Practical way of getting a stable voltage from a supercapacitor bank? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Practical way of getting a stable voltage from a supercapacitor bank?

Move Thread

LAN_403

Artlav

Wed Jul 02 2014, 09:11PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Supercapacitors discharge linearly.
If a bank of them have a 30V max voltage, then 90% of it's energy will be within 10V to 30V range.

So, to use that energy one should be able to accept a wide input voltage.
But most devices use a fixed input voltage - 5V, 12V, 24V, 48V, etc.

Three system design options here.

One - step down.
Your working voltage should be below the 10% mark.
As a rule of thumb, multiply it by 3 to get the bank's max voltage.
Easy to make - buck converters are among the most efficient ones.
Disadvantage - more caps in series needed.

Two - step up.
Your working voltage is above the max voltage of the pack.
Not so easy to make, since step-up converters are not as easy or efficient, and there will be large currents at the input.
Also, you can easily go over the voltage limits of cheap devices, needing more expensive solutions - i.e. powering an LED is easier from 12V than from 48V.

Three - convert to a voltage in the middle.
For 10V to 30V it's 24V, a voltage easy enough to work with.
Problem is - i have no idea what sort of converter can do this, at decent efficiency and 100W range of power.

So, here is the question - how do you design a converter that can take 10V to 30V, and give out 24V at about 100W, with decent efficiency?

Boost to something like 32V, then buck to 24V works, but efficiency is in 50% range.
I also tried looking at SEPIC.
It works, but i can't figure out how to get more than a couple watts out of it.

Any ideas?

Carbon_Rod

Thu Jul 03 2014, 12:29AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

You may be able to get a regulated output. However, the nonlinear rate of discharge may strain the super-cap as it suffers a rapid increase in current draw as regulator Vin drops below Vout. Excluding the increasing conversion losses may also pose a problem.

The lm5118 has a fairly decent conversion efficiency if the correct inductor size if chosen:

a $7 for a 15Watt buck boost module with current regulation for testing the design:

Patrick

Thu Jul 03 2014, 04:55AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I have a great deal of experience with super and ultra caps which use the Helmholtz effect, starting way back in the ancient year of 1997.

But for what purpose is this device needed?

Artlav

Thu Jul 03 2014, 02:54PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Carbon_Rod wrote ...

You may be able to get a regulated output. However, the nonlinear rate of discharge may strain the super-cap as it suffers a rapid increase in current draw as regulator Vin drops below Vout. Excluding the increasing conversion losses may also pose a problem.

The caps are rated for 200A continuous, which is 2kW at 10V. I only want 100-200W or so.
Not a problem.

Carbon_Rod wrote ...
The lm5118 has a fairly decent conversion efficiency if the correct inductor size if chosen

So, a buck-to-boost transition special purpose controller?
Is there a name for this?
"buck boost" gives me predominately inverting converters.

I'd love to find some chip or design that would allow parallel stacking - a converter that alone can deliver i.e. 1A at high efficiency, but that can be duplicated 10 times in parallel (without complex load balancing issues).

Patrick wrote ...
But for what purpose is this device needed?

Experimentation.
I have 12 of 3000F caps, 30WH total, and i want to figure out how to use them as energy storage, get to know the technology.

They are charged from a small solar panel, and should provide a stable voltage (24V?) to some LED lighting and an AC inverter.

Pieces needed:
Charger - straightforward. CC, CV from 32V, boosted from a panel.
Balancing and monitoring - trickier, but LiPo stack solutions exist and fit the task. MAX14920 + Atmega328P.
Output - ?

I have already tried case 1 - buck - successfully, for 8S supercap banks to charge a cell phone (5V).

The solar panel charges the caps to 18V with a basic MPPT charger and passive balancing with zenners, and they are discharged down to 6V with a buck converter that gives out 5V.
As simple as it gets, but not always practical and common.

Now i want to try case 3, which sounds to be the most common case - the target voltage being in the middle of the discharge range.

Carbon_Rod

Thu Jul 03 2014, 05:07PM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

Not sure why the lm5118 URL was lost, but it is a little wide-input-range regulator switcher chip:

IIRC, the current rating of 4A in the app notes seem below the control limit older designs noted.

Paralleling SMPS is an old technique:
1. Put large schottky isolation diodes on outputs (should be rated twice peak Vout)
2. load balancing by matching the exact Vout of each supply
3. control loop sharing across each sensor line

The buck boost chip will unlikely be very reliable because of the lm5118 error correction loop in cases 2 & 3. However, adjusted 2A current-regulation in the cheap LM2577+LM2596 (under $4.45 in bulk) module posted should automatically limit the asymmetry of the loading.

It may be less hassle if you can find a solid 10A output converter for under $30.
There are also numerous larger DC-DC modules available used in electric bikes that may meet your needs.

Dr. Dark Current

Sat Jul 05 2014, 11:21AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

It's always best to design the converter / capacitor bank so that the output voltage is either at all times higher than input voltage (and then use a boost converter) or vice versa (and use a buck converter).

If charging of the capacitor from the regulated voltage is also required, a buck-boost converter can be used (basically a half-bridge).

In most cases a simple inverter based around the UC384x IC will work. For a two-way converter (buck-boost) the TL494 can be used.

Electra

Sat Jul 05 2014, 10:10PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

What about a transformer based flyback converter ( it's is really two coupled inductors ), you can set the turns ratio to whatever you want then.
Discontinuous mode may be easer to design for the feedback loop.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob