Parallel voltage regulators on Propeller Proto Boards

Duane Degn · 2011-05-30 20:46

I'm using two Propeller Proto Boards stacked on top of each other for one of my robot projects. I wanted to use one battery pack to power both boards. I had read in several voltage regulator data sheets that they were okay to use in parallel (I don't remember which voltage regulars these data sheets were for).

I connect Vin, 5V, 3.3V and ground on the first board to the corresponding voltages on the second board. The regulators on the second board seem to get hotter than the first board.

Is this a mistake to connect the regulators in parallel? I suppose I could just have the boards share Vin and ground if I need to.

I'm attaching the data sheet for the LM1086 regulators used on the PPBs. I couldn't find anything about using them in parallel.

Thanks,

Duane

LM1086IS-3.3-ND_LM1086.pdf

Duane C. Johnson · 2011-05-30 20:59

In general it is OK to run linear regulators in parallel.

The reason one is running a bit hotter than the other is it's output is set to a little higher voltage, probably due to tolerances.

This is OK as it will eventually go into over temperature fold back, which is normal and no damage will occur. The other regulator will take up the slack.

Duane

LoopyByteloose · 2011-05-30 21:06

From what I've read, linear regulators in parallel tend to be problematic if they are providing power to one item down line that exceeds the current abilities of one regulator. Not only can one run hotter, they can begin to oscillate destructively. If you need more power, you should use an added power transistor.

There are some regulators that have enough protection so that they will shut down before permanent damage occurs. These are not really functioning properly in parallel, though they may avoid damage.

There is nothing to be gained by having them parallel if you are only getting power equivalent to one regulator.

But there are ways to use multiple Proto-boards in parallel that are not destructive (safe). Just allocate +5 and +3.3 to separate devices according to available current demands. The only common connections should be Vin and one Ground point.

Beau Schwabe · 2011-05-30 21:36

Duane C. Johnson,

"In general it is OK to run linear regulators in parallel." - If the output of the regulator was MOSFET in nature I would tend to agree with you, but since the output of most linear regulators use a bipolar transistor configuration, I'm going to have to politely disagree with your statement for similar reasons that it's not a good idea to parallel LEDs.

Duane Degn · 2011-05-31 11:25

Thanks everyone for your advice and help.

It looks like I'll need a different solution for my power needs. I had doubled up for a couple of reasons. One was that I had some high power needs and I thought the two regulators working together would take care of that.

I will probably remove the linear regulars completely and use a couple (one for each voltage) of switching regulators for my power needs. This should also help with battery life.

Thanks again.

Beau Schwabe · 2011-05-31 21:30

If you follow the application notes, there are ways to obtain more driving current by using an external transistor, and this would be the correct way to deliver more current.

Duane Degn · 2011-05-31 21:49

Beau, I've looked here for an application note about using external transistors.

I'll keep looking. Maybe it's somewhere in Parallax Semiconductor's site. I couldn't find the SD card application note in Propeller download section either.

Duane Degn · 2011-05-31 21:55

I found a bunch of appnotes at Parallax Semiconductor. I'll look through them.

Thanks,

Duane

Leon · 2011-05-31 22:10

The regulator data sheets should have details of how to add an external transistor. If not, it's in the 7805 data sheet.

Beau Schwabe · 2011-05-31 22:37

Duane Degn,

I should have been more specific... the app notes would be particular to the voltage regulator that is used and not have any direct ties to Parallax. Leon has the right idea. The method used for the 7805 regulator should work for the 3.3V and 5V regulator used for the Propeller.

Duane Degn · 2011-05-31 23:02

Beau,

Thanks for the clarification.

I did find a lot of great appnotes at Parallax Semiconductors. Is there a reason these appnotes aren't in the Propeller download section?

Thanks to you too Leon. I'll take another look at the data sheets.

I'll probably still go with switching regulators. The first of my two boards controls a bunch of 595 shift registers that are physically located on the second board and the second board sends signals to motor controllers that are on both the first and second boards. I'm afraid I'll get myself in too much trouble trying to keep each board's 5V and 3.3V line separate.

As I understand, the switching regulators are a lot more efficient. Besides, I'm pretty sure I already have some at the correct voltages and power ratings.

I do want to make sure and learn about the transistor solution. I learned a lot of electronics theory (electron theory is probably a better term for what I learned) when I minored in physics but I'm still unclear about all the uses for PNP, NPN, MOSFET and the rest. (I think I remember something about Eli the Iceman?)

Thanks again for everyone's help.

Duane

Leon · 2011-05-31 23:25

The Nat Semi Simple Switchers are easy to use, and they have nice on-line design software that selects a suitable device for you.

kwinn · 2011-06-01 00:20

You should be able to connect the battery to the input on both boards without problems. Connecting the regulator outputs on the other hand will be problematic due to the small differences in output voltage of the regulators.

Take a look at page 24 of the attached data sheet for details on adding a parallel transistor.

RobotWorkshop · 2011-06-01 11:40

Sharing the Vin and ground's is something I've done on may of my robots when I have multiple boards. I just use the built-in 5V and 3.3V regulators to power the Propeller and some of the logic. If I need more power then I add an extra regulator to power those. Just because you need more power doesn't mean it all has to come from the same source on the board. The outputs of the regulators are separate and just power different things. On my latest two robots I do this and the second regulator handles the power for the servo's, PING, LED's ,etc. The load seems to be spread out well. You could easily use a small switching PS as well.

Robert

Duane Degn · 2011-06-01 13:40

One concern I have about keeping the 5V and 3.3V lines separate is (as I earlier explained) the signal lines to devices on each board are not necessarily from the same board the device is powered by.

I'm starting to think this shouldn't be too much a concern. If a 595 shift register received 3.2V as Vdd and received a signal voltage of 3.3V, would this cause a problem? I doubt the regulators on the two boards would have a 0.1V difference. I'd think the voltages should be close enough to have devices powered by one board and receiving there logic signal from the other. Do the rest of you agree this shouldn't be a problem?

The 595 data sheet lists the max Vi (input voltage) as Vcc. It also lists IIK (input diode current) as |(conditions) Vi < -0.5V to Vi > Vcc + 0.5V | MAX +/-20mA
I'm I right to interpret this as I have to keep the input current less than 20mA if the input voltage is more than 0.5V greater than Vcc? So if the input voltage were 0.1V higher than Vcc, I don't need to worry about it?
(I just checked. The data sheet quoted above was for a Philips chip. I'm pretty sure the chips I'm using are made by Fairchild. I couldn't find a similar "Vi > Vcc + 0.5V" statement it the Fairchild datasheet.) I'll post the data sheets if anyone whats to look at them. I just don't expect people to dig through the data sheets for me.

I know I could just take Vcc from the same board as the board providing the signal but that would increase the rat's nest of wires I already have.

I'm hoping some of you will have enough experience with this type of stuff to tell me "yes, don't worry about a possible small voltage difference" or "no, you're asking for trouble. Build a bigger rat's nest."

Thanks to everyone for taking time to answer my questions.

Duane

Toby Seckshund · 2011-06-01 13:51

When I were lad.....

It was always good to have one common ground point, so that the heavy load curents couldn't flow up into sensitive and fragile bits. On my Nascom 2 (back in the 1980s) all the TTL dragged over 2 Amps and the added floppy pulsed heavilly, I had a single bolt that had all the grounds comming from it and then all was good.

( I left that computer behind, when I divorced, and I miss it dreadfully)

RobotWorkshop · 2011-06-01 14:04

I'm using a set of four 74HC595 chips for LED lighting and other outputs. I added an extra 5V regular (powered off Vin) which shares the ground connection with the rest of the board. I'm running the 595's and LED's all from the secondary 5V regulator I installed. I added a 2.7K ohm resistor on the serial data out into the first 595 chip. For the Clock and Latch lines I found that the Propeller running at 3.3V could drive one but when I connected all four it wasn't reliable and had trouble driving them. To fix I added an 74HC14 chip (5V) and used four of the inverter gates. Two pairs (invert and invert again) which buffered those lines. Again I have a 2.7K ohm resistor to the single input of the 74HC14. With it driving all four 595 clock and latch lines it has been reliable so far.

Robert

Parallel voltage regulators on Propeller Proto Boards

Comments