Author |
Message |
Cyeos
Joined: Mar 17, 2010 Posts: 7 Location: Baltimore
|
Posted: Fri Aug 11, 2017 8:24 pm Post subject:
Voltage Regulator Overheating - Heat Sink Needs? Subject description: Troubleshooting big circuit bent project running off LM7806. |
|
|
Hello!
Potential newb question.
Having trouble finding existing related topics at the moment, and other electronics forums touch on the subject but are often going over my head.
I have a bent circuit using an LM7806 voltage regulator off of a 12V DC wall wart to get 6V DC. It is a TO-220 package, currently using this little heat sink:
http://www.mouser.com/Search/ProductDetail.aspx?R=230-75ABvirtualkey56720000virtualkey567-230-75AB
Other parts of the circuit want 12V, so that's why I'm not just using a 6V adapter to begin with, which would obviously be easier.
It works, but I measured about 320mA being drawn from the 6V line, at which the regulator gets too hot to touch very quickly. The bulk of the heat must be coming from the 6V difference from supply(?)
If I understood what I read around at all, I am needing to dissipate around 2W.
1) Should this overheating be happening at that level of current draw?
2) If the overheating is expected, I imagine I need a bigger heat sink, but I do not understand the "thermal resistance" system to figure out how much bigger I need to go than the heat sink I am currently using. Can anyone tell me how to calculate?
Thanks in advance! |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
Posted: Sat Aug 12, 2017 3:35 am Post subject:
Re: Voltage Regulator Overheating - Heat Sink Needs? Subject description: Troubleshooting big circuit bent project running off LM7806. |
|
|
2W seems about right and yes they can get pretty hot with that kind of power dissipation. I've never done
any calculations myself and it's not that easy as there are a lot of variables. Usually I just put a heatsink on
it and if it gets too hot to touch I use a bigger one. A quick search led me to this page which might be
helpful and shows how many variables there are.
Something else you might be able to do is reduce the current draw for the 6V section. For example if there is
an amplifier in it which is drawing most of the current you might be able to replace that with a 12V one. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
|
gdavis
Joined: Feb 27, 2013 Posts: 359 Location: San Diego
Audio files: 1
|
Posted: Sat Aug 12, 2017 10:28 pm Post subject:
|
|
|
I played around with the calculator in PHOBoS' link with the numbers from the heat sink and LM7806 data sheet and the 2W of dissipation (sounds about right). Came up with a heat sink temp of about 40C which would be pretty warm (edit: used wrong value, see two posts down). According to a quick google, too hot would be about 50C to 60C.
The problem is that voltage regulators are meant to do just that - regulate. You've trying to convert 12 volts to 6V which is a 50% drop. That's asking a lot. As you would probably surmise, you can do it as long as you keep the current down and/or heat sink adequately, but it's less than ideal.
You may want to look into a DC-DC converter which is more efficient and may be accurate enough depending on what your powering, A combination of a DC/DC converter and regulator might be something to consider if you really need an accurate 6V.
Like PHOBoS I also wonder what you're powering and if it really needs 6V. _________________ My synth build blog: http://gndsynth.blogspot.com/ Last edited by gdavis on Sun Aug 13, 2017 1:56 pm; edited 1 time in total |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
Posted: Sun Aug 13, 2017 4:20 am Post subject:
|
|
|
hmm 40C actually seems pretty low with 2W dissipation and such a small heatsink. I did try some
calculations myself and it depends a lot on the ambient temperature you use. Without a heatsink it
gets easily above 100C. A DC-DC converter might indeed not be such a bad idea it's definitely more
efficient, although it could cause some unwanted noise. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
|
gdavis
Joined: Feb 27, 2013 Posts: 359 Location: San Diego
Audio files: 1
|
Posted: Sun Aug 13, 2017 1:16 pm Post subject:
|
|
|
PHOBoS wrote: | hmm 40C actually seems pretty low with 2W dissipation and such a small heatsink. |
Oh! I was using the 7.5 C/W quoted in the mouser page. I just looked at the data sheet and that value is with forced convection. Natural convection is 57C over ambient at 2W which is 82C with 25C ambient. Using 28.5C/W in the calculator gives the same value. _________________ My synth build blog: http://gndsynth.blogspot.com/ |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
Posted: Sun Aug 13, 2017 2:03 pm Post subject:
|
|
|
I hadn't even noticed that and used the 7.5C/W too and I got similar results (around 40) untill I read this part:
Quote: | Some people choose 20 °C for the room temperature, some others choose 25 °C or even 27 °C (300 K), but these temperatures are too low for our calculation. One should always consider the maximum possible working temperature of the circuit: in hot summer day you may easily exceed the above values. For home applications, values around 50 °C are usual, for industrial applications is not uncommon to go up to 60 °C or more.
Than, if you don't have a blower and all the cooling is done by natural convection, it's safer to further increase it by 5 to 10 °C, especially if your heat sink is internal. The reason is because without a blower the temperature in the neighbourhood of the heat sink is higher than in the rest of the room. |
_________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
|
gdavis
Joined: Feb 27, 2013 Posts: 359 Location: San Diego
Audio files: 1
|
Posted: Sun Aug 13, 2017 2:28 pm Post subject:
|
|
|
I used 25C since that most likely matches the situation of sitting there testing out your new circuit with the chassis open and putting your finger on it to see if it's getting hot. But ya, at some point the different ranges need to be taken into consideration before you pack it in a chassis and expect it to work over a variety of conditions. _________________ My synth build blog: http://gndsynth.blogspot.com/ |
|
Back to top
|
|
|
stolenfat
Joined: Apr 17, 2008 Posts: 476 Location: Sunny Oakland California
Audio files: 1
|
Posted: Thu Sep 14, 2017 9:37 pm Post subject:
|
|
|
I don't want to discourage your exploration with circuit bending, and I'll say i don't know how much experience you have with circuitry (given your posting here about THIS question and given the very intellectual back and forth you likely know whats up) but to any folks still leaning... I'd say you were bending in dangerous territory. Circuit bending can lead to some whackiness and you never know with bending - things can happen and you might get hurt.
Anyways... I don't think it's been suggested but why a 12v adapter... could you just maybe get a 9v or lower? Maybe that might save on your heat issue _________________ home made noise and electronic ill-logic |
|
Back to top
|
|
|
Cyeos
Joined: Mar 17, 2010 Posts: 7 Location: Baltimore
|
Posted: Mon Oct 30, 2017 9:15 pm Post subject:
Update |
|
|
Hi again!
Thanks everybody.
What do you guys think of this?
http://www.futureelectronics.com/en/technologies/electromechanical/power-supplies/dc-dc/Pages/1010233-OKR-T-3-W12-C.aspx?IM=0
You say they cause noise? Will the usual cap filters fix that?
Not sure I understand how a DC-DC converter works.
I did manage to drop to 9V supply - the higher voltage was originally to suit a Synthrotek Vac-Pak CV circuit, but they work fine on 9V. The circuit bent part of the project itself only wants 6V.
At 9V supply the 7806 regulator doesn't get hot enough to burn you, but still uncomfortably hot with the current heat sink. I'm sure it's lifespan will be affected. |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
|
Back to top
|
|
|
GlowSignal
Joined: Dec 09, 2018 Posts: 1 Location: EU
|
Posted: Tue Dec 11, 2018 10:17 am Post subject:
|
|
|
Getting right heatsink seems like more difficult than building good psu, although i found one has to understand only a few basics to get sufficient heatsink. I tend to regulate heaters which can be a liittle hard at first because of lots of heat in some cases. Not so new series of cheap TL/LM108X regulators offer lower dropouts compared to ussual variable LM317, or other fixed ones.
After my stock of huge heatsinks, instead of understading how to calculate what's needed dried up, this article teached me and showed further reading for understanding how to buy buy cheapest one that work well:
http://sound.whsites.net/heatsinks.htm#s22
HT or lower voltages can be had from very simple capacitance mulltipliers where zeners help avoid heatsinks and still provide extremely low ripply if done properly. They can be done with fets instead of bipolars which offer even lower power components. Such circuits work very good for voltages over 40V which can't be done with ussual regs. Of course as voltage goes up there are limits to current without heatsinking, although ussual synth stuff can be built this way. Power amps can't:
https://static.flickr.com/139/324693062_1bcc2f4ae2.jpg
Hope it helps people who don't wan't to rely on now cheap subminiature tubes that will become very expensive with time, as did even ussual NOS as ECC82. I rather don't get into pentodes like EF86 which are very expensive NOS (of course what's left are mostly rejects or very expensive), fortunately both types and many more are still in production sounding the same as NOS, lifetime is a bit shorter....
Cheers |
|
Back to top
|
|
|
|