High current DC power converter

Thread Starter

Pete 001

Joined Jul 14, 2021
6
First time post - hello all.

I'm looking for a power supply solution that I assume (?) already exists - I don't think I'm trying to re-invent the wheel but can't find a product? I'm quite capable of building electronics projects but I'm somewhat limited on the design capability. This problem seems frustratingly simple but has left me without a resolution.

I have an instrument - normally powered by a 12V car battery.
It outputs short duration pulses of 1A or 8A (power on for max 20 ms and off for up to 50 ms - full cycle 2 to 8 minutes - max 20 times in a day)
In order to boost the output current (16A or 20A (this is the max)) we add another 12V and a 6V battery, thus providing 24V or 30V in total.

I'm looking to take one 12V car battery into the field but have some kind of DC/DC converter(s) that I can hook up to the 12V battery and select the output for either 12V, 24V or 30V - be that 3 separate converters (in one neat box I can build) or output voltage selectable by say a rotary switch changing resistor values, etc.

The max output from the instrument is 20A at 30V, which is 600W. So I guess the input current to supply that capacity from one 12V battery, in the worst case scenario, would be 50A plus a bit for losses due to inefficiencies.

Any thoughts? I have Googled a lot to no avail. The units I've found have too high input voltage, not enough power capability, wrong output voltages, etc.

Many thanks
Pete
 

Papabravo

Joined Feb 24, 2006
17,310
What you have run up against is the immutable law of DC-DC power conversion: the power out of any DC-DC conversion scheme will ALWAYS be less than the power in; sometimes it will be much less. A power supply that deals with high power levels will always be more difficult to find or design. My question is how much of a budget do you have for a solution?
 

Thread Starter

Pete 001

Joined Jul 14, 2021
6
What you have run up against is the immutable law of DC-DC power conversion: the power out of any DC-DC conversion scheme will ALWAYS be less than the power in; sometimes it will be much less. A power supply that deals with high power levels will always be more difficult to find or design. My question is how much of a budget do you have for a solution?
Thank you for the reply but I'm not sure what you mean?
The 600W required (plus a bit) is significantly less than a car's starter motor power requirement?
Would that not mean that the car battery should be more than capable of providing the power needed (given the short duration of the operation)?

Am I missing something?

I did find a boost/buck converter with the following spec:
  • Input voltage: 9V to 60V (80V short time)
  • Output voltage: adjustable from 0V to 60V
  • Input and output current: 50Amax; adjustable current limitation/regulation
  • Efficiency: 96%

But the Input current is right on the cusp - no margin for safety.
Knowing my luck it would go pop!

I guess another approach would be to build a Li-ion battery pack, which would have a larger capacity than the lead acid and a significant weight saving and could incorporate a 12V, 24V, and 30V switchable output. But Li-ion isn't so aircraft friendly and you can easily source car batteries wherever the work takes you.
 

AnalogKid

Joined Aug 1, 2013
9,498
Your timing requirements are not at all clear. 20 ms on and 50 ms off is 70 ms. Over 8 minutes, That's 6800 cycles, not 20. Please post a worst-case timing diagram so we can see the total energy required. It might be that you don't need a 20 A converter, but just a 2 A converter with a large output capacitor.

Separate from that, there are many low-cost boost converter modules available on ebay. For more money, there are several high-density DC/DC converter companies, such as SynQor. XP Power and other more traditional power supply companies also have DC/DC bricks and assemblies that can be paralleled for higher output current. For this project, I would stay away from Vicor unless you have some strange physical environment requirements.

Another reading of your timing indicates a max total on time of less than 160 minutes per day, approx 11% of 24 hours. Again, this indicates that a full-blown power supply capable of 600 W continuous duty might not be necessary.

You state the max off time during a cycle is 50 ms. What are the max and min on and off times?

ak
 
Last edited:

Papabravo

Joined Feb 24, 2006
17,310
Thank you for the reply but I'm not sure what you mean?
The 600W required (plus a bit) is significantly less than a car's starter motor power requirement?
...
Cars do not have DC-DC converters in them. I mean that 600W is a power level where managing voltage and current levels is a difficult design problem. You can't just slap components together on a breadboard and expect to make a working DC-DC converter. Also the greater the boost ratio the more stresses the components will have to endure. The boost ratio for going from 12V to 30V is 2.5. For back of the envelope calculations we use 80% as a realistic efficiency. for 600 Watts output power, that would require 750 watts of input power, which at 12V would be 62.5 Amperes (continuous). Designing an inductor that can handle that current without saturating the core is not a trivial process. Designing a PCB for those power levels is quite far from trivial.

Your inability to find a COTS (Commercial Off The Shelf) unit may be due to several factors, but should be a clue that this might be a low volume specialty item. I'm not saying it doesn't exist, but I am saying that these units are not going to be available in high volume at low cost.
 
Last edited:

Thread Starter

Pete 001

Joined Jul 14, 2021
6
Your timing requirements are........

Yes, apologies - I could (should) have been clearer.
The off time is misleading and unnecessary for descriptive purposes.

There are many timing scenarios unfortunately:
Short current pulse for a designated on period (usually between 1ms and 10ms)
Switch off current......wait (that's the off time and doesn't add anything to this)
Start recording.......again at different frequencies and for different periods (no pulse during this phase)
Switch pulse back on and repeat

Typically, 1000 to 2000 pulses in a given cycle.
You can repeat the cycle normally between 3 and 5 times per session
A session takes between 360 and 810 seconds


20 per day was the fact the all the above (a session) can happen up to 20 times in a working day - where after you can re-charge the battery.

The worst case scenario that I currently looking at is:
On for 10ms, 2000 pulses over 360 second period

I will have a look at the high-density DC/DC converters you mentioned as well - thank you.
 

AnalogKid

Joined Aug 1, 2013
9,498
Example. Start with 12 V. With a non-isolated boost converter, boost it up to 105 V. This charges a 5700 uF capacitor. The cap drives a 20 A linear regulator with a 30 V output. With a fully charged cap, it can drive a 20 A output for 20 ms while it discharges from 105 V to 35 V (leaving 5 V of headroom for the regulator - with zero energy from the 12 V source during the discharge. So the question is how long does the boost converter have to recharge the cap before the next output pulse.


IOW, what is the worst case duty cycle of the output waveform. In this case, "worse case" is the max on time and min off time. This sets the lower limit of how much power the DC/DC converter has to move. For example, if the duty cycle is 29% (20 ms on / 50 ms off), then that 600 W requirement drops to 172 W (because the converter still is running during the discharge phase). Plus efficiencies, design margins, yaddi yaddi

ak
 

AnalogKid

Joined Aug 1, 2013
9,498
The off time is misleading and unnecessary for descriptive purposes.
Actually, the off time is critical.
The worst case scenario that I currently looking at is:
On for 10ms, 2000 pulses over 360 second period
Working with your worst case, 10 ms x 2000 is 20 s. At 600 W peak power, that is only 6 w-s per pulse, and 12,000 w-s of total energy

Evenly distributed (???) over 360 s, that is a duty cycle of 5.6% and an average power of 33 W per pulse, a way more reasonable number.

BUT, this works only if the pulses are evenly distributed. So again, on a cycle-by-cycle basis, what are the worst case on and off times? IOW, what is the least possible time the circuit has to recharge itself before the next output pulse?

A side question - how tightly regulated or shaped do the output pulses have to be?

ak
 
Last edited:

Thread Starter

Pete 001

Joined Jul 14, 2021
6
Sorry Papabravo, I seem to have upset you???

"I mean that 600W is a power level where managing voltage and current levels is a difficult design problem"
Yes, I agree with you. Your initial comment lead me to believe that you thought I was creating power from thin air, which is why I commented and asked for clarification

"You can't just slap components together on a breadboard and expect to make a working DC-DC converter"
Ever so slightly condescending and really doesn't help.

"Your inability to find a COTS......"
Could be many factors yes, which is why i was asking for assistance - and having found one that would almost do the job there presumably are others out there.

"Cars do not have DC-DC converters in them."
I think you'll find they do!

" not going to be available in high volume at low cost "
I haven't mentioned budget but you have twice??
There is an obvious point at which dragging 2.5 x 12V batteries up and down dale is outweighed by the purchase of something that will do the job - what is that budget.......honestly, I don't know but I suspect if I find a solution I'll either buy one immediately or continue to lug batteries
 

Thread Starter

Pete 001

Joined Jul 14, 2021
6
Actually, the off time is critical.

Worst sequence is thus:
ON 1ms 1A
OFF 1.22ms
Record 20ms
ON 1ms 1A
OFF 1.22ms
Record 20ms
ON 10ms 20A
OFF 10ms
Record 20ms
ON 10ms 20A
OFF 10ms
Record 20ms
Repeat for designated time window

The output pulses are not important in so far as the instrument outputs the max (20A) current it can for the voltage connected to it. In none ON mode the instrument still needs power but the draw is very minimal but it needs to be the designated voltage - ie 12V, 24V or 30V
 

LowQCab

Joined Nov 6, 2012
1,350
20-Amps X 30-Volts = 600-Watts, at a ~50% Duty-Cycle = 300-average-Watts for
8-Minutes duration, every 60 minutes.

~20 to ~25-Amps is a job for a Car-Battery(ies).
You have 50-Minutes every Hour to recharge the Battery(ies) at around ~5 to ~10-Amps.

Will the device that You won't describe for us withstand say ~45-Volts ?,
if so, You need 3-Car-Batteries and a custom made Battery-Charger.
( I recommend Optima-Gel-Cell Batteries, about ~$250.oo each, and worth twice as much )

The Charger will need to be able to provide ~10-Amps at 43.5-Volts, or ~450-Watts.
That's a stout Battery-Charger, and it will probably weight about ~60lbs.
.
.
.
 

Ian0

Joined Aug 7, 2020
3,773
It outputs short duration pulses of 1A or 8A (power on for max 20 ms and off for up to 50 ms - full cycle 2 to 8 minutes - max 20 times in a day)
In order to boost the output current (16A or 20A (this is the max)) we add another 12V and a 6V battery, thus providing 24V or 30V in total.
Could you please clarify?
So, about once an hour, this device runs for between 2 and 8 minutes, during which time it outputs 1A or 8A at a duty cycle of between 28% and 100% ? There's a factor of four difference in the duty cycle
There's a big difference in energy requirement depending on whether it is 1A or 8A! (a factor of 8)
Could it possibly run for 8 minutes every time it starts? Or is it randomly spread 2 minutes to 8 minutes? (there's another factor of 4 here)
This is a very difficult target to hit, as the requirements could possibly vary by a factor of 128.
Then you want to boost the output current to 20A. That's a further variation by a factor of 2.5, now there's a variation of x320 between lowest and highest possible requirements.
Until you have accurately quantified your requirements, designing it will be tricky.
 
Last edited:

Papabravo

Joined Feb 24, 2006
17,310
Not between the battery and the starter motor.

Bob
Sorry Papabravo, I seem to have upset you???

"I mean that 600W is a power level where managing voltage and current levels is a difficult design problem"
Yes, I agree with you. Your initial comment lead me to believe that you thought I was creating power from thin air, which is why I commented and asked for clarification

"You can't just slap components together on a breadboard and expect to make a working DC-DC converter"
Ever so slightly condescending and really doesn't help.

"Your inability to find a COTS......"
Could be many factors yes, which is why i was asking for assistance - and having found one that would almost do the job there presumably are others out there.

"Cars do not have DC-DC converters in them."
I think you'll find they do!

" not going to be available in high volume at low cost "
I haven't mentioned budget but you have twice??
There is an obvious point at which dragging 2.5 x 12V batteries up and down dale is outweighed by the purchase of something that will do the job - what is that budget.......honestly, I don't know but I suspect if I find a solution I'll either buy one immediately or continue to lug batteries
For the record I'm not upset, and I'm perfectly content to let you wander around in the wilderness to your heart's content.
 
Top