SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Reply
Thread Tools
Old Jun 15, 2012, 02:15 AM
Promoting Model Aviation...
Murocflyer's Avatar
United States, CA, Tehachapi
Joined Nov 2005
24,921 Posts
Discussion
Calculating Servo Draw or How Can I Tell I Need a UBEC?

I'm trying to understand how a person can tell if he needs to use a UBEC or not.

I am building a plane that will run on a 3S LiPo and have 4 servos. 3 ea HS-81s and 1 ea HS-85BB.

I am sure there is a calculator on this somewhere, but being somewhat new to electrics (been a while since I have been an active e-powered flyer) I am unsure on how to figure this out.

Does anyone have a good method to solve this issue?

Thanks,

Frank
Murocflyer is offline Find More Posts by Murocflyer
Reply With Quote
Sign up now
to remove ads between posts
Old Jun 15, 2012, 02:19 AM
Stuart
srnet's Avatar
UK, Cardiff
Joined Dec 2008
3,223 Posts
Quote:
Originally Posted by Murocflyer View Post
Does anyone have a good method to solve this issue?
Get an Eagle tree logger with the servo current measuring add on and find out what the servo current really is.
srnet is offline Find More Posts by srnet
Reply With Quote
Old Jun 15, 2012, 07:31 AM
Dave the Rave
dmccormick001's Avatar
Joined Jun 2007
873 Posts
The current draw on those 4 servos is going to be fairly small, as those are all miniature servos designed for small, lightweight models.

What size ESC are you using, and doesn't it have an on-board BEC? Almost all ESCs have a BEC built into them, no need to add an external one. The main exception to that is in the case of some very large ESCs, if your ESC doesn't have an on-board BEC then of course you'd have to use an external UBEC (Universal BEC). But for a plane that's using the size and type of servos you listed, the BEC that's built into the ESC should be more than sufficient.
dmccormick001 is offline Find More Posts by dmccormick001
Reply With Quote
Old Jun 15, 2012, 09:59 AM
Promoting Model Aviation...
Murocflyer's Avatar
United States, CA, Tehachapi
Joined Nov 2005
24,921 Posts
I'll be using the CC T-Bird 36 http://www.castlecreations.com/produ...nderbirds.html

I did see this at Servo City http://www.servocity.com/html/hs-85b...hty_micro.html

"Current Drain (4.8V): 8mA/idle and 240mA no load operating"

What does "no load operating" mean?

Thanks,

Frank
Murocflyer is offline Find More Posts by Murocflyer
Reply With Quote
Old Jun 15, 2012, 10:25 AM
"Simplify, then add lightness"
Raleigh,NC
Joined Nov 2000
2,701 Posts
The current draw from servos is highly variable and depends on load and motion. Peak currents can be high, but average current which causes the BEC to overheat is usually the limiting factor. According to the spec sheets those servos draw around .25A when moving with no load, and less than 10 mA when idle. If all the servos were moving at the same time even with no load you could draw around 1 amp. For the linear BEC in the typical ESC, the difference between the battery voltage(12V) and the BEC voltage(5V) multiplied by the current gives you the wasted power which is converted into heat. In your case with all the servos moving with no load you would be wasting around 7 watts of heat in the BEC. The typical linear BEC in the typical ESC will handle somewhere around 1 to 3 watts of continuous power before its thermal protection shuts it down. If all of your servos were moving continuously, the BEC would overheat and shut down.

Typically in an airplane you are not constantly moving all the servos at the same time. With a stable flier you would probably be safe with the internal BEC on a typical ESC, but with an aerobatic plane doing a lot of servo movement, the BEC could overheat and shut down.

PS idle current is when the servo is sitting in one position with no mechanical pressure on the servo arm. No load operating current is when the servo is moving with no mechanical resistance on the servo arm.
jeffs555 is offline Find More Posts by jeffs555
Reply With Quote
Old Jun 15, 2012, 10:35 AM
Promoting Model Aviation...
Murocflyer's Avatar
United States, CA, Tehachapi
Joined Nov 2005
24,921 Posts
Very good. Thanks for the info. This will come in handy for future airplanes as well.

Frank
Murocflyer is offline Find More Posts by Murocflyer
Reply With Quote
Old Jun 16, 2012, 07:51 AM
Dave the Rave
dmccormick001's Avatar
Joined Jun 2007
873 Posts
Quote:
Originally Posted by jeffs555 View Post
For the linear BEC in the typical ESC, the difference between the battery voltage(12V) and the BEC voltage(5V) multiplied by the current gives you the wasted power which is converted into heat. In your case with all the servos moving with no load you would be wasting around 7 watts of heat in the BEC. The typical linear BEC in the typical ESC will handle somewhere around 1 to 3 watts of continuous power before its thermal protection shuts it down. If all of your servos were moving continuously, the BEC would overheat and shut down.
The only disagreement I have with what Jeff has posted is in what he has listed as a "typical" BEC. His explanations of servo current, etc. are all right on the money, but I think he has underestimated what most common ESCs, even the cheaper ones, are capable of handling in the way of servo current. I went to HobbyKing's website and picked one of the first ESCs I found there, a very common line they call "Red Brick" ESCs, and the BEC in that ESC is rated for 3 amps @ 5 VoltsDC. Using Jeff's formula that would be (12v - 5v) 7v * 3a = 21watts. And that ESC sells for only $8-9. The idea that the BECs in most of the smaller ESCs will only handle 1 to 3 watts of power is too low, I think. With a plane/motor/ESC combo like the one the OP is describing, any decent ESC should have a BEC built in that will handle those servos, even for 3D. I've built a truckload of planes of that size and type, and that's always been my experience, a UBEC is only necessary when you get into the much larger planes and sevos.

No disrespect intended for what Jeff has posted, he's way over my head most of the time in circuit theory, but I think I'm right about this one?
dmccormick001 is offline Find More Posts by dmccormick001
Reply With Quote
Old Jun 16, 2012, 10:51 AM
"Simplify, then add lightness"
Raleigh,NC
Joined Nov 2000
2,701 Posts
Dave,
While manufactures like to advertise a simple single number as a rating, sizing regulators like most things electronic is rarely simple. Most often the numbers you see advertised are the absolute best case numbers and in the real world can be achieved only under ideal circumstances. Also, current rating and power rating are two different things and you can't just multiply current rating by voltage rating to get power rating.

The power rating is determined by the maximum chip temperature which is usually between 125C to 175C. If the chip exceeds this temperature it could be destroyed so most regulators have a thermal shutdown which cuts the output off when temperatures get too high.

Power is converted into heat. The chip temperature is determined by how well you can get this heat away from the chip which is determined by the package the chip is mounted in and whatever heat sink(if any) is used. The typical D2PAK packaged regulator mounted by itself on a PC board in free air will raise the chip temperature at least 50DegC above ambient temperature for each watt of power. At 50DegC per watt, it only takes 3 watts to get it up to 150C(your 21 watts would have it over 1000DegC). That temperature is in addition to any other heat on the board like from the mosfets powering the motor.

There are so many factors that affect temperature that you really can't pick any one number. If mounted in the airstream with a giant heatsink added, the numbers could approach your 21 watts. If closed inside a foam fuselage, the numbers could be near one watt.

Jeff
jeffs555 is offline Find More Posts by jeffs555
Last edited by jeffs555; Jun 16, 2012 at 11:02 AM.
Reply With Quote
Old Jun 30, 2012, 07:25 PM
Got shenpa?
flieslikeabeagle's Avatar
Los Angeles
Joined May 2004
10,940 Posts
Quote:
Originally Posted by dmccormick001 View Post
...the BEC in that ESC is rated for 3 amps @ 5 VoltsDC. Using Jeff's formula that would be (12v - 5v) 7v * 3a = 21watts. And that ESC sells for only $8-9. The idea that the BECs in most of the smaller ESCs will only handle 1 to 3 watts of power is too low, I think.
Truthfully, I think you have just pointed out why it is a very bad idea to trust the published ratings of most (linear, not switching) BEC's.

Let me put it this way: you can get hot enough to melt solder with an iron drawing less power than the 21 watts you calculated! Here's an example: http://www.radioshack.com/product/in...ductId=2062728

Now consider that that 15 watt soldering iron has far, far more surface area exposed to cooling air than the tiny little BEC chip in an ESC does. That means the BEC chip will get much hotter than the soldering iron tip for the same power dissipation!

What I'm trying to say is this: Never mind 20+ watts, even just a watt or two is enough to generate alarmingly high temperatures in a tiny BEC chip. In most ESC's there is no real heatsink for the BEC to help it dissipate heat - all it has is a few square millimetres of copper (on the circuit board).

Semiconductors tend to get very unhappy if they get much hotter than boiling water (100 deg C or 212 deg F). And if you want them to be reliable, it's best to keep them much cooler than that - 60 deg C is about as hot as I want any semiconductor device to get.

If you're game, there is an easy test you can try: hook up some servos and an RC receiver to an ESC, but no motor (for safety). Turn on the Tx and power up the ESC and servos.

Now find the BEC chip on the ESC and put a fingertip on it. With your other hand, wiggle one of the transmitter sticks to get those servos moving. Be prepared to pull that finger off if things get too hot!

Chances are you'll find that BEC chip gets very toasty very quickly, and you may have to pull your finger off in a hurry to avoid pain.

An old rule-of-thumb I've found useful: if you have to yank your fingertip off a hot semiconductor device within less than 3 seconds to avoid pain, that surface is too hot (over 60 deg C).

And that was with no load on the servos, and no motor connected to heat up the rest of the ESC (that additional heat also has to be dissipated by the same tiny copper layers in the ESC circuit board). Imagine how much more stressed the BEC is with friction and aerodynamic loads fighting the servos?

Personally, coming from a long background of dealing with solid-state electronics, I've always felt the tiny linear BEC was one of the weakest links in our RC planes. It was acceptable in the days of 6-cell NiCd and 2S Lipo flight batteries, but as the battery voltages used in most electric models went up (3S lipo is so common now), the linear BEC became a disaster waiting to happen IMHO.

So when switching BEC's dropped to very affordable prices (some years ago they dropped to under $7 from a few sources), I simply started using them on every RC model I built.

And that is actually what I recommend. If your model uses a 3S or higher voltage lipo pack, just use a switching UBEC on it, and eliminate the weakest and one of the most deadly central failure points of any RC model. (A failed BEC usually means a destroyed model, and/or property damage or injury to people in harms way from the out-of-control model.)

-Flieslikeabeagle
flieslikeabeagle is offline Find More Posts by flieslikeabeagle
Reply With Quote
Old Jul 01, 2012, 12:08 AM
Stuart
srnet's Avatar
UK, Cardiff
Joined Dec 2008
3,223 Posts
Quote:
Originally Posted by flieslikeabeagle View Post
So when switching BEC's dropped to very affordable prices (some years ago they dropped to under $7 from a few sources), I simply started using them on every RC model I built.
Fine, but if you have a linear BEC in the ESC that is adequately heat sinked, and\or the particular model has a limited BEC current load, would not adding the switching BEC make the system more unreliable ?

A switching BEC is many components, a linear BEC normally only one, so assuming both are working within ratings which is going to be the more reliable ?

You also need to take account that adding the switching BEC adds in extra wires etc, so that too decreases reliability.
srnet is offline Find More Posts by srnet
Reply With Quote
Old Jul 01, 2012, 01:03 AM
Dave the Rave
dmccormick001's Avatar
Joined Jun 2007
873 Posts
Quote:
Originally Posted by flieslikeabeagle View Post
........... Let me put it this way: you can get hot enough to melt solder with an iron drawing less power than the 21 watts you calculated! Here's an example: http://www.radioshack.com/product/in...ductId=2062728
Not that I disagree with everything you posted, but this comparison is a bit far-fetched. You're comparing a 110 volt AC device that's designed to get hot, with a 12vDC device that isn't. It's like comparing a helicopter to a Bi-plane.
dmccormick001 is offline Find More Posts by dmccormick001
Reply With Quote
Old Jul 01, 2012, 03:38 AM
"MAYONNAISE"
Acetronics's Avatar
Le Treport, France
Joined Jun 2004
1,404 Posts
http://www.mnbigbirds.com/Servo%20To...0Caculator.htm : to get required torque
+
http://www.aeroglide.net/2010/10/ser...-octobre-2010/ : to convert torque to current AND verify servo is fast enough under load ... ( Humour ... )

may be ...

Alain
Acetronics is offline Find More Posts by Acetronics
Reply With Quote
Old Jul 01, 2012, 04:02 AM
Got shenpa?
flieslikeabeagle's Avatar
Los Angeles
Joined May 2004
10,940 Posts
Quote:
Originally Posted by srnet View Post
Fine, but if you have a linear BEC in the ESC that is adequately heat sinked, and\or the particular model has a limited BEC current load, would not adding the switching BEC make the system more unreliable ?

A switching BEC is many components, a linear BEC normally only one, so assuming both are working within ratings which is going to be the more reliable ?
Most of the switching BEC's I've seen are based on a single chip, just like linear ones. There are two or three external components in both cases - typically a couple of capacitors and resistors, and one inductor in the case of the switching BEC. Usually resistors and inductors and even capacitors are quite reliable - the really unreliable thing is a hot semiconductor. Semiconductors are wonderfully reliable when cold, but that changes in a hurry if they get too hot.

I don't think there's much evidence to support the hypothesis that a switching BEC is less reliable than a linear one - they are everywhere these days, in the chargers for your laptop, cellphone, camera, and other electronic devices, as well as in virtually all the wall-wart DC power supplies made in the recent past. And all of those are much higher power switchers than a UBEC - in other words, a UBEC should be much more reliable than all these other very reliable switching supplies all around us these days.

I would agree with you that a linear BEC within its (true) rating is as reliable as a switching one - the trouble is that the true current rating is usually too low to be useful on a 3S (or higher voltage) system. You cannot trust the "2 A" or "3 A" manufacturer rating - just try the finger test I outlined above and you'll see your yourself. Even a tenth of that current draw (0.2 A) is often enough to cause drastic heating of the BEC chip.

I don't know where ESC manufacturers come up with their linear BEC ratings. I think they take a chip that the manufacturer rates for, say, 2 amps maximum current when attached to an infinitely large heatsink (zero thermal resistance). They then put that chip on a tiny ESC circuit board with almost no heatsinking, so that it overheats with maybe only 0.2 continuous current draw - but still rate their BEC at 2 A.
Quote:
Originally Posted by srnet View Post
You also need to take account that adding the switching BEC adds in extra wires etc, so that too decreases reliability.
A wire is probably ten thousand times as reliable as a semiconductor device under these sorts of circumstances. The basic physics of it is that semiconductors have leakage currents that are exponentially proportional to their temperature - as they get hotter, they become progressively leakier and less and less reliable, and can often enter thermal runaway (the temperature will not stabilise, but will continue to increase until the chip either shuts down or is destroyed).

Wires simply do not have this sort of temperature sensitivity. Think about the RC failures you've encountered over the years - how many were related to failure of a wire? Chances are most failures you've encountered were either mechanical (stripped gear tooth, failed pushrod connector) or electronic (overheated BEC or ESC, failed driver chip in a stalled servo, etc).

-Flieslikeabeagle
flieslikeabeagle is offline Find More Posts by flieslikeabeagle
Reply With Quote
Old Jul 01, 2012, 04:26 AM
Got shenpa?
flieslikeabeagle's Avatar
Los Angeles
Joined May 2004
10,940 Posts
Quote:
Originally Posted by dmccormick001 View Post
Not that I disagree with everything you posted, but this comparison is a bit far-fetched. You're comparing a 110 volt AC device that's designed to get hot, with a 12vDC device that isn't. It's like comparing a helicopter to a Bi-plane.
Actually, the comparison isn't far-fetched at all. The voltage is completely irrelevant - doesn't matter if its one thousandth of a volt or fifty thousand volts. It's the wattage (power) that matters.

This comes down to basic thermodynamics, which is a science that's been quite well established for well over two centuries. The fact is that only two things matter: the power dissipated in the device (in watts), and the rate at which that heat escapes from the device into the environment around it.

Here's how it works: imagine 15 watts of energy dissipated in any object. It could be an elephant, a soldering iron, or a BEC chip, it makes no difference at all. In response to those 15 watts of energy, the temperature of the object starts to rise.

As soon as the object gets warmer than its surroundings, it starts to lose heat to the surroundings. The hotter it gets, the faster it loses heat energy. At some point it will be hot enough to lose energy exactly at the same rate as it is gaining energy - fifteen watts of heat coming in, fifteen watts of heat going out. When this happens, the temperature will stabilise - it won't get any hotter, but will level off.

Now, common sense tells you the elephant won't get as hot as the BEC - and common sense is correct in this case. The elephant has an enormous surface area from which to shed heat - and its temperature will barely rise the tiniest fraction of a degree, because by then it will already be losing fifteen watts of heat to the air around it and the temperature will stabilise.

The soldering iron, of course, will get much hotter than the elephant, because it has much less surface area with which to shed heat. In fact, the tip of the iron is small enough to get very, very hot - hot enough to melt solder.

How about the tiny BEC chip? It's smaller than the tip of the soldering iron, and won't shed heat as well. Without a heatsink, it would have to get hotter than the tip of the soldering iron before its temperature would stabilise. In reality, it will quickly get hot enough to either shut itself down, or to fail completely.

In reality the BEC chip on an ESC does have a tiny little heatsink - some of the copper in that tiny little circuit board acts to wick away heat from the BEC chip and dissipate it into the air. But this is not very much of a heatsink at all, and therefore the BEC chip will get very hot with very little power dissipation in it. Hot enough to fail or shut down long, long before you reach 15 watts of power dissipation in it - see, the 15 watt soldering iron wasn't really a far-fetched example at all!

As jeffs555 said, there are some things you can do to help the weeny little heatsink which is all the BEC has - most crucially, you can try to provide cooling air flow over it. This will help, but only a little - not enough to make a linear BEC powering four servos off a 3S lipo pack a particularly good idea.

In a nutshell, there was a time when electric fliers had two choices: carry the extra weight of a receiver pack (battery pack to power servo and receivers, like the glow pilots use), or cross your fingers and hope the (much lighter) linear BEC doesn't overheat and destroy your model.

Today we have a third choice: add a tiny little switching BEC, and escape both the weight penalty of the receiver pack and the dangerous unreliability of the linear BEC in one go.

-Flieslikeabeagle

P.S. By the way, like all good science, that bit about the balance between heat gain and heat loss dictating the temperature of an object applies to an incredibly wide range of situations - it explains everything from the average temperature of the earth to the reason why blankets work to keep you warm.
flieslikeabeagle is offline Find More Posts by flieslikeabeagle
Reply With Quote
Old Jul 01, 2012, 08:09 AM
Stuart
srnet's Avatar
UK, Cardiff
Joined Dec 2008
3,223 Posts
Quote:
Originally Posted by WAA-08 View Post
I'm trying to understand how a person can tell if he needs to use a UBEC or not.

Does anyone have a good method to solve this issue?
You do actually need to measure it flight to be sure.

I measured (using an Eagle Tree logger) the ESC temperature, Linear BEC output voltage and Servo current on a plane fitted with two HS81s and a HS645 (the big one). All was fine, BEC current well within the capability of the linear regulator (3A) output voltage a flat line. ESC barely getting warm.
srnet is offline Find More Posts by srnet
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion How can you tell if a lipo battery is "puffed"? mondo1948 Batteries and Chargers 9 Jul 18, 2012 10:19 PM
Discussion How can i tell if a charger will charge through balancing port? aviator4934 Batteries and Chargers 4 Jun 04, 2012 08:54 PM
Discussion Dymond D47 servo, Which version do you have on hand, how can you tell??? TDboy Hand Launch 5 Apr 21, 2012 02:26 AM
Discussion Ii was wondering someone could tell my how or if i can edit my thread title. skarface Chit Chat 4 Nov 06, 2011 11:58 AM
Discussion How can you tell when your car battery is worn out? Eddy Batteries and Chargers 39 Sep 12, 2011 11:58 PM