Calculating Servo Draw or How Can I Tell I Need a UBEC? - Page 3 - RC Groups
Jul 06, 2012, 06:48 PM
Got shenpa?
http://www.servocity.com/html/hs-81_micro.html

Here's a snippet from that page:
Quote:
 Current Drain (4.8V): 8.8mA/idle and 220mA no load operating Current Drain (6.0V): 9.1mA/idle and 280mA no load operating
Note that at a 5 volt operating voltage the specified no-load operating current is three times higher than srnet's estimate of 70 mA. When operated at 6 volts, Servo City's specified current draw is four times higher than srnet's estimate of 70 mA.

I could not find HS-81 current draw data on Hitech RCD's website (other specs for the HS-81 are here: http://www.hitecrcd.com/products/ana...ini/hs-81.html). So I presume the current data on the Servo City website came from measurements made by someone at Servo City.

The term "no load operating" is not defined on that page, but presumably it means moving the servo (operating it) with no additional load connected. In other words, linkage and hinge friction, and aerodynamic loads on the flight controls, would raise the servo current even higher than those measured numbers of 220 mA / 280 mA @ 5V and 6V respectively.

If these numbers are correct, four servos being operated at the same time with actual flight loads on them (this happens when flying rolling circles, for example) would together draw more than 880 mA in the best case - it's probably conservative to guess at one ampere total current draw.

With a 3S battery - say 11 volts into the linear BEC and 5 volts out - that would translate to around 6 watts of heat dissipation in the BEC chip(s).

Simple thermodynamics says that 6 watts of heat dissipation in a tiny BEC chip is definitely worthy of a raised eyebrow and a quiet whistle of surprise.

Tests? The simplest test that will tell you something useful is the one I keep mentioning. Plug in all the servos and the battery (but no motor), identify the BEC chip on the ESC circuit board, put your fingertip on it, wiggle the transmitter sticks to get the servos moving. (It's not a bad idea to have cold water standing by to dip your burned fingertip into after the test.)

This test is so simple that it's worth trying for anyone who's even a little bit curious about this issue of linear vs switching BEC's. In a nutshell, if it's too hot for your fingertip, it's too hot for a silicon semiconductor device to operate reliably.

I worked as an engineer in an electronics company at one time. Lots of testing and measurement went into the final heatsink design on all our products. But the first and simplest test at the early prototype stage was usually the finger test - if if didn't pass that, back it went for a re-design before we even bothered with the more sophisticated testing.

We can sit around here and argue all day about whether linear BEC's might or might not cope with the loads we put on them. We can design measurements and tests that tell us just how close to meltdown linear BEC's can be pushed without actually crashing our models. We can play Russian Roulette by flying our models with linear BEC's and hoping they won't overheat and shutdown, this time, as long as we don't wiggle the sticks too much.

Or we can slap on a light, inexpensive UBEC that completely eliminates any worries about the BEC overheating, and simply go fly.

I chose the last route, not because "the truth is not acceptable", but because simple physics tells us that a linear BEC is an obvious weak point in our RC models electronics. It's vital to the model, and it is also very highly stressed, and susceptible to failure as a result.

There was a time when there was no other alternative to living life on the edge with an overstressed linear BEC, unless the model could carry the weight of a receiver (battery) pack. But thankfully that time has long gone, with the arrival of small integrated circuits that allow simple and efficient switching UBEC's to be built at low cost and minimal weight.

So these days it is cheap and easy to eliminate that weak point in our model's electronics. From my point of view, it makes no sense to sit around and worry if I might just be able to get away with a linear BEC. I already know there is a much more trustworthy alternative - and it's affordable and lightweight too. Why on earth wouldn't I use it?

I fully realise that it is human nature to argue and disagree; universal agreement will never happen on anything, not even the colour of the sky. So if you prefer a linear BEC, fine, just be very aware of the BEC heat issue and do what you can to avoid falling victim to it. If you prefer a switching BEC, fine. You have no need to worry, just go fly!

-Flieslikeabeagle
 Jul 08, 2012, 06:26 AM Registered User I've had Becs stop working before in flight on power glders. Not too much of a problem as you can shut down and regain control. Not a problem with a glider. With a heli or an aerobatic plane a loss of power can be disasterous especially with computer based rxs which can take a few seconds to recover. You have missed one point. Some BECS use regulators like the 7805 which go into foldback at 1 amp current limit. I had one of these on my Twister Storm heli which was allegedly designed for 4 small servos. Current draw depends on how hard the servos are being driven and stiff hinges or servos hitting the end stops will increase the current dramatically. There is a big difference between something that works most of the time and something that works almost all the time. I would certainly advise any beginner to go for a 3 amp SBEC and to test it on the ground first to see if it really is 3A capable. Too much of our equipment is badly or optimistically specced. But it is cheap and by not using it to the spec but underrating it we can get get reasonable reliability. I have a STBlaze with a built in BEC which may or may not be SBEC. It appears to work. I use it because I know what might happen and I accept the risk. A beginner won't know and should be more conservative.
 Jul 12, 2012, 04:58 AM Just Learning I've just been reading this thread to decide whether to install a separate ubec or not in my powered glider. I may be missing something here, so please correct me as I could be wrong. The graphs of current draw by servos mentioned earlier in this thread at http://homepages.paradise.net.nz/bhabbott/Servo.html shows peak currents for various servos. He also states in his text: "Analogue servos only power the motor once per servo pulse (~50 times per second), so they tend to buzz and lose position under high loading." This is because the signal to each servo is PWM (pulse width modulated) between 1ms and 2 ms within a frame of 20ms. My thinking is that if current draw by each servo really is only present during the pulse width for that servo, then each servo in turn will draw current during a different time slot within the overall frame rate. So, regardless of how many servos are being driven, the peak current draw from the ubec will only ever be as high as the one servo drawing most current. If this is true, then just about any ESC with BEC (normally providing 2amps from the BEC at least) is more than adequate for any number of servos in a typical model.
Jul 12, 2012, 08:05 AM
Quote:
 My thinking is that if current draw by each servo really is only present during the pulse width for that servo, then each servo in turn will draw current during a different time slot within the overall frame rate. So, regardless of how many servos are being driven, the peak current draw from the ubec will only ever be as high as the one servo drawing most current. If this is true, then just about any ESC with BEC (normally providing 2amps from the BEC at least) is more than adequate for any number of servos in a typical model.
It doesn't work like that and the current to the servo motor does not occur only during the servo pulse from the receiver.

In the typical analog servo, the input pulse goes into a pulse width to voltage converter. The conversion to a voltage occurs during the 1-2mS servo pulse and at the end of the pulse the voltage is captured and held until the next pulse. This voltage is compared to the voltage on the feedback pot and if they are different, current is applied to the motor. The time that the current is applied to the motor depends on how far the motor needs to turn. If you are moving the stick slowly, like in the waveforms on Bruce's page, the current will occur in short pulses. If you are moving the stick faster, the pulses will be much longer. If making a lot of rapid stick movements the current will be continuous.

Again, the peak and average currents required by a servo depend on the mechanical load on the servo arm and the rapidity of the stick movement. Worst case might likely occur during recovery from a spiral dive, which would also be the worst time for your receiver and servos to shut down.
 Jul 12, 2012, 03:58 PM Stuart What was the situation before we had readily available, and cheap, switching regulators ? Were linear BECs to dangerous to use then also ?
Jul 12, 2012, 05:42 PM
Got shenpa?
Quote:
 Originally Posted by srnet What was the situation before we had readily available, and cheap, switching regulators ? Were linear BECs to dangerous to use then also ?
Yes. Very much so, for some years. So much so that I was buying surplus automotive cellphone chargers from electronics surplus stores, taking them apart, and modifying them into switching UBEC's to use in my models.

Search RCG and you should find a couple of threads from 2005 or so where I shared this information here on RCG, setting off a hunt by several RCG members for the best and cheapest source of hackable cellphone chargers. Lucien Miller (now owner of Innov8tive Designs, the US distributor for Scorpion motors) hit paydirt, finding some inexpensive, high-current chargers down in the San Diego area that were easily modified to make very nice switching UBEC's for RC planes. He created a popular "how to" thread on these here on RCG, complete with detailed photos showing how to hack them for use as UBEC's.

A little RC history, pre-dating the era of hacking cellphone chargers into UBEC's:

Linear BEC's arrived in the days of six-cell NiCd batteries powering brushed Speed 400 motors on big, slow airframes with limited flight performance due to the high weight and limited power of those early electric power systems.

On models of that era, linear BEC's did okay. Incoming voltage was low at around 7.0 - 7.2 volts, so the BEC only had to drop around 2 volts across itself, limiting heat dissipation. Servos were costly and heavy so many electric models of the time only used two servos. Airspeeds were limited and aerobatics were gentle, so aerodynamic loads were minimal. Under these conditions, power dissipation in a linear BEC was acceptable.

Then brushless motors and lithium-polymer flight batteries arrived. Almost overnight typical battery voltages shot up to 11+ volts instead of 7. With the BEC still putting out 5 volts, there were now 6 volts dropped across the BEC instead of 2 volts, nearly tripling heat generation in the BEC. At the same time servos got lighter and cheaper, and four-servo models became more common, doubling the load on the BEC. Airspeeds and aerobatic capability increased, putting heavier aerodynamic loads on the flight controls, increasing servo current draw and making the poor BEC suffer even more.

At this point in RC history linear BEC's had gone from "adequate" to "desperately in need of improvement".

ESC manufacturers responded with the cheapest and easiest to implement "improvement": they started to solder two BEC chips one on top of the other, and pretended that this doubled the current capacity of the BEC. In practice it does no such thing, because stacked chips are very poor at dissipating heat. I still remember the string of in-flight BEC failures I and several of my flying buddies experienced at the hands of the Castle Phoenix 25, circa 2005 or 2006.

At around the same time microcontrollers started to make their way into formerly all-analogue RC radio receivers. First we had DSP 72 MHz receivers like Bergs and later Castle Creations ones, then all-digital 2.4 GHz receivers from Spektrum and others. All of these receivers shared a couple of key differences from the previous generation of all-analogue ones: digital receivers tended to draw more current to power the onboard microprocessor, and digital receivers were much more sensitive to brown-outs.

A few millisecond long BEC glitch on an analogue receiver would go completely unnoticed. With a digital receiver, the same BEC glitch could reboot the microcontroller, in some cases causing loss of radio control for several seconds while the microprocessor rebooted and resumed operation. More than long enough to crash the model.

Go back and search RC Groups circa 2005/ 2006, and you'll find much discussion on digital RC radio receivers from Berg and Spektrum that supposedly "locked up", causing the destruction of many models. Almost certainly many (if not all) of these lock-ups were actually caused by BEC failure, not receiver failure. I myself saw multiple Berg-equipped models crashed due to loss of radio control, and after the original Spektrum DX6 came out I also saw the same happen to models carrying Spektrum receivers.

I had been a Berg user, and experienced at least one lockout using one. When the DX6 came out, I switched to Spektrum. Not long after I myself lost one model (an aerobatic four-servo Hotpoint EP) to a "Spektrum lockup" that was most probably actually a linear BEC shutdown.

At that point in time, linear BEC's had very much failed us in this hobby. Technology had pushed them from "okay" to "a very bad idea", and they were causing many lost models.

And that's when I started making my own switching UBEC's out of surplus cellphone chargers, and later buying inexpensive ones from Hobby City (now Hobby King).

Guess what? From the day I first switched to UBEC's, I have never experienced a single loss of radio control. I continued to fly my old DX6 and AR6000 receivers, but there were no more "Spektrum lockouts". Later I upgraded to a DX7 and used larger Spektrum receivers on some of my bigger and heavier models, always using a UBEC to power them. Hundreds of flights later, I had not experienced a single glitch, or a single instants loss of control.

Srnet, by now your point of view is quite clear. It seems you've already made up your mind about linear BEC's and are not about to change your opinions based on any new information received. Nor are you going to try the super-easy finger test to find out for yourself how hot linear BEC's get under quite light servo loads.

So I wish you well, and I hope your linear BECs reward your faith in them, and never betray you so that you have to stand by and watch while your favourite model smashes into the ground, completely unresponsive to your controls.

I, on the other hand, HAVE experienced linear BEC failure and the loss of a model, not to mention the danger to the members of the public all around the park I was flying at. Having subsequently investigated the issue, found the cause, and found a way to eliminate these treacherously unreliable components, there is no way I'm going back to them - except on small, light models using 2S lipo packs and no more than 2 servos.

-Flieslikeabeagle
 Jul 13, 2012, 11:28 AM Just Learning Thanks flieslikeabeagle for a very thorough explanation of the history of the esc with bec through to independent ubec. Now I know this there's no more room for letting feelings get in the way, but a solid factual understanding. Also thanks to jeffs555 for the theory of operation. I'm now definately going to use a decent ubec from now on. Only decision now is whether to go for the cheaper Hobbywing units or pay a lot more for something like the Dimension Engineering ones. Does anyone have any experience of the reliability of either of these?
 Jul 13, 2012, 11:35 AM Promoting Model Aviation... Here are two quality ones that I know of: https://www.rcgroups.com/forums/showthread.php?t=748220 http://www.dimensionengineering.com/products/parkbec I'd also like to know if one stands out more than the other for which one to use. Thanks, Frank
 Jul 13, 2012, 11:45 AM Just Learning Anyone know a source of the Castlecreation ubec in the U.K.?
Jul 13, 2012, 02:07 PM
Got shenpa?
Quote:
 Originally Posted by mpeill Only decision now is whether to go for the cheaper Hobbywing units or pay a lot more for something like the Dimension Engineering ones. Does anyone have any experience of the reliability of either of these?
I've used plenty of the cheap ones from Hobby King, and have never had the slightest problem with any of them. They have worked perfectly for me.

I usually get the UBEC's that can be set to either 5 V or 6 V, and I prefer to set them at 6 volts. That gives a little more headroom to make sure the digital radio receiver never reboots due to undervoltage. As a bonus the servos are a little quicker and stronger as well.

This UBEC was one of my favourites, because it was light, inexpensive, and capable of 5V or 6V operation: Turnigy 3A UBEC

I have not examined the Castle or Dimension UBEC's, however there's a very good chance the same low cost, high quality, mass produced switching regulator chips are used in both the cheap and the expensive UBEC's. These switching regulators (which is what a UBEC is) are now everywhere, in chargers and power supplies for all sorts of electronic devices, from cellphones and computer monitors to laptops and GPS units. There are even switching regulators inside the cellphone itself, for example to control the brightness of the LED's illuminating the display. (Linear regulators waste too much power, and power is precious in a cellphone.)

So by now the technology is well tried and well developed, and we can expect good levels of reliability.

Back to our application (UBECs for RC models), if you need truly extreme levels of reliability, it is actually possible to use two independent UBEC's, with the outputs wired together using a pair of inexpensive Schottky diodes so that if one fails, the other seamlessly takes over.

If I was ever flying a turbine jet, or a large and expensive (and potentially deadly) giant scale model, or an aerial photography platform carrying lots of very expensive camera gear, I'd probably use dual UBEC's on it for added insurance against BEC failure. (You can also use dual receiver battery packs, or pair a UBEC with a receiver pack, using the same two-diode technique.)

But I never flew any models like that, and I'm pretty sure those cheap Turnigy UBEC's are more reliable than the new weakest link in the system: me, the imperfect and fallible human RC pilot!

-Flieslikeabeagle
 Aug 19, 2012, 03:03 PM Promoting Model Aviation... I finally had the chance to do the test. I hooked up my battery to my CC T-bird 36 ESC and cycled the servos all around for about 5 seconds. The ESC still got warm. This confuses me. I did not think any servo current should ge going through the ESC when my CC BEC is hooked up. Any thoughts? Frank
 Aug 19, 2012, 05:20 PM "Simplify, then add lightness" When you use an external BEC, you need to pull the center(red) wire out of the connector on the ESC that goes to the receiver and put heat shrink or tape over it. If you don't, both BECs will be trying to supply power.
Aug 19, 2012, 05:35 PM
Promoting Model Aviation...
Quote:
 Originally Posted by jeffs555 When you use an external BEC, you need to pull the center(red) wire out of the connector on the ESC that goes to the receiver and put heat shrink or tape over it. If you don't, both BECs will be trying to supply power.
I did that and that is why I am surprised about the heat in the ESC. I'm going back out now and try it again and see what happens. Perhaps it may have been still warm from the motor running.

Frank
 Aug 19, 2012, 05:40 PM Promoting Model Aviation... OK, after about 15 seconds of moving the servos, the ESC did feel a little warm. I also did not see the amp draw go above one (1) amp. Thoughts? Frank
 Aug 22, 2012, 10:11 PM Promoting Model Aviation... Any ideas on why the ESC should be getting warm? Thanks, Frank