Fox-Dickson-Boost DC/DC no special parts el cheapo

Hi all,

Ok so here I am doing exactly what everyone else is doing and that is mainly building programmers/debuggers thanks to the hard work from all the usual suspects on the forums.

One thing I did notice was quite a lot of talk about getting Vpp from USB +5v and the use of Maxim and Ti parts that are either impossible to get samples of or cost a fortune, inductors are a pain in the bottom, everyone wants a charge pump one, but they all need oscillators and that seems to be where everyone gives up and gets a maxim/Ti chip AND a bunch of capacitors.

Welp here ya go, I know a lot of the more seasoned users here will recognise the major parts of this circuit but I have not seen this design any place else on the net and its perfect for generating Vpp at a very small mA load such that Vpp presents.

All the bits are bog standard stuff I had laying about stuck to the carpet, in my bed, in draws ment for knifes and forks and other places where electronics end up if you love the art.

As you can see from the occiloscope plot Vpp reaches 11v in under 10ms and seesm to run at about 11.6-12.4 volts stable in the real world depending on the tollerence of your components, I included a 10K pull up (R5) from 5v to simulate a typical load on Vpp which you can remove or keep depending on what you want to do with your Vpp. But remember that the charge pump requires a load to function, largest value for R5 realisticaly is 250KΩ, but better is 100KΩ

I've tested the design on the bench and it seems fine, but I have not used it to program a pic with yet as that'll come later, but its such a simple & cheap design any issues should be easy to work out.

(worst case ripple at Vpp is 154mV, you could reduce this with a larger value C5, with C5 at 10µF, time to 11volts is 39ms and ripple goes down to worst case 7.5mV )

(Changing R1 and R4 from 330Ω to 33Ω gives more mA output but is at the limit of oscillator stability(according to the simulation anyway), play with the values if you need more power)

Edit: R1 and R4 at 100Ω seems to give the best real world performance

**broken link removed**Ω

Fox-Hickson-Boost DC/DC no special parts el cheapo

hey! what program is this, where u simulating ur schematics?

Re: Fox-Hickson-Boost DC/DC no special parts el cheapo

BlackOps said:

hey! what program is this, where u simulating ur schematics?

Click to expand...

Its National Instruments Multisim, part of the NI Ultiboard package. I use it for commercial work but I "think" there may be an educational version.

**broken link removed**

The Ultiboard package links with the Multisim schematic capture so you can forward and back annotate to a PCB design, quite a nice package. Its supposidly the only real competition to OrCAD which I personally hate.

If you can use eagle you can use this, very intuitive and now comes with a MCU sim addon too so you can simulate the PIC, 8051, etc. in your circuit as well as the component level stuff. All based around the SPICE simulator engine so you have all the normal quirks like maximum time step problems and such, but it does come with a nice auto fix feature that makes (most of) the problems go away

Just use a MAX232 as in my Inchworm design for VPP generation from 5V. Cheap & common parts like ST232 work fine.

Yup, normally I use a max product of some description to do it, but I thought it would be fun to design a solution costing <50cents excluding the capacitors you have to use even with the max products, also useful if you don't RS232 have on the board. I priced the parts up for it here in england and even from rip off RS the cost comes to 36 pennies

Your discrete design probably works fine for generating a boost voltage like VPP that does not have any significant current requirement.

If you need a boost voltage without some amount of current, your design will be challenged. The issue is the 330 ohm resistors in the collector circuits. All capacitor charge currents for the second stage are going through these capacitors. In IC charge pumps, they use a push-pull or totem pole transistor arrangement so that there is always a low impedance path to the capacitor terminals.

Of course, two resistors are much cheaper than two additional transistors, so you design economically exploits that fact.

Hi banjo, yes I totally agree with your assessment of the circuit, I have been playing with the R1 and R4 values and found 100Ohms to be a pretty good compromise, using that value the circuit is able to generate about ~3.2mA at the Vpp tap while maintaining a voltage of ~10v.

I have been looking at the Push Pull Boost systems employed byn Maxim/Ti etc. but am still at some of a loss as to how they are doing it in silicon as all of the examples of that method I can find require a either a sizeable inductor or a coil of some sort.

Around a year ago a friend of mine began making a low wattage DC/AC converter for driving a small water pump and he used many aspects of the push pull design with a 555 timer generating the clock and later a PIC 10F200, his initial efforts were somewhat less than successful but eventually the design worked with the use of a hand wound transformer with centre taps.

I am sure it must be possible with only minor modification to increase the output of the posted design to create 10-20mA, dropping R1 and R4 to 33 Ohms does increase power delivered by quite a a lot and would put it in-line with commercial charge pump chips but the oscillator starts to become unstable at those values and can even stall depending on load, I think a good play with the transistor types Q1 and Q2 along with adjustments to possibly higher quality Tantalum caps for C1 and C2 as the price for these has dropped dramatically recently.

My goal with this design is 20mA at 10v from 5v in at a total cost below 1$

You might want to take a look at charge pumps made with hex inverter ICs. With six gates in a package, two gates are used to form the oscillator. The remaining gates are parallelled to for the capacitor switching portion.

At what volume is your design goal? I can get a single hex inverter IC from Digikey for $0.50. However, if I buy thousands, the price drops below $0.20.

, i am only talking 10's off, and the price is to include all parts caps included.

The project is not a commercial thing just an attempt to create something of use for fun. My electronic background is purely digital, my analogue design friends call me a philistine as I trust auto routers LOL so I am just messing about with fairly simple circuits without any IC's or easy ways out to give me a better understanding, I have a little free time on my hands after a large project so I am having fun

OK, quick update for you, if R1 and R4 are changed to 20Ω and R2 and R3 are changed to 500Ω then the circuit will produce a minimum output of 10mA at 10V, you can change the resistor R10 to 500Ω and run the simulation to get the same figures as I have also changed C5 to 10uF

**broken link removed**