### Author Topic: high voltage DC-DC converters  (Read 7540 times)

0 Members and 1 Guest are viewing this topic.

#### XTY

• Contributor
• Posts: 5
##### high voltage DC-DC converters
« on: January 01, 2012, 11:29:38 pm »
Here it is:
what would be the best way to produce up to 3KW nominal power output at 300 dc voltage from an 12 volt batteries?
what topology and what would be the best way to go about building such a project?

#### slateraptor

• Frequent Contributor
• Posts: 833
• Country:
##### Re: high voltage DC-DC converters
« Reply #1 on: January 02, 2012, 12:36:57 am »
what would be the best way to produce up to 3KW nominal power output at 300 dc voltage...

A generator...really.

#### Psi

• Super Contributor
• Posts: 7132
• Country:
##### Re: high voltage DC-DC converters
« Reply #2 on: January 02, 2012, 12:45:35 am »
Designing something at that power isn't going to be an easy project.
You will definitely spend more time and money building it than you would spend buying a 3kw 12v inverter and rectifying the output to get your 300VDC

One issue you will have is that most high power inverters use 24V or 48V simply because at 12V you'd need 333A  (assuming 75% efficient).
333A @ 12V is so large that you would need multiple batteries in parallel and if you had multiple batteries you'd be better to put them in series to reduce current and increase efficiency.

If you want to see what kind of switchmode system are used at that current google for inverter circuit diagrams.
You should be able to find some examples.

I imagine most of them use a push pull fet drive into a toroid at 50-200khz.
« Last Edit: January 02, 2012, 08:42:16 am by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)

#### IanB

• Super Contributor
• Posts: 9489
• Country:
##### Re: high voltage DC-DC converters
« Reply #3 on: January 02, 2012, 05:09:21 am »
I can confidently say this is the most efficient circuit for the job:

I'm not an EE--what am I doing here?

#### XTY

• Contributor
• Posts: 5
##### Re: high voltage DC-DC converters
« Reply #4 on: January 04, 2012, 08:34:51 pm »
I can confidently say this is the most efficient circuit for the job:

Wow how come i didn't see that one?

#### slateraptor

• Frequent Contributor
• Posts: 833
• Country:
##### Re: high voltage DC-DC converters
« Reply #5 on: January 04, 2012, 08:46:49 pm »
Wow how come i didn't see that one?

Because the naive sol'n doesn't incorporate any form of regulation, doesn't compensate for series resistance, and doesn't necessarily satisfy your nominal power specification to say the least.

#### XTY

• Contributor
• Posts: 5
##### Re: high voltage DC-DC converters
« Reply #6 on: January 04, 2012, 08:54:34 pm »
Designing something at that power isn't going to be an easy project.
You will definitely spend more time and money building it than you would spend buying a 3kw 12v inverter and rectifying the output to get your 300VDC

One issue you will have is that most high power inverters use 24V or 48V simply because at 12V you'd need 333A  (assuming 75% efficient).
333A @ 12V is so large that you would need multiple batteries in parallel and if you had multiple batteries you'd be better to put them in series to reduce current and increase efficiency.

If you want to see what kind of switchmode system are used at that current google for inverter circuit diagrams.
You should be able to find some examples.

I imagine most of them use a push pull fet drive into a toroid at 50-200khz.

Thank you for the eye opener...guess i need to revise my ideas to take in account practicality and efficiency...thank you

#### IanB

• Super Contributor
• Posts: 9489
• Country:
##### Re: high voltage DC-DC converters
« Reply #7 on: January 04, 2012, 09:39:05 pm »
Because the naive sol'n doesn't incorporate any form of regulation, doesn't compensate for series resistance, and doesn't necessarily satisfy your nominal power specification to say the least.

If we are talking lead acid batteries, as we usually are when considering 12 V batteries for higher power demands, then the above points are not much of an issue.

Lead acid batteries have strong self-regulation due to their very low internal resistance. The voltage hardly drops at all under load leading to a stable output voltage.

There is no compensation for series resistance required. In fact, I'm not even sure what this means? When putting batteries in series the resistance of interconnects has the same consequence with 1 battery or 100 batteries. It's a constant factor that doesn't change.

The nominal power specification of 3 kW requires 10 A at 300 V. Any suitably sized lead acid battery can supply 10 A with ease. As noted in an earlier post, a reduction in the number of batteries will only increase the current demand, meaning that the inverterless solution will always satisfy the power requirement more efficiently than any other solution.

The objection to be made, if you want to get serious, is one of practicality. How to purchase, store, maintain and charge that many batteries. Having said that, I have been in a room with hundreds of lead acid batteries sitting on wooden shelves connected by metal bus bars as part of an emergency power backup system. It can be done.
I'm not an EE--what am I doing here?

#### Time

• Frequent Contributor
• Posts: 725
• Country:
##### Re: high voltage DC-DC converters
« Reply #8 on: January 04, 2012, 09:46:30 pm »
-Time

#### Psi

• Super Contributor
• Posts: 7132
• Country:
##### Re: high voltage DC-DC converters
« Reply #9 on: January 04, 2012, 11:01:54 pm »
Greek letter 'Psi' (not Pounds per Square Inch)

#### slateraptor

• Frequent Contributor
• Posts: 833
• Country:
##### Re: high voltage DC-DC converters
« Reply #10 on: January 04, 2012, 11:19:05 pm »
Because the naive sol'n doesn't incorporate any form of regulation, doesn't compensate for series resistance, and doesn't necessarily satisfy your nominal power specification to say the least.

If we are talking lead acid batteries, as we usually are when considering 12 V batteries for higher power demands, then the above points are not much of an issue.

Lead acid batteries have strong self-regulation due to their very low internal resistance. The voltage hardly drops at all under load leading to a stable output voltage.

There is no compensation for series resistance required. In fact, I'm not even sure what this means? When putting batteries in series the resistance of interconnects has the same consequence with 1 battery or 100 batteries. It's a constant factor that doesn't change.

The nominal power specification of 3 kW requires 10 A at 300 V. Any suitably sized lead acid battery can supply 10 A with ease. As noted in an earlier post, a reduction in the number of batteries will only increase the current demand, meaning that the inverterless solution will always satisfy the power requirement more efficiently than any other solution.

The objection to be made, if you want to get serious, is one of practicality. How to purchase, store, maintain and charge that many batteries. Having said that, I have been in a room with hundreds of lead acid batteries sitting on wooden shelves connected by metal bus bars as part of an emergency power backup system. It can be done.

I have no idea what mechanism you're referring to when you say a lead-acid battery has"strong self-regulation." The internal resistance of a battery is a dynamic parameter, and varies with temperature and current draw amongst other factors. Surely you can see that as current draw increases, internal resistance becomes more significant.

A battery's internal resistance has nothing to do with interconnect resistance, and since the naive approach arranges 12V batteries in series, internal resistance will likewise be additive and thus even more significant.

The comment "a reduction in the number of batteries will only increase the current demand" doesn't make any sense in the context of the proposed naive sol'n since the batteries are connected in series, viz. removing a battery in series decreases voltage and does nothing to current since no power regulation is incorporated within the design. The comment is applicable to 12V batteries in parallel in order to reduce current draw from each in parallel, but your diagram's 12V x 25 --> 300V surrenders that benefit of the doubt.

To be frank, I'm not even certain where Psi is going with the whole inverter suggestion to begin with (perhaps he'd be willing to clarify). I mean I'm not a power engineer, but proposing DC-->AC-->DC immediately suggests poor efficiency. The question is of DC-DC conversion, and a flyback switchmode topology would seem to be a proper sol'n where it not for the high 3kW power requirement (hence my generator wiseass comment).

The objections you've pointed out, although legitimate, are of secondary concern in light that the primary goal of achieving 12VDC to 300VDC @ 3kW regulated power has yet to be satisfied.

#### IanB

• Super Contributor
• Posts: 9489
• Country:
##### Re: high voltage DC-DC converters
« Reply #11 on: January 04, 2012, 11:41:52 pm »
I have no idea what mechanism you're referring to when you say a lead-acid battery has"strong self-regulation." The internal resistance of a battery is a dynamic parameter, and varies with temperature and current draw amongst other factors. Surely you can see that as current draw increases, internal resistance becomes more significant.

The voltage regulates itself. Simple as that. In that example battery linked above, the impedance is listed as 3 milliohms. Consider for practical purposes that the internal resistance is about the same number, then the voltage drop with a current of 10 A will be 30 mV. Insignificant. It doesn't matter if it goes up and down a bit with temperature, it is still insignificant. The battery voltage itself will vary more than that with state of charge. Also, since the original requirement didn't specify any tolerance or allowable range for the 300 V output it is impossible to say whether any actual range of output voltages is or is not satisfactory.

Quote
A battery's internal resistance has nothing to do with interconnect resistance, and since the naive approach arranges 12V batteries in series, internal resistance will likewise be additive and thus even more significant.

This is not so. The voltage drop as a percentage of output voltage will be the same with one battery as with 25 batteries. Hence 30 mV in 12 V will become 750 mV in 300 V. Just as insignificant. As I said, when you put batteries in series then series resistance, whether internal or external, is a constant factor.

Quote
The comment "a reduction in the number of batteries will only increase the current demand" doesn't make any sense in the context of the proposed naive sol'n since the batteries are connected in series, viz. removing a battery in series decreases voltage and does nothing to current since no power regulation is incorporated within the design. The comment is applicable to 12V batteries in parallel in order to reduce current draw from each in parallel, but your diagram's 12V x 25 --> 300V surrenders that benefit of the doubt.

It makes perfect sense in terms of the requirements. If you reduce the number of batteries you will need a DC-DC voltage converter of some kind to get the 300 V output, and this will increase the current draw on the remaining lower voltage battery supply compared to the 25 batteries in series for equal power outputs. Like for like, fewer batteries demands more current from each battery.
I'm not an EE--what am I doing here?

#### slateraptor

• Frequent Contributor
• Posts: 833
• Country:
##### Re: high voltage DC-DC converters
« Reply #12 on: January 05, 2012, 02:53:49 am »
The voltage regulates itself. Simple as that.

You are aware that, unless explicitly designed with additional peripheral hardware, all batteries are unregulated power sources, right? By your argument, the simplest form of active regulation (aka buffer) serves no purpose.

In that example battery linked above, the impedance is listed as 3 milliohms.

That specification is given under controlled laboratory conditions and is pretty much useless for all practical purposes. Like I said earlier, internal resistance is a dynamic parameter and a function of multiple variables to include load, temperature, internal design, age of device, etc. If we assume that internal resistance remains constant or even doubles, then sure, it's relatively insignificant. But it won't. In fact, a 100-fold or 1000-fold increase in internal resistance isn't uncommon. And just when you think you're done, there's feedback consequences in the form of internal heat dissipation which will exacerbate the situation.

Also, since the original requirement didn't specify any tolerance or allowable range for the 300 V output it is impossible to say whether any actual range of output voltages is or is not satisfactory.

The comment "3KW nominal power output" suggests some form of regulation. Admittedly, the point is moot.

This is not so. The voltage drop as a percentage of output voltage will be the same with one battery as with 25 batteries. Hence 30 mV in 12 V will become 750 mV in 300 V. Just as insignificant. As I said, when you put batteries in series then series resistance, whether internal or external, is a constant factor.

Once again, a battery's internal resistance is not a constant parameter and will increase under various operating conditions and over the lifespan of the device far greater than you might think.

It makes perfect sense in terms of the requirements. If you reduce the number of batteries you will need a DC-DC voltage converter of some kind to get the 300 V output, and this will increase the current draw on the remaining lower voltage battery supply compared to the 25 batteries in series for equal power outputs. Like for like, fewer batteries demands more current from each battery.

Naturally, by argument of conservation of energy. However 1) your sol'n does not include a transformer or any other form of DC-DC converter, and 2) the comment made by Psi was originally made in the context of batteries in parallel, of which your sol'n does not include.

As an aside, I could have sworn Dave did a blog some time ago on the topic of batteries.

#### IanB

• Super Contributor
• Posts: 9489
• Country:
##### Re: high voltage DC-DC converters
« Reply #13 on: January 05, 2012, 03:25:09 am »
You are aware that, unless explicitly designed with additional peripheral hardware, all batteries are unregulated power sources, right? By your argument, the simplest form of active regulation (aka buffer) serves no purpose.

No, regulation does not have to be electronic to be effective. Batteries regulate their output voltage by physical and chemical means. A lead acid battery may be considered an adequately regulated DC power supply for most practical purposes without any electronic additions. You are yourself aware that electronic regulation is never perfect, and the output voltage goes up and down with load, right?

Quote
That specification is given under controlled laboratory conditions and is pretty much useless for all practical purposes. Like I said earlier, internal resistance is a dynamic parameter and a function of multiple variables to include load, temperature, internal design, age of device, etc. If we assume that internal resistance remains constant or even doubles, then sure, it's relatively insignificant. But it won't. In fact, a 100-fold or 1000-fold increase in internal resistance isn't uncommon. And just when you think you're done, there's feedback consequences in the form of internal heat dissipation which will exacerbate the situation.

I think you are getting a bit carried away with theory divorced from practicality here. The fact is that if the internal resistance of a battery went up anything like 100-fold it would be considered dead as a dodo and due for replacement. Also, increasing temperature in a voltaic cell will always decrease the internal resistance and improve performance. Batteries like warmth (though they may not appreciate excessively high temperatures).

Quote
Once again, a battery's internal resistance is not a constant parameter and will increase under various operating conditions and over the lifespan of the device far greater than you might think.

You misunderstand. Series resistance is constant in its relative effect when varying the number of batteries in a series arrangement. If you put ten batteries in series the total system resistance will go up tenfold, but the voltage goes up tenfold as well. The two effects cancel out, requiring no extra compensation.

Quote
Naturally, by argument of conservation of energy. However 1) your sol'n does not include a transformer or any other form of DC-DC converter, and 2) the comment made by Psi was originally made in the context of batteries in parallel, of which your sol'n does not include.

The whole and fundamentally important point of my solution is that it does not include a transformer or any other form of DC-DC converter, while all other topologies do require such additional components. It is an engineering solution to a problem that simplifies the design by removing extra components. That reduces the bill of materials, improves the efficiency, reduces the potential failure modes and reduces the design costs.

I'm not an EE--what am I doing here?

#### Time

• Frequent Contributor
• Posts: 725
• Country:
##### Re: high voltage DC-DC converters
« Reply #14 on: January 05, 2012, 04:58:23 pm »
http://www.northstarbattery.com/1.0.1.0/215/NSB%20180FT%20Red%20Battery%20SES-542-46-01.pdf

Any idea what those cost?

The silvers might be more affordable actually.  I have seen a few 90 Ahr NSB90s for about 180 bucks a piece.  The reds are meant for a long float life while the blues are more designed for high discharge rate applications with lots of cycling (more abusive) - I think.  Good batteries if you have the budget.
« Last Edit: January 05, 2012, 05:02:51 pm by Time »
-Time

#### Zero999

• Super Contributor
• Posts: 12741
• Country:
• Hero999
##### Re: high voltage DC-DC converters
« Reply #15 on: January 05, 2012, 06:48:10 pm »
You are aware that, unless explicitly designed with additional peripheral hardware, all batteries are unregulated power sources, right? By your argument, the simplest form of active regulation (aka buffer) serves no purpose.

No, regulation does not have to be electronic to be effective. Batteries regulate their output voltage by physical and chemical means. A lead acid battery may be considered an adequately regulated DC power supply for most practical purposes without any electronic additions. You are yourself aware that electronic regulation is never perfect, and the output voltage goes up and down with load, right?
It depends on what level of regulation is required.

Off load and hot off the charge a sealed lead acid battery's voltage might 14.5V, but when it's fully loaded and discharged the voltage could be 10.5V - that doesn't sound like very good voltage regulation to me.

Still you're probably better off using many batteries in series (240V to 360V) and a step-up/down converter rather than a single 12V battery powering a step-up converter.
« Last Edit: January 06, 2012, 06:00:23 pm by Hero999 »

#### XTY

• Contributor
• Posts: 5
##### Re: high voltage DC-DC converters
« Reply #16 on: January 05, 2012, 10:16:25 pm »
You are aware that, unless explicitly designed with additional peripheral hardware, all batteries are unregulated power sources, right? By your argument, the simplest form of active regulation (aka buffer) serves no purpose.

No, regulation does not have to be electronic to be effective. Batteries regulate their output voltage by physical and chemical means. A lead acid battery may be considered an adequately regulated DC power supply for most practical purposes without any electronic additions. You are yourself aware that electronic regulation is never perfect, and the output voltage goes up and down with load, right?

Quote
That specification is given under controlled laboratory conditions and is pretty much useless for all practical purposes. Like I said earlier, internal resistance is a dynamic parameter and a function of multiple variables to include load, temperature, internal design, age of device, etc. If we assume that internal resistance remains constant or even doubles, then sure, it's relatively insignificant. But it won't. In fact, a 100-fold or 1000-fold increase in internal resistance isn't uncommon. And just when you think you're done, there's feedback consequences in the form of internal heat dissipation which will exacerbate the situation.

I think you are getting a bit carried away with theory divorced from practicality here. The fact is that if the internal resistance of a battery went up anything like 100-fold it would be considered dead as a dodo and due for replacement. Also, increasing temperature in a voltaic cell will always decrease the internal resistance and improve performance. Batteries like warmth (though they may not appreciate excessively high temperatures).

Quote
Once again, a battery's internal resistance is not a constant parameter and will increase under various operating conditions and over the lifespan of the device far greater than you might think.

You misunderstand. Series resistance is constant in its relative effect when varying the number of batteries in a series arrangement. If you put ten batteries in series the total system resistance will go up tenfold, but the voltage goes up tenfold as well. The two effects cancel out, requiring no extra compensation.

Quote
Naturally, by argument of conservation of energy. However 1) your sol'n does not include a transformer or any other form of DC-DC converter, and 2) the comment made by Psi was originally made in the context of batteries in parallel, of which your sol'n does not include.

The whole and fundamentally important point of my solution is that it does not include a transformer or any other form of DC-DC converter, while all other topologies do require such additional components. It is an engineering solution to a problem that simplifies the design by removing extra components. That reduces the bill of materials, improves the efficiency, reduces the potential failure modes and reduces the design costs.

Thank you all for you for some light on this subject, from what i've understood from pro's and cons it's actually better to use say 4 batteries in series and have more groups of these in parallel...or some of this type of wiring...it would sure be more long lasting and more reliable to have them in groups series-parallel...and it does not have to be regulated, it must be able to output 3KW for 70-75% of it's charge...as for voltage i'd say that it could be between 290 and 325dc it's a good idea to put them in series, but it would more difficult to be reliable, then more links in the chain the more problems could arise...especially if say just one of them has a rise in it's internal resistance...then the whole chain practically fails...

#### IanB

• Super Contributor
• Posts: 9489
• Country:
##### Re: high voltage DC-DC converters
« Reply #17 on: January 05, 2012, 10:24:13 pm »
it must be able to output 3KW for 70-75% of it's charge

One thing to be aware of is that lead acid batteries are really not good at charge cycling. If you discharge them, charge them, discharge them then sooner or later they will die, probably sooner. Some designs of battery can withstand this treatment better than others, but none of them are ideally suited for it. That's why when you buy those cheap halogen spotlights with a lead acid battery inside them you always end up with a failed battery.

Lead acid batteries are best used as a floating power reserve where they are kept close to full charge. If they must be used to supply power, as in an off-grid solar power solution, then it is best to use batteries specifically rated for this duty and not to take them below 50% charge even then.
I'm not an EE--what am I doing here?

#### Zero999

• Super Contributor
• Posts: 12741
• Country:
• Hero999
##### Re: high voltage DC-DC converters
« Reply #18 on: January 06, 2012, 06:10:37 pm »
it's a good idea to put them in series, but it would more difficult to be reliable, then more links in the chain the more problems could arise...especially if say just one of them has a rise in it's internal resistance...then the whole chain practically fails...
You'll have the same problem with an SMPS running off 12V: in this case the SMPS itself is probably most likely to be the weakest link in the chain.

The battteries should be pretty reliable, especially if they're charged in parallel so their voltages can equalise.

#### XTY

• Contributor
• Posts: 5
##### Re: high voltage DC-DC converters
« Reply #19 on: January 20, 2012, 10:26:46 pm »
it's a good idea to put them in series, but it would more difficult to be reliable, then more links in the chain the more problems could arise...especially if say just one of them has a rise in it's internal resistance...then the whole chain practically fails...
You'll have the same problem with an SMPS running off 12V: in this case the SMPS itself is probably most likely to be the weakest link in the chain.

The battteries should be pretty reliable, especially if they're charged in parallel so their voltages can equalise.
Since the feedback of all you wonderful members of this forum i've decided to use something like 48V for my inverter design....so i'll try to keep my inverter performance to a maximum by increasing the input voltage and decreasing the current....thank you....and i don't realy need voltage regulation since i.m trying to power SMPS directly without rectifier bridge that often goes from 150 to 350DC, now i'm thinking on the best way to build the inverter.....that would be stable enought...

#### sonicj

• Frequent Contributor
• Posts: 752
• Country:
• updata successed!
##### Re: high voltage DC-DC converters
« Reply #20 on: January 21, 2012, 12:23:55 am »
xWh of lead acid storage is going to consume approximately the same volume & weight regardless of voltage. if the target is 3kW @ 300VDC and the source is lead acid batteries, a series string at 300VDC nominal is the most reasonable solution in my opinion.

boosting anything less than the target voltage will add unnecessary complexity, losses in the conversion, cost of the interconnects, stress on the batteries, etc.... with a nominal 300VDC input, the 3kW output can be regulated (if needed) with a relatively simple pwm design.

telecom backup systems are a good example... the equipment operates at 48VDC & the backup batteries are wired in series to produce 48VDC. a single humongous 12V battery or wiring a bunch of 12V batteries in parallel and then boosting to 48VDC would be silly. my 2ยข fwiw.
-sj

#### damo

• Contributor
• Posts: 48
• Country:
##### Re: high voltage DC-DC converters
« Reply #21 on: January 23, 2012, 06:42:21 am »
I can confidently say this is the most efficient circuit for the job:

This is how they do it in electric cars and they draw far more than 10A

#### Zero999

• Super Contributor
• Posts: 12741
• Country:
• Hero999
##### Re: high voltage DC-DC converters
« Reply #22 on: January 27, 2012, 05:48:26 pm »
This is how they do it in electric cars and they draw far more than 10A
The trouble is come small electric vehicles such as golf carts need to use lower voltages (typically 48V) for safety reasons.

#### Psi

• Super Contributor
• Posts: 7132
• Country:
##### Re: high voltage DC-DC converters
« Reply #23 on: January 28, 2012, 12:29:36 am »
Home built electric cars are normally done with multiple 12v batteries in series to get between 72 and 144V DC

This is then feed into a motor controller which PWMs the 72-144V into the electric motor for speed control.
The design of these motor controllers is quite interesting. Its normally done with a large bank of fets or igbts in parallel.
But due to the massive currents involved they have to be connected in parallel with solid copper bars to stop one fet getting more current than others.
Greek letter 'Psi' (not Pounds per Square Inch)

Smf