Solid State Buck Boost Convert SMPS Design

[codex653]

ok people, i'm new to Electro Tech here and i thought i would try out the forums and see how much people can help me with me next project!

anyways, my next project that i have is a buck-boost converter switch mode power supply. I'm going to be using it to supply both a positive and negative voltage for my 150 watt RMS power amp that i built (that's actually a bit underrated...it's closer to 170w RMS). The reason why i am choosing a switch mode supply vs using a transformer is because 1. space reasons 2. cost 3. weight and 4. i don't want the transformer's electric field induction into the wires of the amp

i have the basic outline for the buck-boost converter as shown here:

http://www.powerdesignersusa.com/InfoWeb/design_center/articles/DC-DC/dc_f5.gif

now, granted i will not be using a Bipolar transistor as my switch, but instead a power mosfet...in this case, a IRF640.

I understand the basic theory behind the buck-boost in that the output voltage is determined by the duty cycle of the PWM and you can get different output current and voltage ripple regulation by the selection of the inductor and capacitor values. I know about what continuous and discontinuous modes are (even tho i dont really know how to get the inductors to physically operate in those modes)...basically i know some theory about it.

Now i know basically how i want to build this (parameter wise) but the part i get confused on is component value selection...especially on the output inductor and capacitor. Here is what i want for the supply.

I'm going to be building two separate supplies, one with a negative output and the other with a positive output but the ratings will be the same.

-Power Rating: around 140 watts continuous per polarity
-Voltage Requirements: Mains 125vac input, 30vdc output
-Voltage Ripple: preferably less than 5% under load...i want to keep a steady power
-Output Current: 3A average, 6A peak
-Output Current Ripple: again, preferably less than 5%...but i don't know what is typical for a SMPS supply so i may be going a bit low...i have the ability to make my own inductors and measure their values at the frequency i will be running them at.
-Operating Frequency: probably somewhere between 40Khz and 80Khz
-Efficiency: anything above 75% sounds good to me

I don't think i am missing anything in the basic parameters, but if i am, then please tell me

now for my driver circuit for the mosfet, i'm not gonna be using any fancy chip, just the classic 555 timer connected for PWM. That will be driving the gate of the IRF640 for the switching, and then the rest of the buck-boost just kinda takes over from there.

As for the mosfet, i believe i have chosen a good mosfet for the job as it can handle 12A and 200v, which is more than enough than what the circuit will be requiring, including some headroom for spikes too... I have also gotten the P-Channel version of the mosfet, the IRF9640, for the construction of the positive polarity supply. a buck boost converter inverts the polarity of the original dc input so i'm gonna need a p-channel for when have the input at a negative voltage. I don't need to worry about the polarity of the gate drive circuit because on both versions of the mosfet, they can be driven from a positive or negative source.

The diodes i have selected are going to have to be paralleled because i haven't found any fast diodes that can handle some high current. the diodes are rated at 1000v and 3A, but i am going to parallel 3 of them together, each with their own resistor to help equal out the current "hogging" between the diodes. also i am considering placing them on a small heatsink..but that's only if they get warm under operation. They aren't anywhere close to being the fastest diodes on the market, but this is my first SMPS so i'm willing to sacrifice a little bit of speed and some efficiency just so i can get it working.

I'll be needing a little help with the negative to positive buck-boost, as i know the component arrangement will have to be slightly different because of the reversal of the polarities compared to the positive to negative one. Other than that, can anybody help me with my design? I hope i've been pretty thorough to give as many specs as i can think of that i want for this thing...i just need some more help and guidance in understanding the final design, cause the stuff they have on google for how to do this is ridiculously confusing mathematics to me..like college level stuff that i haven't even learned yet
Thanks for all the help!
Codex

 

[misterT]

This is a useful article: https://www.electro-tech-online.com/custompdfs/2010/12/Likely_Chris_July_2003.pdf

 

[misterT]

now for my driver circuit for the mosfet, i'm not gonna be using any fancy chip, just the classic 555 timer connected for PWM. That will be driving the gate of the IRF640 for the switching, and then the rest of the buck-boost just kinda takes over from there.

Don't use the 555 to drive the mosfet. You need to switch the mosfet on and off very quickly in order to get good efficiency (and to avoid overheating). Here is a very good guide to mosfet gate drive circuits and mosfets in general: https://www.electro-tech-online.com/custompdfs/2010/12/slup169.pdf

 

[codex653]

why wouldn't the 555 be ok? i've seen plenty of videos on youtube of people using a 555 timer as the switching circuitry and it seems to work ok for them...it's too late for me now anyways, i've already ordered the timer and the mosfets/diodes for the project...

 

[misterT]

why wouldn't the 555 be ok? i've seen plenty of videos on youtube of people using a 555 timer as the switching circuitry and it seems to work ok for them...it's too late for me now anyways, i've already ordered the timer and the mosfets/diodes for the project...

You can get everything working with the 555 directly driving the mosfet. But, generating a PWM waveform and driving a gate are two different things. You will get better performance if you use a good mosfet gate-driver (between the 555 and the mosfet). Place the gate driver close to the mosfet and the 555 away from the EMI.

The 555 can sink and source ~200 mA to the gate. 200mA is quite ok, but I would still use a separate gate driver circuit or IC. You can add the gate driver circuit in your design at any time (and if) you want. The 555 is good for generating the PWM pulses, but for a really good (fast) switching action you need a good gate driver. My point is: use the 555 to generate the PWM waveform, but use a good gate driver to switch the mosfet on and off. Read the paper I linked and you will understand why.

Look at this thread also: https://www.electro-tech-online.com...n-off-mosfet-circuit.110404/?amp;goto=newpost

 

[codex653]

I am also considering doing a 2 switch forward converter with a homemade transformer.....i have some very large ferrite cylinders that probably weighs a good quarter lb or so that i can wind (painstakingly by hand), so i'm sure it has enough ferrite to it in order to handle the power requirements...if not, i can always increase the frequency. i'm starting to lean towards the forward converter tho because it would be extremely simple to build a center tapped secondary that would be able to give me the +/- dc voltage i need

 

[codex653]

but also with the 555 timer...would it not also work to use a fast switching transistor (or maybe a darlington?) on the output in order to be able to supply more current to the gate of the mosfet?

 

[crutschow]

Just to clarify, a buck-boost circuit will generate a constant output voltage for an input voltage greater than, equal to, or less than the output voltage. Is that what you need?

A buck-boost circuit is significantly more complex than either a buck circuit, which generates an output voltage less than the input, or a boost circuit, which generates an output voltage greater than the input voltage.

 

[codex653]

yes that would be correct...i need to have a 160v (rectified mains) input be taken down to +/- 30vdc.....that's why i am choosing between the buck boost, because it will invert the input, or the 2 switch forward, because i can use a transformer to get the output voltage and polarities i need....i'm not sure which one to use yet and i need help with inductor and capacitor selection since both of those circuits depend on the inductor and capacitor at the output to get good voltage and current regulation as i understand it.

 

[crutschow]

Buck boost does not pertain to inversion of the power polarity. Re-read my post.

 

[codex653]

i get what you are saying...atleast with the voltage part...but doesn't a buck boost invert the input voltage polarity???? i'm looking at a diagram right now for a buck-boost that does that...it's one of the main reasons why i was looking at the buck boost because i need +/- polarities...i'm sorry i've been so confusing :/

 

[crutschow]

You are correct. I was thinking of the non-inverting buck-boost converter.

You do want an inverting buck-boost design. Sorry for the confusion on my part. Carry on.

 

[codex653]

there is a non inverting buck boost converter?! ooo do tell! every single one of the schematics i've seen online have all been inverting buck boosts...that would be extremely useful for the positive polarity half of the power supply! that way i wouldn't have to go to all the trouble to have to use P-channel mosfets and such, correct?? anyways, do you have a schematic for it possibly? or do all you have to do is reverse the direction of the output diode and capacitor?

 

[codex653]

also, in addition, i'm looking at types of discrete gate drivers for the buck boost mosfet so that i will be able to turn on and off very quickly and therefore increasing efficiency...i found some concept circuits in this Texas Instruments page online...look at pages 12 and 14 and the diagrams. i think they would provide some solutions for faster turn on/off time?

https://www.electro-tech-online.com/custompdfs/2010/12/slup169-2.pdf

 

[codex653]

ok wow so i've done some more research on the topic and have found all the equations needed for a buck converter and also a buck-boost converter....just one thing tho....from someone's experience, how big of a heatsink would i need for a 40 degree F temp rise?...i did some math with the buck converter last night and apparently my mosfet is going to be dissipating around 5 watts max..using the thermal resistance of the case of the mosfet (which is 1 degree C/W), that comes out to be 5 degrees C which translates to 40 degrees F roughly....so..is that a big heatsink or not???? i have a ton of decently sized heatsinks but i have no clue as to the C/W rating...

 

[tcmtech]

i don't want the transformer's electric field induction into the wires of the amp

huh? A switching power supply still produces a inductive field just the same as a transformer does but with far more potential for harmonics and other electrical noise.

how big of a heatsink would i need for a 40 degree F temp rise?...i did some math with the buck converter last night and apparently my mosfet is going to be dissipating around 5 watts max..using the thermal resistance of the case of the mosfet (which is 1 degree C/W), that comes out to be 5 degrees C which translates to 40 degrees F roughly....

You are confusing temperatures here. A 5 C rise is 5 degrees C above what ever the ambient temp is. If its 25 C then the IC would be ruining at 30 C.

I'm going to be using it to supply both a positive and negative voltage for my 150 watt RMS power amp that i built (that's actually a bit underrated...it's closer to 170w RMS).

Also a +- 30 volt power supply will not push 170 RMS into a 8 ohm or even a 4 ohm speaker.


At the moment you sound like you are in way over your head and don't have any real idea of what you are building or how to even go about selecting the correct components to make this work.

 

[dougy83]

i did some math with the buck converter last night and apparently my mosfet is going to be dissipating around 5 watts max..using the thermal resistance of the case of the mosfet (which is 1 degree C/W), that comes out to be 5 degrees C which translates to 40 degrees F roughly....so..is that a big heatsink or not???? i have a ton of decently sized heatsinks but i have no clue as to the C/W rating...

A 5 degree C change is 9 degrees F change. And no, you don't need a big heatsink for 5W - to keep the junction temp below, say, 100^C, a 15^C/W or better is required (@SLC).

As far as using the 555 to drive the mosfet, I think you may find it doesn't work all that well and the mosfet will end up getting hotter than you calculated.

There are integrated switching ICs available that include the control circuit, the voltage reference and even the mosfet available for quite cheap (e.g. FAIRCHILD SEMICONDUCTOR|FSCQ1565RTYDTU|POWER SWITCH, TO-220F-5, 1565 | element14 Australia). The datasheet shows a few examples for flyback converters, which will probably be easier for you to construct than either the 1/2 bridge forward or 2 of the buck/buck-boosts. You only need to wind a single transformer as well & the outputs track each other.

 

[Blueteeth]

My two cents/pennies.

For a step down bipolar power supply, at your power rating, especially for audio, I would strongly recommend a push-pull converter. The transformer as you stated, make it easy to generate both positive and negative supplies. It can also be significantly smaller than a 50/60Hz transformer because of the higher switching frequency, and for >100W a flyback is frankly impractical. There are specific PWM IC's which will make your life a lot easier, although designing any form SMPS can be a chore, in fact, winding the transformer is generally the easy part... debugging can be lot more work.

As another option, (which has its pitfalls) you might be able to buy a premade 250W power supply, and add a winding to the transformer for the negative output. This gives you much less control over whats going on, being at the mercy of existing circuitry, but taking a lot of work out of it. At the very least it will provide a 'kit' of parts, and a PCB but you would have to know whats going on to modify its voltage output.

I would drop the 555. Its a lovely versatile chip, and I've used it in quite a number of power supplies, but never more than 20W, or for high frequency, it jsut doesn't have the speed or control for reasonable efficiency. TL494 and the SG3524 spring to mind as common cheap options as well as the UC38* series from Texas instruments (see https://www.ti.com/lit/pdf/slor091). For a simplified idea of what you're up against, check out fig.5 in this app note: https://www.electro-tech-online.com/custompdfs/2010/12/sg3524.pdf

As your amp is 170W rms, even though I seriesly doubt you'll ever reach that power draw (or even half of it) I think a 200W PSU would be the minimum requirement. Many ATX PC power supplies provide a nice 'base' circuit with components rated for 200W. Rewinding the transformer might not be that tough, as you can use the number of turns for the voltage outputs (12,5 and 3.3V) as a reference. They are also dirt cheap.

 

[codex653]

tcmtech...sorry for the confusion man! haha i've never built any of these types of converters before...well ok one, if you count the joule thief lol ....i had no clue about the whole harmonics thing possibly being GREATER than a transformer! all i knew was that when my guitar got close to the amp's transformer i'm using right now (which is freakin huge btw) everything just starts this crazy buzzing...until i shut off the power...so i attributed that to the magnetic field of the transformer...i thought that surely with the smps being a ton smaller than the other one, it wouldn't put out as much of the magnetic field as the transformer had..(i knew it would still do it some, but i was planning on shielding that)...cause the transformer's field reached across a good 3 or 4 feet!! so every time i went to change a setting on my amp, i either had to turn off the power, or dial down the volume on my guitar

again, sorry about all this...the temperature thing..i have never done it before and i used a so called "calculator" online to tell me what the whole temp rise was cause i was short on time...guess it's lessoned learned to not trust all the calculators out there! hehe

oh and yes a +/- 30v supply CAN put out 150-170 watts RMS into a 4 ohm load .....but only when you have a two channel amp with each channel kicking out on average 80 watts RMS, depending on the volume i'm playing it at! i just pulled one of those things the manufacturer's do haha. " i have a 150-170 watt RMS amp here!....BUT! it's divided into 2 channels"...yah...typical amplifier corporations

 

[codex653]

dougy83..ok i thought i had something funky with the whole temperature changes...thanks! yah, i know the 555 isn't gonna do a great job, people have been telling me that all over the place lol
i want to be able to build this entire thing from scratch..including the mosfet driver...mainly cause i want to LEARN, and also cause the project would be a benefit to my amp...i have no place to take electronics courses around where i live, and my highschool does not offer anything either...so as i've found it, the way i've learned is by researching stuff online, building it, making mistakes, learning from mistakes, etc....so i really don't want to have something that is gonna do all the work for me...i really really want to understand completely everything i'm doing, and ive found for me, that a hands on approach is best