5volt, 5amp TO220 Regulator

[pullin-gs]

It is packaged like a regular 1.5 amp regulator, but the specs claim 5.
Part# MIC29500

With a meaty heat sink this may be a candidate for a simple bec on 12-cell and less ships. And it delivers the amps also.

Ideas?

Pullin'

[Mr.RC-CAM]

The heat dissipation issue is the Achilles heel to linear LDO VRegs. A couple amps on a 12-cell installation would be brutal. How about a DC-DC Buck or SEPIC VReg instead? With a careful layout, the EMI issues can be avoided.

RC-CAM

[pullin-gs]

Yes, that is historically the problem with on-board ESC BECS.
The 1.5amp rating isnt what kills 'em, its the heat.:(

Anyway, the device can sink it if appropriatly mounted.
I figure that 30 watts as heat at 3 amps for a 12-cell pack may be pushing it.
I'm going to mount it up (on the bench) and load up 3 amps while pushing it at 8volts,10volts, 12volts, 14volts and see how it does using my power-supply.

Is this test setup one that could produce data applicable to a servo load in a real application?

Wish me luck. :)

Pullin'

[pullin-gs]

Ran the numbers.
Heat sink (and ability of regulator) needed to sink 20+ watts is not practical for this use.
Possible to use a resistor (2ohm for less than 10 cells, 3 ohm for 10-14 cells) in series rated at 25watts with good results. External regulator heat sink is much smaller, and demands on regulator much less. However weight is still an issue.

I bot a UBEC for $30 instead. :(

Pullin'

[Zlatko]

Hi Pullin,

UBEC is nice but have a look at LM2596 . Don't be afraid of the switchers :)
A few extra components but lots of rewards ;)

Cheers
Zlatko

[vintage1]

Use of a finned sink mounted in the prop draught is very possible on aircraft and will lead to excellent dissipation.

A section cut from a scrap computer CPU is ideal.

[pullin-gs]

LM2596?
Hmmm. I'll take a look.

It's a heli application...not much air movement. So to sink a linear reg with a 4S or 5S pack may prove challenging.

[Acetronics]

:D
Some beautiful position lights bulbs ( 6 v 10 w i.e ) in a series/parallell assembly could also dissipate lots of calories ...
and give your copter the ultimate touch !!!

Alain

[Leftyretro]

LM2596?
Hmmm. I'll take a look.

It's a heli application...not much air movement. So to sink a linear reg with a 4S or 5S pack may prove challenging.

Anyone think of using the 2S 'tap' on the battery back to power the BEC regulator chip?
This would relive the regulator from having to dissipate a lot of voltage from the 4S (or5S) battery output. I guess the biggest problem would be that it might lead to an unbalance condition in the pack as two of the cells would be discharging at a slightly higher rate then the remaining cells. Might be manageable if one always insures that they check and rebalance the pack after every use as some do anyway.

Really effcient switching regulators is the best solution.

Lefty

[Bob Baker]

If you want to try a linear, look also at the 78T05, rated at 3A, which should be enough. These are not Low Drop Out, but that is not needed for 3S, 4S etc; and they are cheap at around $1.70 rather than $7 or so for an LDO. And I'm not convinced that the heat issue is such a problem, at least the way I fly; the average current is going to be way below peak.

The LM2596 looks really simple to use. There is also a LM2576, which runs at 50kHz rather than 150kHz for the 2596; this requires a larger inductor but may (or may not) have a radiated noise advantage.

There are also some evaluation boards around that may be useful. I have an LM3477 board that someone sent me for free. I hooked it up yesterday and it works fine, but the output looked ratty, so I added some filtering. It runs at 500kHz. Its a bit more complicated than the 2596 circuit. It is more efficient, about 90% rather than 80% for the 2596.

I actually prefer the linear solution if heat turns out not to be a problem, since it is simpler, less prone to failure, and less prone to cause interference.

[Zlatko]

:D
Some beautiful position lights bulbs ( 6 v 10 w i.e ) in a series/parallell assembly could also dissipate lots of calories ...
and give your copter the ultimate touch !!!

Alain

Very nice idea and I too like lots of lights but consider what will happen to the filament in the high vibration environment :)
At least you will deffinitely know when you are going to crash :)
"Dood, the light went off .... " " Ohhhhhhh ..... #%@* ! " :)

Cheers
Zlatko

[Acetronics]

Hi, Zlatko

Not too difficult to prevent a bulb failure ... with i.e. diodes in series parallelled to the bulbs.
Same as Schottky diodes used to prevent a Nicad "open circuit " in a batt pack ...

a small security pack ( less than 50 grams ! ) would here be the best : No more lights: time to land ...quickly !!!

Alain

[Zlatko]

Alain,

Thanks for that, I didn't even think about those possibilities :)
I like the back up battery option.

Cheers
Zlatko

[Warske]

Not too difficult to prevent a bulb failure ... with i.e. diodes in series parallelled to the bulbs.
Same as Schottky diodes used to prevent a Nicad "open circuit " in a batt pack ...You would need to use zener diodes on the load side. Each would need to be able to dissipate the power dissipated by the bulb.

Warske

[Acetronics]

Hi, Warske

I wouln't choose Zeners here , but gangs of 1N 540x ...or BY 550 or P600.
I do not like fryed zeners ...

The cost is much less, too !!!

I prefer the security batt ... if you want the truth !!!

Alain

[Warske]

I wouln't choose Zeners here , but gangs of 1N 540x ...or BY 550 or P600....I prefer the security batt ... if you want the truth !!!I can't work out how you could use 1N540x in this application. I understand how they work on the battery side, but not the load side. Sorry, I must still be asleep. A diagram?

The security battery does seem like a good idea.

Warske

[vintage1]

:D
Some beautiful position lights bulbs ( 6 v 10 w i.e ) in a series/parallell assembly could also dissipate lots of calories ...
and give your copter the ultimate touch !!!

Alain

Pulsing in sympathy with servo movement...its a good idea. Bulbs are DESIGNED to dissipate power.

[Warske]

I wouln't choose Zeners here , but gangs of 1N 540x ..OK, I think I figured out what you meant. What threw me was "Same as Schottky diodes used to prevent a Nicad 'open circuit' in a batt pack."

You would effectively build each zener out of 3 amp silicon diodes. So to drop 12v using light bulbs, you would also need about 19 large silicon diodes to protect against the light(s) burning out.

Interesting idea!:)

Warske

[Bob Baker]

I just got some 78T05's, and hooked up one to the receiver on an airplane with 4x HS-81's (esc unplugged). I ran it from a bench supply set to 16V, which would be close to max for a 4S system. The regulator was mounted on a small heat sink; around 2-3 sq in total; in still air.

I yanked the controls around at a much higher rate than I would normally do in flight, and still did not experience any significant heating problem after a couple of minutes. It was a bit warm, but not hot. Of course there were no flight loads; just linkage and control surface loads.

Not a very scientific test, but it looks good enough for me.

[Acetronics]

Hi, bob

Static tests are OK ... but The LM 78xxx series cut the power when overheated ...

If you like to play with a Russian 44 Caliber ... just try it "in flight" ; a switching supply is not so dangerous - if well designed !!!

Alain

[Warske]

Hi, bob

Static tests are OK ... but The LM 78xxx series cut the power when overheated ...

If you like to play with a Russian 44 Caliber ... just try it "in flight" ; a switching supply is not so dangerous - if well designed !!!

Alain Essentially all IC regulators cut the power when overheated, including the switchers. Check the data sheet for the LM2596 switcher that Zlatko mentioned.

The reason for this is simple: If you design a regulator for general use and it doesn't cleanly shut down, it could end up frying the equipment. In the case of a computer system, it can cause data corruption.

In a situation where a linear regulator does the job, using a switcher is actually more dangerous, because it contains more parts that can fail.

If you use the BEC in an ESC, you are most likely using a linear regulator, but one that has several disadvantages:
- It has to contend with the heat load from the rest of the ESC circuitry.
- Its only heat sink is copper traces on the circuit board.
- Its all wrapped in heat shrink tubing, which is a good thermal insulator.

Add to this the fact that the ESC is often mounted inside a poorly ventelated area, not far away from a hot battery. And yet people are flying with these every day.

Using a separate linear regulator, even without a heat sink, is a huge improvement over that situation. You just have to be aware of what your battery voltage and load current is, and how this affects the power dissipation. With high enough current and voltage, you are probably better off with a switcher, but there is a middle range where a linear regulator (or several) makes more sense.

Not a very scientific test, but it looks good enough for me.Bob, yours is a voice of sanity in a wilderness of hysteria. You did more testing than most people who trust their planes to the linear regulator in an ESC every day. Unless you are flying a jet, I doubt your load will change much in the air. Sounds like your heat sink is bigger than it needs to be.

Warske

[Bob Baker]

With high enough current and voltage, you are probably better off with a switcher, but there is a middle range where a linear regulator (or several) makes more sense.

Bob, yours is a voice of sanity in a wilderness of hysteria.

Warske

Yes, at least at 4S or 12 nimh I am convinced that you do not need a switcher. I have several Castle Creations escs, which use linear regulators as you described, and also have thermal shutdown. They work fine at 3S, even with their very limited ability to dissipate heat.

I'm not so sure about the sanity part.

[Acetronics]

Using an external TO3 ballast transistor ( MJ 2955 w/ heatsink i.e) whith a 78xx ( see higher current schemes in the 78 datas ...) could solve the temp. shutdown problem ...and allow higher peak currents too ...so, a simple TO 220 7805 is usable here !!!

Alain

[vintage1]

If you think that an IC engine is probably 25% efficient, and generates maybe 700 watts of power you can get an idea of how hot the other 2.1KW of wasted power makes a lump of finned heastink and the exhaust ...

Good design and a bit of finned heatsink in a good draught will easily take you up to several hundred watts of safe dissipation. Ok you normally have to split that over several devices to avoid hot spots, but in principle its doable, and I have many audio amp designs out there to prove it!

To do 2 amps at say 20 cell level is only 30W dissipation. A speed 400 dumps that and stays below boiling point. Mostly :D

Warske has it in a nutshell. Most BEC's are very poorly heatsinked, with no provision for cooling air to circulate. In my yoof, we used little TO5 transirtotrs - generally 500-800mW rating. Simply clipping on a curly heatsink or some fins would TREBLE the dissipation to safe sort of 1.5W or so. T0220 is similar - about a watt in free air, but stick it on a plate of alloy, and its 5-10W easily. Run some air, and add a couple of fins over it and maybe its 25-30W.

[Bob Baker]

Using an external TO3 ballast transistor ( MJ 2955 w/ heatsink i.e) whith a 78xx ( see higher current schemes in the 78 datas ...) could solve the temp. shutdown problem ...and allow higher peak currents too ...so, a simple TO 220 7805 is usable here !!!

Alain

Note that the 78T05 I am using is a distinctly different part from the 7805; rated at 3A rather than 1A. No extra transistors are needed in this case.

[Acetronics]

:) There's also the old computer's welders 78P05 ... in a TO3 case and rated 5 Amps ... ( note LM 338 K is easier to find now )
TO3 is simply much better to dissipate heat ...
- Just know I have just finished a 5 cells batt coupler - 5.5 v regulator for F3X planes ( rated 5 amps continuous and ...116 peak !!! ... 5.35v output for the batt nominal 5.5 v end of discharge point ...under 5 A)
;) Someone better ???
So, understand here I had a look to whatever is sold as a regulator ...

My option here would be to use a separate battery for the radio ... not really heavier than a regulator plus heatsink ... but much, much safer !!!

Alain

[Warske]

Using an external TO3 ballast transistor ( MJ 2955 w/ heatsink i.e) whith a 78xx ( see higher current schemes in the 78 datas ...) could solve the temp. shutdown problem ...and allow higher peak currents too ...so, a simple TO 220 7805 is usable here !!!
Alain Alain, you raise an interesting point about thermal shutdown. How should it work ideally for an R/C airplane?

If you go into thermal shutdown, the servos will stop in their tracks, and the ESC will probably shut off the motor (true for a CC ESC, anyway). The regulator will start to cool off, and if high enough, you may get control back before the plane hits the ground.

Now what happens if you bypass thermal shutdown using an external transistor as you suggest? When the transistor overheats, it will probably fail shorted (the usual failure mode for silicon BJTs), and your receiver and servos will be in flames from the excess voltage before the plane hits the ground. Not an improvement.

But look at what happens when using an ESC with a built-in BEC: When things start to get hot the motor drive shuts down before the BEC overheats. This reduces the thermal load, and maybe the aerodynamic load on the servos. The pilot is also motivated to land quickly, thinking the battery is about to expire.

So if you use an external BEC, linear or switcher, an interesting approach might be to monitor the BEC temperature and kill the motor when the BEC starts to get hot. But that's only if you're paranoid.:) I'm not convinced there's actually a problem.

Warske

[Acetronics]

Hi, Warske
The problem here is not for the ESC, but the R/C equipement ... if the ESC overheats ...we do not care, here.
So, killing the motor doesn't change anything to our problem ...
NO Thermal ( or other reason ...) shutdown is allowed here, The regulator MUST work !!!
so, we just have to oversize it ... not to hope it will work !!!

External ballast permits high current peaks ... that's the reason I use it. Now, have the regulator on the same heatsink permits a thermal shutdown too.

Vintage says 30 w dissipation OK, seems safe . 30 w and 100 °C temp elevation give a less than 3.3°C/Watt a heatsink.
Say SP 38 model ( 60 x 43 x 38 mm ) or a S25 ( 100 x 80 x 13.5 mm ) or ML 97/2 ( 42 x 25 x 64 mm )
That's thermics - no more -
And is sure to work properly.

Alain

[Warske]

...killing the motor doesn't change anything to our problem ...
NO Thermal ( or other reason ...) shutdown is allowed here, The regulator MUST work !!!
so, we just have to oversize it ...I very much agree with you.

What I was trying to point out is that there is another way, instead of greatly oversizing the BEC.

If you knew the BEC was about to overheat, you could land the plane before it actually does.

One way to signal the pilot that the BEC is starting to get hot is to shut down the motor. Or you could use flashing lights or rocket flairs.

This signal-the-pilot feature is already in a CC ESC, and may help to explain why they work so well even though handicapped as I said earlier.

Having pointed that out, I think the more practical way to design an external BEC is to oversize it, as you say.

Warske

[Acetronics]

I don't think it is , psychologically, a good thing for the pilot to ask himself every 3 sec " will the BEC overheat ??? " , " will I have enough motor " , "will I land correctly "...
Fly wouldn't be a pleasure, in those conditions ...

Alain

[pullin-gs]

I really like the idea of the light-bulb. Maybe a bunch wired in parrellel to avoid the burn-out issue if one blows?:)

I may have to give this a try.....maybe make it a spotlight like the police helis have. :D:D

Pullin'