Switching Boost Regulator + Current Source

[freeskier89]

Hey there,

My goal is to deliver anywhere from 10 to 100mA into a 500-1000 ohm load and I need to be able to set the current with a microcontroller. And to top it off I need to do all this from two series AAA NiMH cells (2.4V). I won't need to deliver 100mA at 100V continuously from two AAAs; the output will be delivered in pulses, where I plan to store and release the charge from a tank capacitor.

So my main question is how feasible is it to design a switching boost regulator that boosts 2V to 100V? Do you guys have any good resources to find more about the practical design of boost regulators? Or do you have any advice yourself?

I am trying to figure out how I am going to layout this architecture. My initial thought is that everything (including the microcontroller) should be powered with 1.8V so I don't have to do dual step up conversions because then I would be multiplying efficiency losses for the two conversions. Also there will be quite a bit of time where the load will be getting no power, and I don't want the wasted power of the first 1.8V to system voltage boost regulator running full bore.

Sorry for that confusing description. I attached a graphic of the general layout of the circuit. Do you think everything looks fine? This super-simplified diagram is only to be looked at to get an idea of the structure I am talking about. There are obviously small things that I omitted and such.

Please let me know if you want me to elaborate/reword. Thanks a ton!

 

[dknguyen]

Hey there,

My goal is to deliver anywhere from 10 to 100mA into a 500-1000 ohm load and I need to be able to set the current with a microcontroller. And to top it off I need to do all this from two series AAA NiMH cells (2.4V). I won't need to deliver 100mA at 100V continuously from two AAAs; the output will be delivered in pulses, where I plan to store and release the charge from a tank capacitor.

I would think that current would kill your batteries very quickly even in pulses. 100V@100mA from a (1.2x2)V source is AT LEAST a 4A draw- way too much for those cells to handle I would think- or pushing them very hard.

So my main question is how feasible is it to design a switching boost regulator that boosts 2V to 100V? Do you guys have any good resources to find more about the practical design of boost regulators? Or do you have any advice yourself?

You would need several stages rather than boosting all in one stage (has to do with transistor switching and other factors, not to mention at high gains very small errors in duty cycle result in very large changes in output voltage- not good). Personally, if you are building for practice I'd just build one stage and if I needed that voltage I'd buy one (but I tend to have expensive projects and other things to work on than the power supply). My prof says a gain of 10 per stage is pushing it- better to stick around 3 or 5 per stage.

Sorry for that confusing description. I attached a graphic of the general layout of the circuit. Do you think everything looks fine? This super-simplified diagram is only to be looked at to get an idea of the structure I am talking about. There are obviously small things that I omitted and such.

LOoks good from what I can tell. I'm a bit shakey on spotting functionality in op-amp circuits though. Not enough stages to provide the voltage gain you want though.

I am trying to figure out how I am going to layout this architecture. My initial thought is that everything (including the microcontroller) should be powered with 1.8V so I don't have to do dual step up conversions because then I would be multiplying efficiency losses for the two conversions. Also there will be quite a bit of time where the load will be getting no power, and I don't want the wasted power of the first 1.8V to system voltage boost regulator running full bore.

Reword please- battery capacity aside, you're going to need more than one stage to reach your desired voltage anyways and it's going to take up a lot of room on your uC. And a switching regulator controlled by your uC couldn't power the uC- it needs power to run the regulator so how can it ever power up to run the regulator to power itself? Even if you could, it wouldn't be making the most of each stage.

 

[ericgibbs]

hi f89,
Is it possible that you can tell us exactly what you are going to do with the current pulses, whats the project for??

 

[ronsimpson]

I would change the coil into a “Zeon Flash Transformer”. A transformer or tapped coil will give more ‘boost up’.

Move the current sense resistor to the ground side of the load.
To go from 2.4 to 100 volts will not be efficient.
Dknguyen is concerned that you cannot get that much power from two AAA batteries. Because of your pulse application, you may need to store energy on a large capacitor, then take the energy from the capacitor not the batteries.

 

[freeskier89]

Thanks for the replies; those are the sort of things I was looking for. I need to stick with high-side current sensing for my application. I like the idea of just making a single stage flyback converter with a xenon photoflash transformer, but, correct me if I am wrong, wouldn't that have the same problem as a single stage single inductor boost converter of not allowing for very precise adjustments due to errors in duty cycles? Also, all the flyback models I can find only give the output voltage as the input voltage times the ratio of the windings where as with a single inductor boost converter, you simply adjust the duty cycle to adjust output voltage. So am I correct in seeing the flyback converter as only a way to multiply the voltage by a constant set by the ratio of the turns on the transformer?

Thanks dknguyen for the figure that I should only aim to increase the voltage by 3 to 5x per stage. I was wondering about where the practical limit was, and I couldn't seem to find any sort of figures in my research. So I guess I will have to use three stages. Knowing that I am going to need to reselect a MCU and select one with at least 3 high-speed PWM channels. Let me know if there are any MCUs that come to the top of your head. I doesn't need to do much more than control this boost converter. Unfortunately, its seems like most dedicated PWM controllers would be hard to set digitally short of hooking up a DAC to the duty cycle control input and such.

For simplicity's sake, we can probably omit the battery dilemma at the moment. If I have to power it with massive lithium ion cells I will. Out of curiosity, what would you expect the efficiency to be for this at 10 and 100mA?

Thanks again!

 

[freeskier89]

Sorry to double post. I just was researching for another project and found this IC caught my eye: https://www.electro-tech-online.com/custompdfs/2007/11/MAX749.pdf. It appears to me to just be the boost converter that I was describing, but they have it hooked up to output a negative voltage. How can this chip pull off a 50x voltage gain? How could it be doing anything special? Am I missing something drastic?

If I can't replace the MAX749 with my own MCU, I am thinking I can probably use this IC if I just put another sense resistor at the output and digitally adjust the output voltage accordingly. 64 steps should be sufficient for resolution for setting the voltage. Do you guys see any problem with this?

 

[dknguyen]

Thanks for the replies; those are the sort of things I was looking for. I need to stick with high-side current sensing for my application. I like the idea of just making a single stage flyback converter with a xenon photoflash transformer, but, correct me if I am wrong, wouldn't that have the same problem as a single stage single inductor boost converter of not allowing for very precise adjustments due to errors in duty cycles? Also, all the flyback models I can find only give the output voltage as the input voltage times the ratio of the windings where as with a single inductor boost converter, you simply adjust the duty cycle to adjust output voltage. So am I correct in seeing the flyback converter as only a way to multiply the voltage by a constant set by the ratio of the turns on the transformer?

A flyback converter is just a buck-boost converter except the inductor has been repalced by a transformer that splits the circuit into two isolated halves-I don't know about the windings. The windings ratios probably let you get some extra voltage gain.

Thanks dknguyen for the figure that I should only aim to increase the voltage by 3 to 5x per stage. I was wondering about where the practical limit was, and I couldn't seem to find any sort of figures in my research. So I guess I will have to use three stages. Knowing that I am going to need to reselect a MCU and select one with at least 3 high-speed PWM channels. Let me know if there are any MCUs that come to the top of your head. I doesn't need to do much more than control this boost converter. Unfortunately, its seems like most dedicated PWM controllers would be hard to set digitally short of hooking up a DAC to the duty cycle control input and such.

**broken link removed**

The dsPIC30F2020- has 8 dedicated PWM channels with all sorts of SMPS stuff like phase shifting PWM, current reset PWM, current limiting PWM, and dithering of the internal clock. CHeck it out especially the SMPS PWM module and the comparator module. Or any other IC in that datasheet actually. But one seems too little to do much (But yours is a very dedicated thing so maybe), and one seems overkill for most things.

 

[kchriste]

The way I would deliver the 100V 100ma pulses would be to charge a capacitor up to 100V via a boost converter and pulse the 100V with another FET. This will keep the peak current in the boost converter to a minimum. Now the question is; what's the max duty cycle of the pulses? This will determine the max current you'll need to draw from the batteries to keep the cap charged up to 100V or whatever voltage pulse level you intend to output.

 

[Hero999]

2V to 100V in one stage shouldn't be a problem but you need a flayback transformer rather than a single inductor. Flyback transformers give a larger voltage gain than just the turns ratio. They rely on the back EMF in the primary being steped up by the turns ratio. For example, if you've got 5 turns on the primary and 100 turns on the secondary and you get a 20V pulse when the current is inturrupted you'll get 400V on the output when you might nomally only expect 40V from a 3V source.

 

[freeskier89]

Thanks guys for the assistance. I will probably end up using a flyback transformer, but I'm still very interested on how the max749 can use a single inductor and do almost exactly what I am wanting. Using a single inductor would be preferable. I keep looking at the typical application circuit in the datasheet and as far as I can gather, it is configured exactly like a normal boost converter but it is backwards so it will output a negative voltage. Am I missing something?

 

[walkura]

Yes i think you are right (good seen)
Thats what we call a inverting buck *note the polarity of the diode *
My approach would be to get the voltage up to about 10 volt & follow that with a sg3525 pushpull .
Simple ,cheap & effective
(try to get switched mode powersupply's by Pressman)
That book will explain everything you need to know & more

By the way use nicad instead of NiMh i think there change of survival is higher

 

[audioguru]

The little AAA Ni-MH cells in my electric RC model airplane power its motor with 5A plus its receiver and servos just fine. The battery is made by Sanyo.

Energizer shows that their AAA Ni-MH cell works fine with a load up to 1A but at 1.8A or more it is in trouble.

 

[freeskier89]

So should I just try to make a prototype with a single inductor? I still can't seem to find anything saying how my original idea of a standard boost converter is any different from the application circuit in the MAX749 datasheet? Maybe I should do an initial boost up to 10V in a preliminary stage to be safe and then boost that.

 

[kchriste]

You could just use a tapped inductor in a flyback config instead of a 2 winding transformer. Google for "autotransformer" for more info.