SMPS boost converter help FET question

[huricaine]

So I have been reading into SMPS topologies and theory and I am working on building a boost converter to boost 1.5V from a few D-cells upto 30V for a portable audio amplifier. I am using a simple circuit for now to get a feel for it.
I have and am using a tl494 IC powered on a 9v for my PWM circuit.
My main question for now is about the mosfet switch. I know that FETs are like capacitors and do not draw current like a normal non reactive circuit but they do still have current flow to charge/discharge the gate right? I read that I must be able to supply enough current to charge/discharge the FET.
My question - I am going from 1.5V to 30V, my FET is a single HUF75344G, how much current do I possibly need to properly charge/discharge the gate during operation or I should say what current ability does my gate driver need to be able to supply? I have put together a bit of the circuit except my gate driver so I need to have an idea of how much current is needed so I can choose my parts/etc..

 

[ChrisP58]

The gate drive current needed will depend on your switching frequency, the mosfet gate capacitance, and how fast you need to turn the mosfet on and off. You might get by with an amp or so, but can't say for sure without some numbers.

You say your input is 1.5V coming from a few D-cells, so I assume that they are in parallel. Can you put them in series? A higher input voltage will make the conversion easier and more efficient.

Also, are you locked into using the tl494? While I'm sure you can make it work, other chips would be better. My choice would be the UCC28C40. It has a one amp output without needing a separate gate driver that the tl494 will need.

 

[huricaine]

Thanks you are a lot of help Chris.

I have two D-cells in parallel. I didn't realize a higher voltage would be more efficient I should have known but yes I will definetly switch to series. The only reason I was using a tl494 is because I have a lot of them I ordered a bit for amp repair a while back. I understand they are not as convenient or meant for this sort of application I just like to use whatever I can so I can get experience with different parts/circuits but I am not set on it.. I also have a few UC3526DW I am not sure the differences since I'm new to PWM circuits but that is what I am trying to learn. What is the difference between the UCC28c40 IC you mentioned and one of these? It looks to have a push-pull driver integrated, I still have to run that output to an external switch right? Is my FET mentioned in the first post a good choice? I chose it because out of all the FETs I have it had the lowest Rdson value, and it has a high current supply ability.

Any Notes and tips you can give me and calculations on how to choose wht I need to design it would be nice as this is all for a learning experience..

I will order a few of those ICs now.

 

[ronsimpson]

Now I think you want to boost 3V to 30V. At what power level?

 

[huricaine]

Yes, My audio amplifier that will be the load will have the 30V rail, the amplifier circuit will be this basically: **broken link removed**
It is an opamp driven class B amp that is supposed to be able to swing closer to rail supply then the opamp normally can due to the way the output transistors are connected. I will guess/assume an efficiency of 65% for overhead, the nominal impedance load for the amp will be 8 ohms, I think it could possibly handle 4 ohms aswell and I would like to design the power supply with the overhead of being able to handle that calculated power.

Oh oops...the amp circuit is rated to produce 15W rms.

 

[ChrisP58]

The amplifier you posted is designed for a split supply of + and - 15V. You can make it work with a single +30, but you will have to AC couple the input, output, and R1. And bias pin3 of IC1 to the junction of two 10K resisters from +30 to gnd.

 

[huricaine]

Yep I forgot to mention that I will be using that amp in single supply fashion. 30 and Gnd not dual supply. I'm going to use an input/output coupling/decoupling cap, PNP transistor grounded instead of -vcc etc... and feed the opamp reference half vcc with a voltage divider.

 

[kinarfi]

Personal opinion, 12 watts from a couple of D cells doesn't sound practical to me and are you trying make an amplifier work or learn about smps, smps are real fun and interesting. 12 watts out of 3 volts is 4 amps and that's a lot for a D cell and unless you you get the correct FET, 3 volts may not turn it on completely, your HUF75344G has a threshold of 2 to 4 volts, the easiest way to boost voltage is a simple switcher, 555 timers are fun also.

 

[huricaine]

I didn't think much about how much current a pair of D-cells would handle the current I figured it by the typical Ah rating that D-cells have and figured I could produce 10W for about an hour. I have built some amateur amps before but this is not at all about the amplifier, it is about smps. I have a few dozen simple switcher buck converters that I have used a lot that work good for 3A and down regulating voltage but I have been reading into discreet smps circuits I am still new to them but I wanted to get my feet wet with a simple boost converter making a portable audio dock station. 10w total(after efficiency etc) is what i'd realistically want to try and meet which means my true audio power is going to be less but is fine by me but like you said around 3A does sound like a lot. Maybe I should just use 4 D-cells, 6V to 30. Will a little under 2A be too much of a load? I know it would be an equivalent of 1 batt as far as current capacity goes. Oh and for now I have been using a 9v battery just for the tl494's supply I got it oscillating with a saw-tooth wave and on the output I got it to about an 80% duty cycle without feedback at over 3v. When I connected it to my boost supply circuit (my FET gate) it didn't work and then I remembered about the current supply. I just wanted to take steps messing with it until I was ready to do everything properly.

As far as the IC supply, can't I boot-strap it by using a large capacitor? I have some large caps that I do not mind putting into my project as space is not a big deal. Btw I do not know much about bootstrapping but it seems to make sense as long as that capacitor will hold onto that 30V all the time(Large capacitance)?

 

[Maged]

What I would do is go like this
1)If you will be satisfy with only 10 watts of output, and efficiency of 65% then a total of 15 watt is needed from the battery (minimum)
15 W / 3 = 5 Amps, will the D cell provide that? the best quality cell will do but may heat up because AH of a battery does not mean necessarily one hour of discharge time but the max current may be less for a longer time.
2) all the regulators discussed above need a minimum of 8 volts for VCC to start running and most of them has a feature to turn off the unit if the supply is low. even the 555 will start at 4.5 volt unless yo use 7555 to start at 2 volts
the solution may be in a blocking oscillator that can start at low voltage

 

[alec_t]

The cost of D-cells being what it is, wouldn't a 6V or 12V SLA battery be cheaper in the long run to provide the current you need?

 

[huricaine]

What's an SLA battery Alec? I just wanted something cheap and available...I bought a bunch of D cells for pretty cheap but there obviously going to be cheap in quality. I always wanted to mess with d cells, especially with boost converters to see what I can push them to accomplish. I am considering a totally different battery but I want it to be something I can buy easily and relatively cheap or if I pay a bit for it it should have good capacity to justify price.

But maybe I should use even more D-cells? I can get a 3-pack for a dollar so I can use 9 for $3 basically. At the same time I am also wanting to contrain the space a relatively small dock box because I also want to see how much sound I can get from a small package, something I have always liked to achieve. I care less about length of the box versus the height and depth. So a bit of D-cells aren't impossible.

If I have to I may say screw it and not worry too much about the enclosure size, and use a bit more speakers to make it louder(or better frequency response with surprising low end extension). I have 1.5" x 1" speakers. If I can build a pretty small package, and sacrifice power due to battery capacity and still get good output from a pair of these speakers then I would be pretty pleased! But the way I am I want to push the envelope and create a small system that has surpising low end frequency extension like I said which I know is nearly impossible. I planned on porting them to a low-ish tuning.

Anyone know a lithium battery or something I can get on ebay to handle decent power/current? I am going to check ebay rightnow.

 

[huricaine]

ooo, check it out: **broken link removed**

I'd be willing to pay the price...I just need to make sure that it can deal with the current draw. Edit: I just realized that the batt in the link isn't AC mains style. Others I have seen for US power are a bit more in price. I looked up SLA battery and realized it stands for sealed leads acid like an AGM batt. Not sure if I found anything great.

 

[ChrisP58]

An audio amplifier does not draw constant power unless it is driving a constant signal into a constant load. It's actual power draw will vary with the volume and dynamics of what you are listening to.

More battery will certainly give you more runtime. How much runtime for how much battery space and weight will be difficult to predict until you have some actual experience with the system.

D-cells are primary devices, meaning they are not rechargeable. SLAs and Lithium Ion are secondary cells that can be recharged. That makes them less expensive in the long run.

 

[huricaine]

I am considering making this a bigger sized dock and not worry about size as much. I want to use a 6.5" peerless HDS woofer I already have. I have been thinking about using a cordless drill battery. Any idea how much power these batterys can supply? I am thinking of going a bit higher in power. Size will not be such a concern this way and I can use it at work.

 

[alec_t]

good output from a pair of these speakers

At 1.5" x 1" are you serious? What is their RMS power rating (not 'music power' !)?

I have been thinking about using a cordless drill battery. Any idea how much power these batterys can supply?

The manufacturers' web-sites should have the info.

 

[huricaine]

Actually they are 3" x 1.5" I believe...I'm not around to measure them but they are small for sure. I always wanted to try and push small speakers to their limit I don't expect anything crazy. But after thinking about it a small ipod-like dock would not cut it so I figure I should just build an enclosure for one of my peerless HDS ported on a bit more power. The voltage will need to be boosted a lot higher, i'll use larger batt(s) and amp power.

Btw those small speakers were rated at 5w each I believe it was the actual RMS rating but I can't be sure...I was going to treat them like it was the RMS rating and see how things went.

So I guess I am moving on to change the project to a larger dock. I want to add a toolbox like handle to the top and keep the height on the short side and make it the proper length and width needed to meet the volume target. I will probly add eq through negative feedback on the opamp aswell which I have dabbled with before. For now my focus is finding the correct battery, decide how much power I can safely push, etc. I have an amp IC already tht I want to use. Would a different IC for the PWM be desired at this point or that UCC one?

 

[dr pepper]

Nimh's can deliver high currents so in theory it'd work, for a short while.

Take care on the kind of amp you use, high frequency noise from the switcher can make an amp unstable and blow up.
The tl494 is a veritable chip and ought to do the job.
I like to use etd series transformer cores, theres an online calc to work out windings, the issue with ferrites is not to saturate them.
Also as you found out the supply voltage for the smpsu controller might be an issue, maybe a little joule thief type circuit just to power the tl494 might be the way to do it from a low voltage supply.
You can get fet gate driver ic's, but they are critical on board layout and bypassing, I tend to use a pair of transistors from the chip to the fet, if you get a fast switching time dissipation from the fet is lower.

This is an interesting project, keep us posted.

 

[huricaine]

Thanks dr.pepper. I am thinking of using my complimentary SMD bjts pzta06/56 for the gate drivers using a number of them in parallel for higher current handling. I have never used multiple transistors in parallel but have always wanted to try it and get experience with it and learn it. As far as I know, I am supposed to parallel all terminals except the emitter, adding ballast resistors with a value under an ohm right? I will use the bjts as an emitter follower pair(with parallel multiples). I noticed tlx94 circuits use a pull-down resistor with the emitter follower pair, is this because the IC cannot pull down all the way to ground?

 

[dr pepper]

I wouldnt have thought you'd need more than one trans on each side for the gate drive, if your using more than one fet maybe, but still you'd be better off using different higher rated trans.
However if you did emmitter resistors are a good idea, start with 0.1r, the resistance balances out the diffo in base emitter voltage variances and ensures the load is shared.
The tl494 can pull either way, there are 2 uncommitted tranny's in the '494, you can connect the emmitter to ground so the collectot will pull to gorund, or you can connect the collector to + so that the emitter will pull up, if you configure the chip so that the o/p's alternate you can connect one side to switch + and the other to switch ground giving true push pull, if you do this dont forget to make use of the deadband input on the chip to reduce run through.
Heres a circuit I use on the o/p of the '494, r2 and r3 I use 10 ohms, the others 2k2:
http://blog.savel.org/2005/08/29/tl494-magic-chip-part-5/
Note another cool way to run 2 mosfets on the same load with this ic is to use alternate mode so that one fet switches on one cycle and the other on the next, I tried this on a dc motor controller worked great, no worrying about interaction between the fets or drives.