How to make a dc amplifier which will convert 3v to 9v

[Tricker]

i am working on a project which requires a power supply to operate a servo motor of rating 19v, 4amp. maximum voltage i am allowed to use is 12v DC. so i want a circuit to amplify a 3v supply to minimum 9v. i have searched a lot for this circuit but that require an inductor. please suggest me a circuit without inductor.Any suggestion for this help are welcome.

[Dave_Oblad]

Hi Tricker,

Here is a link to an old style converter circuit, it's a charge pump (for conception). I've used variations of this design many times over the past years.

http://electroschematics.com/wp-content ... ematic.gif

Otherwise, the newer single chip versions have very few external parts and very high efficiency. I assume you have "Google" where you live? Just Google Voltage Doubler Chips. Maxim makes several versions that take advantage of super high speed switching that reduces the external cap sizes. Most do not require inductors.

You do know that to get 19VDC @ 4 amps from 12V supply, that the 12 volt source will have a draw over 8 amps?

Power = Volts times Amps. No way to get around that fact. (Include losses due efficiency)

so i want a circuit to amplify a 3v supply to minimum 9v.

Why? What does this have to do with the initial requirements of 12V=>19V. Also, you have to be careful about ground commons. Transformers at least provide some isolation, which it seems you are not allowed. Is this a school project?

Best wishes,
Dave :^)

[Dave_Oblad]

Hi again,

Here is a link to higher powered converters if you don't mind using a small simple coil Inductor. It's a Buck-Boost approach and doesn't use a transformer. An LT3972 seems to be able to output about 8 amps, from the selection table. But anyway, lots of choices available. Your 3V to 9V target isn't explicit about the amount of current you will require.

http://parametric.linear.com/html/step-up_(boost)_regulators

The pump charge approach above (in my last post) have some issues with high current outputs because of the speeds required to charge larger caps for more current abilities. These boost converters however, can deliver a lot more current and are still very small, if size was an issue. I use them a lot. My last one output 4 amps at 7.5 volts and the inductor was just a small coil about 5mm in diameter and height. Over all, this type of converter is much smaller than any other approach I know of.

So..Why are you avoiding the use of an inductor?

For additional info:
http://en.wikipedia.org/wiki/Boost_converter

Good luck..

Regards,
Dave :^)

[Tricker]

thanks dave for your posts. bt i need a circuit in which maximum input shold be 6v and output should be 19v. thye circuit which you suggested requires an input of 12v. please if possible suggest me a circuit which takes input 6v.
waiting for your reply.......

[Dave_Oblad]

Hi Tricker,

Besides changing the spec to.. what is it now.. 6V => 19 volts, you still haven't specified the current level in amps that you will need. Why not? It is a major aspect of the design.

Anyway, try actually clicking on the link I provided and enter the specs into the little boxes. It will then tell you what device(s) will meet your specifications. Here is the link again:

http://parametric.linear.com/html/step-up_(boost)_regulators

For example, I entered 3 to 6 volt input, 19 volt output, and 4 amps. It suggested a Linear chip called: LT3757A... as 3 to 40 volts input, 3 to 350 volts output, and up to 25 amps, depending on the voltage you want. Of course it does require using a small Inductor coil, but you still haven't explained why you are loath to use one. I already pointed out that voltage doublers have power (amps) limitations and offered the better design choice as being a booster circuit. They are small, easy to use and rather cheap.. so what's the problem?

Click on the device name, download the specs in PDF format, fill in the blanks and build it. If you want something already pre-built then you may have to shop around a bit.. and expect to pay 10X what the parts cost alone.

Sorry, that's the best I can do.

Regards,
Dave :^)

[Tricker]

Actually there is a problem to find the inductor which is not available in the market and also in the college.We can't order too & we cant build it too.

[Dave_Oblad]

Hi Tricker,

So this is a School Project? I hope you are not trying to Trick someone into doing your homework ;-P

Anyway, I've shown you the path.. it's up to you to walk it.

Best regards,
Dave :^)

[DavidCyr2000]

This technique is interesting for low to medium current draw, but we would like to draw as much as 80 Amps at 18 Volts for our application for periods of 30 seconds max. Is there any way with this method of achieving such high current draw? Could we put several circuits in parallel? Are there high current solutions today that weren't available when this thread was started? Other possibilities? Thanks!

[Natural ChemE]

DavidCyr2000,

Welcome to the forums!

While it's a neat question to address here, I'd forewarn that you'll want to put some good consideration into that particular setup.

One issue's the amount of electricity you're drawing. For 80 Amps over a 18-Volt drop, you're looking at 1440 Watts, which is closer to at least 1600 Watts after you factor in the loss of AC to DC conversion. Many residential circuit breakers are set to trip at 1800 or 2400 Watts, which is pretty easy to hit, especially during the startup spike in draw.

Anyway, 18V power supplies are pretty common. And, yeah, you're supposed to be able to run them in parallel to increase their total current rating, so in principle four 20A power supplies should give you 80A.

Just a standard disclaimer that this is a lot of electricity, which can be dangerous to life and property.

[DavidCyr2000]

Thanks for your reply, and yes we appreciate the fact that this is an extremely large power draw, so appropriate precaution will be designed into the application.
Could you please recommend a circuit with specific components that would convert 12 Volts to 18 at 20 Amps?
Thank you!
David Cyr

[Natural ChemE]

DavidCyr2000,

Could you describe the situation a bit more?

I mean, if it's in an indoor setting, just getting a power supply would seem to be the ticket. But I'm guessing that this is a mobile application? What's providing the 12V current? And whatever that source is, is it already a series? For example, if your source is already a bunch of 6-Volt elements in series of two, that's 12V, but you could rearrange them to be in series of three to get 18V.

Also, if you've got 80 Amps at 12 Volts, that's only 960 Watts. At 100% conversion efficiency to 18 Volts, you'd only get about 53 Amps.

[DavidCyr2000]

We are in the design choice phase of this project, so a couple of possibilities are;
1) The application could be driven by a 6 volt battery in series with the 12 volt vehicle battery/alternator. A grouping of 6 volt lithium-type batteries could be dynamically switched (regrouped) to 12 volts for recharging from the 12 volt vehicle alternator and then, "on demand", used in series with the vehicle 12 volt vehicle battery to supply the required 18 volts, or,
2) Use the 12 to 18 volt converter we are discussing that would eliminate the need for the extra batteries and complex switching circuitry.

Specific to your question; "...if your source is already a bunch of 6-Volt elements in series of two, that's 12V, but you could rearrange them to be in series of three to get 18V". The 12 volt supply is a single mobile vehicle battery with a conventional alternator keeping the 12 volt battery charged. We appreciate that this current draw is beyond the capacity of the alternator to "keep up" with the short-term demand, but that is acceptable since this demand will be infrequent so the main 12 volt battery won't become discharged.

If the charge pump approach is not feasible, then we will have to investigate the switching methodology for additional on-board batteries... I hope this provides adequate information to help you recommend the converter approach! Thanks! David Cyr

[Natural ChemE]

DavidCyr2000,

Okay, gotta say, this thing that you're working on sounds like fun! If you stick around, it'd be neat to hear about how the design competition goes.

I don't know much about the product market for devices like this. From a quick check on Google, there's stuff like this "SMAKN® DC Converter Module power adaptor Regulator 12V(8-16V) Step up to 18V 15A 270W" on Amazon.com, which seems like you could put a few in parallel. But for all I know, there's a company selling a more reliable, higher-load solution for less money elsewhere.

[DavidCyr2000]

Functionally, the SMAKN is the type of device we are trying to design and build. The issue is whether we can scale it up far enough to achieve 120A+ at 12V IN and max 80A OUT at 18V. It appears there are N-channel MOSFETs that will handle up to 600A and P-channel up to 180A. So, a charge pump SHOULD be able to produce 80A out assuming the circuit is done right. But that becomes the question. What circuit and what specific things we'll need to do to handle the physics of such a circuit that will no doubt dissipate a significant amount of heat.

Some high current MOSFETs:
600A N-channel - http://www.digikey.com/product-detail/e ... ND/6053919
200A N-channel - http://www.digikey.com/product-detail/e ... ND/1880374
180A P-channel - http://www.digikey.ca/product-detail/en ... ND/5413742
Charge Pump Circuit - http://electroschematics.com/wp-content ... ematic.gif
Do you think this circuit is an over simplification for illustration purposes or would it work as is? The values in this diagram give a frequency of 875 Hz. It would be interesting to see the switching wave form at the output. The higher the frequency, the smaller the capacitors need to be, but the greater the heat loss…
Do you think this charge pump actually doubles the 12V? We only want 18V. I guess making the capacitors “too small” would effectively reduce the output voltage, but make it choppy, but an inductive motor shouldn’t care…
Thoughts???
David Cyr

[Dave_Oblad]

Hi DavidCyr2000,

You have indicated that your end system doesn't need to source supply such heavy currents for long duration's. Since one can't get more power out than the available power in.. then surges are usually overcome by adding Large Caps as energy storage containers to the Input side to accommodate surge requirements.

The devices I suggested earlier in this thread do not produce Free Power. But they sidestep Ohm Law as far as heat loss is concerned. I convert 18VDC to 5VDC all the time in my circuits using these. Power lost to heat is minimal.

If your power output exceeds the load capability of the battery source such that the battery output drops below a specific voltage then it seems your are fighting an up-hill battle.

You indicate 1440 watts in and 1440 watts out with no loss.. ok maybe.. But how long can the source supply 1440 watts of power? How fast does the Battery output degrade under full load? Batteries have a built-in resistance that limits output power rather dramatically and rather quickly.

In reference to your Gif link above..
I used the following circuit to produce -12V from my 12VDC source:

I turned it sideways to keep as much detail as possible.
Right Click on it and open in a New Window.

I run it at 8KHZ and it allows about 250mA loading before the Caps (470uF) can't sustain it. Getting the driver Transistors to switch properly without them biting each other can be a bit tricky...lol.

I think such a circuit is a dead end for your massive power requirements.. to be honest.

Regards,
Dave :^)

[DavidCyr2000]

Hi Dave,
The circuit is to be driven by a heavy duty car battery with the alternator running at high RPM, so I would expect the battery bus voltage to drop less than a volt below the running 13V. Starting a car motor draws 300A - 400A with a drop not more than 2V below 12V. Since the 80A output current draw at 18V will last between 20 and 30 seconds, (once per hour), capacitors won't; "...accommodate the surge requirements".
Yes, there will be losses. The heat loss will increase as 555 timer frequency increases. So there is a trade-off between capacitor size and heat sink size that we can likely only optimize with experimentation. Given the input capacitance of high current MOSFETs, we will probably have to drive them with an intermediate impedance matching stage. The capacitors should be ESR for best operation; type TBD.
Not ready to give up yet!
David Cyr

Reference from: http://www.aa1car.com/library/charging_checks.htm
ALTERNATOR AMPERAGE OUTPUT (excerpt)
In addition to checking the alternator's voltage output, you also need to check its current or amperage output. Amperage is how much current the alternator generates at a specified voltage and speed. Not long ago, an 80 amp alternator was considered a high output unit. Most late model alternators produce 120 to 155 amps or more. Current output increases with engine speed, from around 20 to 50 amps at idle up to the unit's maximum output at 2,500 RPM or higher

[Dave_Oblad]

Hi again,

Oh.. I was under the impression that a battery was the only power source allowed. But if you can augment it with gasoline powered electrical generator(s) then I fail to see the challenge.

My above circuit (modified) might implement super fast Fets and Megahertz Frequencies, but it will never yield more than 50% of its input power because it uses a 50% duty cycle for creating the power output.

The other types I've used employ an inductive kickback to get an increased voltage without a transformer.. but again, while they tend to run rather cool, they still exhibit a significant drop in power-in to power-out efficiency.

Anyway, good luck and let us know how it goes...

Sidebar: My first real job in late high school (summer vacation) was at an auto-electric repair shop. My most frequent customers where those with Blown Alternators. They had left their car running while giving someone a jump start. The regulator causes the alternator to carry the lions share of the current to maintain battery voltage. The diodes in them are not meant to carry the amount of current necessary to start a car. We made a fortune on replacing Alternator Diodes..lol. Not sure how much of that still applies today in modern cars.. but I never jump start someone with my engine running.. to be safe.

Regards,
Dave :^)

[DavidCyr2000]

Interesting that you mention the "half wave" duty cycle issue just as I took a look at making it "full wave" by having two circuits; one as previously described [http://electroschematics.com/wp-content/uploads/2009/12/555-voltage-doubler-schematic.gif] and the second with an inverter driven off the 555 of the first circuit, driving circuit two MOSFETs. This should halve the current each circuit needs to provide and also reduce the size of the single output capacitor. See attached. Does this make sense?
We will definitely try much higher frequencies than that used in the sample circuit.
Agreed; always shut off motor when boosting another car...

[Dave_Oblad]

Hi again,

Good idea. You will have an issue with the transistors fighting each other. When the base(s) are at 6 volts.. both will be on full.. and clamping a short across the power supply. It is tricky to get rid of that effect and still run fast enough.

The bias resistor in the 2nd circuit will unbalance the base drive current.. might be better to use a 2nd 555 triggered out of phase with the first one.

Best wishes,
Dave :^)

[Dave_Oblad]

Hi again,

Ignore the first comment in my previous post. I hadn't noticed the configuration of the transistors avoids overlapping ON states. More about that in a moment.

Have you analyzed the circuit I supplied for the strategy of switching the base currents? You may note the symmetry of my circuit is broken by R14.. that's because the drive source/sink attributes of the 555 are not symmetrical (sinks more than can source, current wise). So R14 is an assist on the high side.

C27, CR5, CR6 do most of the work in speeding up the switching times while R15, & R16 sustain the on/off state as the current through C27 tapers off (charges). C25 and C26 quiet down the base noise on Q4 & Q5.

The efficiency of the whole is related to the main frequency of course, but higher Frequency means more issues during the slope periods from the switching driver.

In using Fets.. you will still have the same issues at high frequencies.. plus a new issue. Bases in Transistors look resistive by nature and are pretty fast. The Gates of Fets look like Caps with a charge and discharge time built in. They will add additional current consumption from the driver as the Frequency goes higher, and of course, can over heat or be damaged if run too fast.

So, while output efficiency goes up with frequency, power loss at the Gates will also go up.. obviously defeating what you hoped to gain by using higher frequencies.

Of course wise selection of the best characteristics for the drivers and the strategy employed to prevent them from fighting each other will be key to a successful (best) design (but don't expect any miracles..lol).

Using the devices I suggested in my first post is a better avenue of pursuit IMHO. They get a voltage increase from inductive kickback and many allow use of external drivers (fets) for max current outputs with minimal losses due to heat build up. They side step ohms law regarding heat.. meaning when full ON then current = max while resistance is virtually zero and when full OFF current is zero with virtually infinite resistance. That's a pretty neat trick..lol. They regulate the desired voltage by rapidly switching between two ideal states (phase modulation) that bypasses heat build up (loss).

BTW.. I like the technique used in your version to avoid both T1 & T2 from turning on at the same time. The downside is that using smaller values for C4 reduces the Emitter/Base currents and thus the available current that can be drawn from this design (I'm pretty sure). The smaller C4 becomes, the more like variable resistors (rather than switches) the transistors will appear.

That's clever.. but it treats both as variable resistors and is why you don't get +24V output.. besides the Diodes and Base/Emitter threshold drops. Available current from using transistors as variable resistors will have some restrictive effect on available current output. My circuit gives a hard -12V output with decent current because it uses the Transistors more as Switches than Variable Resistors. The downside is figuring out a simple strategy that avoids both from having over lapping ON states.

See.. my design required 24V from 12V at a minimum of about 270 mA and minimum loss due to heat (my load floats). I tried your circuit first (text book source).. yes it gave me almost the voltage I needed but the available current drawn sucked (that's a technical term...lol). Also note my circuit is over 30 years old, things have improved since then.

Anyway.. I leave the rest up to you.. have fun and good luck.

Best wishes,
Dave :^)

Ps. This is like my 10th edit of this post.. I'll stop now.

[DavidCyr2000]

Hi Dave,

If the power devices are regular transistors as they are in your example, then the diode/resistor/capacitor arrangement ahead of the power section should work fine, however, with MOSFETs, the gate voltage is very different so I think we would need some other anti-fighting scheme.

We are also looking at the LTC3787 as you suggested: http://www.linear.com/parametric/Step-Up_(Boost)_Regulators#!cols_1038,1036,1646,2167,1035,1034,1367,1033,1032!s_1035,0!gtd_!1032_<=12!1033_>=14!1035_>=60!1034_>=18!1646_Boost|Buck-Boost
They are rated at 60 amps although their example shows a circuit with max 10 amps output. Not sure if it can be scaled up to 60 amps or not...

Thanks again for your guidance!
David Cyr

[Dave_Oblad]

Hi David,

I think the path towards the LTC3787 (and like devices) is best. May lead to a smallish highly efficient solution that doesn't require an anvil for a heat sink...lol.

Good luck again,
Dave :^)

[rossweinberg21]

Good project!! Keep up your good work Tricker. I gain a lot of knowledge here. Thanks for all contributors especially Dave. He shared some tremendous information here.