0
\$\begingroup\$

I'm working on a project where I am planning to use two 5V USB Power Banks because one cannot supply sufficient current. These Power Banks are 20,000mAh each and have built in circuits to regulate their voltage down to 5V and are intended for recharging devices such as cell phones via USB. I don't know what these circuits look like however, but my multi-meter tells me they are (when fully charged) ~5.05V +/- 0.05V .

It is a matrix of 396 RGB LED's (the WS2812) controlled by a micro controller. To side step the issue of one being a slightly different voltage than the other; If I divide the display into two equal parts, keep ground common to both halves, but not the +5V, will that effectively block one pack from charging the other since they will be isolated from each other?

My main concern here is if the micro controller (Arduino) will be able to effectively communicate with all the LEDs which are using a proprietary daisy chain communication bus as far as I can tell. In short, the first LED passes instructions to the next and follows a precise timing diagram to know when to latch the data in it's own buffer. All this allows all the LEDs to be controlled with just 1 signal wire.

My power banks each have a 2.1A and a 1A USB port. Is it safe to tie those together to create a single 3A supply to one side?

More numbers: 396 Tri-Color LED's can draw a maximum of 60mA*396=23.76A!!! For what should be obvious reasons, I'd like to add that I will not be using all of the LEDs at the same time and will be trying to minimize the current draw by not running them at full brightness (controlled by software). Only one colour at half brightness (10mA)*396 draws a much more reasonable peak current of 3.87A. Since it's all digital, I plan to carefully calculate the current draw of my animations to keep them under the theoretical maximum draw of 6A that my two battery packs can source.

Thanks! __ They are Aukey brand power banks which I got off of Amazon. They contain Li-Po batteries, a charging circuit with all the nice protection stuff built in and what I believe is a switching power supply circuit. The reason I believe it is a switching supply and not a linear one is because when I have it hooked up to a number of LEDs flashing in unison I can hear high pitched squeals that shift up and down in frequency as it tries to reach 5V again. I've found that the smaller 8000mAh Aukey power bank I'm using with a test load of 75 LEDs has trouble maintaining 5V under load. I've tried adding some large and small capacitors to reduce the fluctuations seen by the source, but generally after connecting the battery the voltage falls from 5V to 3.4V where the micro controller becomes unstable and crashes. According to the marketing, the Power Bank should be able to source 2.1A... So I'm asking myself if somehow the changing nature of the load degraded one or more of the voltage regulating components in the power bank?

Hmmmm, perhaps this is a useful clue: (From the Aukey.com website) "AIPower technology utilizes the state of the art Texas Instruments TPS2513 evaluation module. Incorporating class-leading technology, the TPS2513 is a Dual Channel, Auto Detect USB Charging Controller Evaluation Module that drives AIPower to deliver the charging speeds you deserve."

Googling the TPS2513 I found that in the datasheet it is typically used in combination with the TPS2561A. From the Texas Instruments TPS2561A webpage "Accurate 2.1A Min / 2.5A Max Setting (Including Resistor)". What are the odds that this is where the 2.1A current limit comes from? The TPS2513 is also putting voltage on the data lines for some reason. It seems this has to do with compliance to battery charging specifications and I'm not sure if it has a detrimental effect if I'm using a usb to barrel plug cable where the two data lines are probably open circuits.

tl:dr The TPS2561A is a constant current supply with fast cut off in over current situations. It's not expecting a variable current draw and the TPS2513 is trying to figure out what on earth is going on. Fortunately, it also supports shorting the data lines so a continuity test of my cable should show if I need to take action there. It might be too late if I've damaged the TPS2561A by attempting to draw excessive current.

I'm posting this here because I'm not sure my analysis is correct. If anybody can point out where I've gone wrong, or how I could make these power banks work, that'd be greatly appreciated. Or maybe, I can at least help somebody else avoid a similar mistake.

Cheers!

\$\endgroup\$
  • \$\begingroup\$ Splitting the display as you suggest should be OK. \$\endgroup\$ – Brian Drummond Oct 1 '15 at 11:33
2
\$\begingroup\$

Opening up the 8000mAh power bank I found out it is a standard Li-Po battery which is 4.35V when fully charged. They're using a step up buck converter to boost the voltage up to 5V. Some thing, some where in their design has failed, I'm not going to waste time figuring out what I broke. Considering how cheap it was, perhaps they got burned with a fake part that can't meet the specs?

I'm left with two options, 1st, I can buy RC batteries that are 7.2V but can source large currents and use a quality DC-DC step down converter OR a linear regulator with a beefy heat sink.

2nd I can break open my nice new and unbroken power banks and harvest the batteries. This is probably a terrible idea. The batteries are soft prismatic cell types. If I bend or puncture them I risk setting myself on fire. I'd call that a lose-lose.

\$\endgroup\$
0
\$\begingroup\$

You can connect them together if you add a diode to each of their outputs. Without knowing how the power banks regulation circuit looks like no more hints can be given. However from the description of your project is sounds like you are better off combining the batteries alone and then feed them to a single voltage regulator that has enough current capability.

\$\endgroup\$
0
\$\begingroup\$

Running the led display with two separate power banks (and no diodes) may be better, since then you won't have any loss on the non-existing diode.

  • conect GND of all battery packs to the GND of Arduino
  • from the Arduino led data out, put a 1kohm resistor in series, and drive the first bunch of leds. The power of these leds go to the first power bank.
  • from the output of the first bunch of leds, have another 1kohm resistor to the next bunch of leds which is powered by the second power bank.

The problem happens if the two power banks are draining out differently. For example, if the first power bank runs out, then it can happen that Arduino drives the leds which have no power, and therefore drive the data in out of the specified range. The series resistors help on addressing this, and let the leds survive even if you disconnect the power banks (but still drive the leds with Arduino).

Alternatively:

  • Test those leds! I am sure no one wants to run them at 3x20mA, they're a way too bright.
  • Use different power supply solution, e.g. a 12V SLA battery and a step-down regulator which can output 5V 5A. You can find lots of those in robotic shops readily available.
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.