to 70 Volt and the feedback had very little control. This sound very similar to your symptoms.
Top of pagePrevious messageNext messageBottom of page Link to this message

gajjer
Frequent Contributor
Username: gajjer

Post Number: 226
Registered: 05-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Sunday, 05 July, 2009 - 04:23 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Bruce
so whats this for - an electric fence round the dolls house!

Since you have determined that the max current is only 240mA I am guessing you have a scope. If the current is rising linearly then the core is not going into saturation. When a core saturates its inductance falls and the current rises more rapidly.
Are you sure the transistor is hard on? And of course that the resistance of the winding is low.
Remember V= Ldi/dt + IR

I'd also agree with alexr about the diode.

How's the dolls house doing?

cheers
gaj
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 147
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 08 July, 2009 - 07:55 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hello Guys,
Thanks for all that. BTW Gajjer, it certainly is for a dolls house electric fence: I dont want any riff-raff.
I did make a mistake in the calculations: the energy IN is related to Ton, but the energy OUT is going out all the time, not just during Toff
Nevertheless, there was still a shortfall of energy IN.
The choke wasnt satutating: there was a linear ramp, but the slope was low. Why? Well, I dont know, but........
I reduced the inductance of my homemade choke to 600uH. I got my 100V, but it really should have been much higher than that. I tried a 680uH choke. This wasnt a torroid, but a bobbin type pulled from a TV set. It outperformed my homemade, which must go to show that measured inductance aint the whole story.
I also tried a homemade 1mH choke using a 15mm torroid. It was green, plastic covered, but with a high specific inductance. Despite its high measured inductance, it acted like a short circuit and torched the 10Amp FET in seconds.

What does this all mean. I dont know, but clearly the maths I'm using which underpins the action of chokes is too simplistic. I just wish I knew more.

Its at times like this that you realise the importance of a sense of humour.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

epithumia
Frequent Contributor
Username: epithumia

Post Number: 670
Registered: 06-2006

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 08 July, 2009 - 09:21 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

At a guess, I'd say the 1mH choke saturated.

Recently I discovered saturable reactors, where the core material is designed to rapidly switch into the saturated state. One moment the coil has a very high impedance, the next it's just a short circuit.

They are used as efficient power switches. The example I worked on was a multi-output step-down converter. Since only one output of the transformer was inside to control loop, the other outputs used saturable reactors in series with the transformer output to regulate those outputs.

Of course, the look just like ordinary inductors. Very confusing if you've never seen them before. confused

I hate magnetics. I've studied them in depth and gained a good understanding of them, but the information doesn't seem to stick!

Epi
If you need me, Neil and me will be hanging out with the Dream King. - Tori Amos
Top of pagePrevious messageNext messageBottom of page Link to this message

ant
Frequent Contributor
Username: ant

Post Number: 446
Registered: 05-2006

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 08 July, 2009 - 10:19 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hello Bruce et al,

Surely you need to be a squatter just to see in through the windows!

Regards Ant
Top of pagePrevious messageNext messageBottom of page Link to this message

zeitghost
Frequent Contributor
Username: zeitghost

Post Number: 1294
Registered: 01-2006

Rating: N/A
Votes: 0 (Vote!)

Posted on Thursday, 09 July, 2009 - 08:36 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Everyone hates magnetics... that's why that smpsu design book is £300...
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 149
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 20 July, 2009 - 02:06 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Everybody knows that magnetics is an invention of the Devil. Ant, your jokes get worse and worse. Good God man, they're almost as bad as mine.
I've had a bit of a play with the circuit. I'm going to try and build a particular device of my own design, partly to see if I really do understand what is going on. Watch this space.

Bruce
Top of pagePrevious messageNext messageBottom of page Link to this message

bruce
Frequent Contributor
Username: bruce

Post Number: 157
Registered: 04-2008

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 05 August, 2009 - 09:52 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hello guys,
I've been playing with this flyback a bit more.
One thing I do now understand: the MC34063 is designed to work in continuous mode, whereas I'm using discontinuous; the maths is different.

I made a circuit which, predictably, didnt work right. But it didnt help that the square wave I was using was mutilated by the capacitance of the MOSFET. The oscillator uses CMOS gates which have limited current output, of course. How could I sharpen up the pulse? I was thinking of just using an npn transistor.

Bruce

Administration Administration Log Out Log Out   Previous Page Previous Page Next Page Next Page