Simple high power boost converter

[dermot]

I have a requirement to generate around 100V at 1.5A into a resistive load from a 14.4V LiFePO4 pack (rated at 60A continuous).

I've tried a couple of 600W and 1200W eBay off the shelf boost converters but they:

a) fall a bit short on voltage, typically 80V max.
b) have a maximum *input* current limit of 10 - 15A maximum continuous which limits power a bit.
c) are a bit bulky, bearing in mind this is for a portable application.

(but those converters worked just fine to trickle charge my 72V ebike battery from 12V in my car)

Since I'm only running power into a ceramic-cased heating element, I don't need much in the way of regulation or even mind if it is ac or dc output. A very quick check of the sealed element shows that it has a low inductance, so it shouldn't mind a reasonably high frequency.

I'd wondered about a beefed up version of a power blocking oscillator (a giant Joule Thief) based on a mosfet and ferrite torroid. As it happens, my junk box has a stack of IRBF4110 power mosfets (used to repair ebike controllers) and some 40mm or so torroids of the same sort as the eBay converters. Since those fets have a maximum operating voltage of 100V, I'm assuming I'd need to put a secondary winding on the torroid to step up to safely achieve my 100 or so Volts output.

Anyone have thoughts on a minimum parts count way of doing this, before I warm up the soldering iron and blow up a few components?

 

[ronsimpson]

I've tried a couple of 600W and 1200W eBay off the shelf boost converters but they:
............................
b) have a maximum *input* current limit of 10 - 15A maximum continuous which limits power a bit.

A 1200 watt eBay "thing" should pull 100A from the 12V battery. If they have 10 - 15A input current then we are talking 120 to 200 watts.

 

[dermot]

A 1200 watt eBay "thing" should pull 100A from the 12V battery. If they have 10 - 15A input current then we are talking 120 to 200 watts.

Indeed so, but those modules are designed to work over an input range up to 60VDC so can achieve somewhat higher power than on 12V. Being cheap eBay kit, the spec is a bit over-optimistic too.

The efficiency falls quite a bit at lower voltage input; for a given power output, they get *much* hotter on 12V than using the 48V I also tried them with.

 

[ronsimpson]

they get *much* hotter on 12V than using the 48V I also tried them with.

Something to remember when you start building. That is how things work.

ceramic-cased heating element

Probably made to work at 110VAC? Is there any way to get a different heat element?

a giant Joule Thief

I think it is not very efficient.

 

[dermot]

Unfortunately not, size and shape is rather specific for the load, it is rated for 1KW at 240VAC, but I only need the rather lower power that 100V or so will provide.

I'm winding the torroid with quadrifilar windings of ten turns of 18ga wire and will drive those in pairs with a pair of mosfets in push pull using the complementary outputs from an LS628 vco I have laying around.

I'll vary the frequency to see where it looks happiest (ie lowest no load current) then try an overwound secondary for a step up. I'm guessing somewhere from a few tens of KHz up to a few hundred.

All a bit of a voyage of discovery, really; I was hoping to find someone that had done something a bit similar already.

 

[ronsimpson]

Why not get a 12V to 110/220VAC "car inverter"?
I have some. If your were here it could be working in minutes.

 

[AnalogKid]

Don't know where you are located, but in the US we have this thing called Harbor Freight, a chain of discount tool stores. I have one of their 200 W inverters. The output amplitude and frequency wander, it generates some RF noise, etc. But it works and is over 75% efficient at high loads. Not bad for $20. Also - ebay.

ak

 

[dermot]

You may not have seen my point c) in my original post about it being for a portable application and the eBay modules at 6x6x9cm are already a bit big.

Last year, I tried one of the single unit 120v inverters that plugs into a cigarette lighter socket - it caught fire after a couple of minutes. I've searched online fairly diligently and if there was an off the shelf widget available, I'd hope to have found it by now, hence feeling the need to roll my own.

I have 45 years experience in various areas of electronics (a lot of RF) but not a great deal with inverters/SMPS, hence my looking for ideas. I'll widen my net a bit and see if there are other resources that might help.

 

[dermot]

Don't know where you are located, but in the US we have this thing called Harbor Freight, a chain of discount tool stores. I have one of their 200 W inverters. The output amplitude and frequency wander, it generates some RF noise, etc. But it works and is over 75% efficient at high loads. Not bad for $20. Also - ebay.

ak

How big? The eBay module I'm playing with is 6x6x9cm which is bigger than I'd like (And can't handle the current required at 12V). One car inverter I played with didn't seem to rated for continuous operation close to full power into a resistive load, more intended for low duty cycle or for running a laptop, I guess. I'll have a look at Harbour Freight since I travel to to the US from time to time.

 

[AnalogKid]

Linear Technology has a wide variety of switching power supply controllers. I don't know if they have a multi-phase boost circuit, but they do have single-phase boost parts with an internal 10 A switch. I've used it very successfully, but not with two in parallel. Still, getting that to work might be easier than dealing with a custom transformer. With an input current of at least 13 A, that would be a tricky little puppy.

ak

 

[dermot]

Linear Technology has a wide variety of switching power supply controllers. I don't know if they have a multi-phase boost circuit, but they do have single-phase boost parts with an internal 10 A switch. I've used it very successfully, but not with two in parallel. Still, getting that to work might be easier than dealing with a custom transformer. With an input current of at least 13 A, that would be a tricky little puppy.

ak

Hmm, perhaps two in parallel with the outputs summed by a pair of Schottky diodes? Do you think that would help keep it stable? Can you recall the part number?

 

[AnalogKid]

LT1270. It is in a 5-lead TO-220. These are not super-fast parts, so the inductors will be large. If you use their parametric search for boost converters, it will lead you to it, and maybe to newer better parts.

ak