# DC-DC step-up

Discussion in 'The Projects Forum' started by JDeen, May 24, 2014.

1. ### JDeen Thread Starter New Member

May 20, 2014
11
0
Hi!

Im looking to drive a 100-250v piezo device off a low voltage power supply (can be higher than 5v, only mentioned as point of reference). The output amperage needs to be in the vicinity of 10uA.

The output waveform should be about 100Hz (square/sine/sawtooth, pref customizable).

Im not very used to working with voltage converters, and would appreciate som pointers to where I should start the design.

Links to litterature is more than welcome.

I understand Im probably looking for a switching dc dc converter, but the number of different designs baffles me and Im unsure of where to start.
Should i start with choosing a driver (what specs would i look for? Freq alone?) or the inductor/mosfet/capacitor?
Does transformators yield any other bonuses apart from isolation/efficiency?

Thanks,
JDeen

PS:
<del>
If I multi-posted it was because I thought my scrawny iPad bugged out on submit stage. Apologies.
</del>
Sorted.

Last edited: May 24, 2014
2. ### t06afre AAC Fanatic!

May 11, 2009
5,939
1,227
Do the high voltage part need to adjustable or is any (fixed) voltage in the 100-250 range ok

3. ### JDeen Thread Starter New Member

May 20, 2014
11
0
Hi,

EDIT:
Essentially I need to generate a (custom) waveform in the 100Hz-range.
Ideally this can be modified, if it complicates matters to a greater extent I'd settle for only adjusting frequency and keeping voltage/waveform constant.

//JD

Last edited: May 24, 2014
4. ### t06afre AAC Fanatic!

May 11, 2009
5,939
1,227
I think your project is somewhat more complicated than you think. As I see it you will need a dual high voltage supply(+/- voltage). And a high voltage amplifier. It would help if you could better explain what you are trying to do. Perhaps you have some references/links that can elaborate your question

5. ### JDeen Thread Starter New Member

May 20, 2014
11
0
More than likely!

It is for a (two phase) piezoelectric pump, and I don't think you a dual polarity setup - however $\uparrow$ I have a worrisome tendency to oversimplify things.

I will look through data sheets and see what more I can supply of information.

//JD

6. ### Papabravo Expert

Feb 24, 2006
11,737
2,483
In this application you should start with a boost converter. A boost converter is one where the output voltage is strictly greater than the input voltage.

The next step is to appreciate that power out will always be strictly less than power in. This fact is encapsulated in the efficiency of the converter. I like to start with 80% to get the ball rolling.

Once you create the appropriate DC output voltage (100-250 V), you can add additional circuitry to build the appropriate oscillator at the 100 Hz. frequency.

If you run into roadblocks with a straight boost converter, then try a flyback design.

7. ### Bernard AAC Fanatic!

Aug 7, 2008
4,925
541
Might try 555 oscillator driving a small audio x former, turns 1:10 to 1:20.

8. ### JDeen Thread Starter New Member

May 20, 2014
11
0
Gotcha. Will I have an advantage by using a boost converter controller IC or will a 555 do, with appropriate circuitry?

Looking at the multitude of converter controllers I still can not find any within relevant voltage supply range. I'd be glad for a suggestion here...

As for technical data - I'm not sure if they let me re-publish the data sheets, however here are relevant key points:
• Piezo-electric diaphragm for a pump*. No reverse polarities.
• Waveform (for output, not the switching wave): Square/sine/sawtooth/custom wave, NOT AC (ie: single polarity). This could be implemented through other circuitry, assuming the power supply can handle a discontinuous drain. 20-300Hz should be possible, or a set value to start off with.
• Voltage: range should, preferably, be adjustable through 50-300V (this can be disregarded for a start)
• Amperage: ~500μA @ 300V.

I can set up a circuit design and do the appropriate maths if I can just get a starting point, hope you can lend a hand up to that point

Thanks!

JD

*PS: Forgot to mention it is a two stage pump - 2 diaphragms 90deg out of phase.

Last edited: May 24, 2014
9. ### Papabravo Expert

Feb 24, 2006
11,737
2,483
The sheer number of off the shelf components makes a definitive answer unreliable. I do know that voltages exceeding about 36 volts on an integrated circuit are rare. This suggests that any conversion scheme to achieve voltages in the range of 100-250 volts will require external components that can handle such voltages.

The one application area that I am familiar with is Automotive Paint Spray booths. Their power supplies start with 36VAC followed by DCD doublers and triplers. They end up with 10 kV at 0-50 μA

10. ### JDeen Thread Starter New Member

May 20, 2014
11
0
Yes, and I expect this.

I've been looking at the formulas for boost circuits-

Minimum Duty Cycle is defined as follows:

$D_{min}=1-\frac{V_{in(max)}}{V_{out}}$

$D_{min}=1-\frac{12}{300}=96%$

Equivalently, maximum is defined as such:

$D_{max}=1-\frac{V_{in(min)}}{V_{out}}$

$D_{max}=1-\frac{5}{300}=98.3%$

Note that neither of these take into consideration the inefficiency of the system, nor voltage drop across mosfet and corresponding power loss in the circuitry.

I get duty cycles which are high enough to indicate a switching frequency in the regions of 500kHz+. A lot of boost converter controller ICs are rated with both frequency and duty cycle. Luckily there are a selection of controllers rated for duty cycles in the range of 95-99.5%.

Based on Texas Instruments LM3488 datasheet here:

The inductor value must be set such that the inductor current does not drop to zero at any point (ie. continuous mode), and puts the minimum value at:

$L=\frac{D_{max}(1-D_{max})V_{in(min)}}{2I_{out}f_{s}}$

$L=\frac{98.3%\ast(1-98.3%)\ast5V}{2\ast500{\mu}A\ast700kHz}$

$L=1.98{\mu}H$

Does this make sense so far?

Thanks,
JD

11. ### Pebri New Member

Dec 2, 2013
1
0
Do you mean 100 kHz? My understanding is that piezoelectric transducers are for high frequencies usually in the ultrasonic region. That 10uA also has me concerned as it's extremely low. What is your application?

I once made a driver for a 40 watt piezoelectric transducer by looking up the data sheet for its capacitance, then using the usual formula for a series resonant circuit to calculate the required inductance at the desired frequency of operation.

Then I powered the circuit with square waves at the resonant frequency. This made the circuit extremely efficient.

Being a series resonant circuit, a small input voltage produced a larger output voltage so if the inductor resistance is low then your desired voltages could be met, although I suggest some form of output voltage limiting should be considered.

12. ### JDeen Thread Starter New Member

May 20, 2014
11
0
As I tried to clarify earlier, these are not speaker membranes, they are designed to pump water (not air). By modifying frequency/amplitude one would be able to adjust flow speed. And yes, I'm talking Hz, not kHz. You are right regarding the amperage - this should be more around 500μA. I mention this in a later post, I apologize for the confusion.

I am not looking for the resonant frequency. Indeed the flow of water in the pump has a signature "resonance" and supposedly conforms nicely with a special driver waveform which resembles a mix between sine and square (from what I can deduce, based on the SRS of the system) - yet this is not required to drive the pump.

Any tips on how to implement this?

13. ### Papabravo Expert

Feb 24, 2006
11,737
2,483
I am deeply suspicious of any design with such a narrow duty cycle range. The whole idea in a DC-DC converter is to have a sufficient duty cycle range to regulate the output voltage. Please don't conflate the creation of a DC voltage for a power supply and the waveform to drive the piezo. They are different animals. The duty cycle range for a boost converter needs to be much wider than a couple of %-age points.

I would break this problem in two and solve them separately.
1. DC Power Supply
2. Wavfrom generation
It will be harder to wrong if you do it this way.

14. ### JDeen Thread Starter New Member

May 20, 2014
11
0
Hi again!

Thanks for pointing that out.
I tried to communicate this earlier, but unsuccessfully so. Mainly the piezo waveform is mentioned to indicate what kind of load the circuit will be driving.

Maybe I should have modified the post to only concern the DC-conversion, but I was not sure if they could be connected by design with benefit.

Thanks,
JD

15. ### Papabravo Expert

Feb 24, 2006
11,737
2,483
I don't know the answer to that question. On the theory that we must learn to crawl before we walk or run a marathon I would focus on how to get to a working prototype quickly. Once you get there a process of successive refinement will get you the rest of the way.

My point:
Would it kill you to start with an AC-DC power supply capable of outputting 100-250VDC so you can get the driving waveform correct and only then tackle the problem of DC-DC conversion?

16. ### JDeen Thread Starter New Member

May 20, 2014
11
0
That might be possible. I have to admit I was most curious about the DC-DC part which is entirely new to me.

Let me give the waveform generation some thought and I'll get back to you!

17. ### Bernard AAC Fanatic!

Aug 7, 2008
4,925
541
" Simple 12 V - 180 V boost converter using the 555 as controller."
http://www.dos4ever.com/flyback/flyback.html Flyback Converters for Dummies.
I have used similar ckts for 5 V to 120 V, 5 V to 24 V using a 4.5 mH inductor with added winding.

absf likes this.

Aug 7, 2008
4,925
541
19. ### Bernard AAC Fanatic!

Aug 7, 2008
4,925
541
Short outline of quarurature control: 555 osc, 2 ea toggle flip-flops, & inverter for clock, or 555 osc. driving 4022 or 4017 count to 4, 2 ea OR s; outputs use 2 high V transistors. R depends on capacity of piezeos, prob. around 1M to 200k, wastefull but simple.

• ###### DC-DC Step-up 00000.jpg
File size:
76.1 KB
Views:
44
absf likes this.
20. ### ronv AAC Fanatic!

Nov 12, 2008
3,659
2,817
I didn't think you could do it with a conventional boost regulator, but you can.
Do you have a link to your micro pump? I'm curious about the capacitance. If it is to high this won't work.
You might need to make R1 say a 500 ohm pot with 100 ohms in series with it so you can adjust the duty cycle to make up for tolerances.

Edit. I used 12 volts as the input. It's hard to get to 300 starting with 5.

File size:
82.5 KB
Views:
40