heat sinks and LM1875

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I don't really understand how to calculate the amount of heat sink I'd need. I picked up a heat sink from a computer processor for $1, split into two sections of fins, each half is about 1x 1.3 x 2.5". The data sheet specs 1.4dC/W. Looking in a Mouser catalog, I see heatsinks rated at 3.5dC/W, 6dC/W etc. How do I use the spec to figure out how much I need. Thought about just putting the kits together on a breadboard and see how hot they get.
Any tips on this?
Thanks:smash:
 
Sparky OR said:
Thought about just putting the kits together on a breadboard and see how hot they get.
Hi,
that's as good a way as any, if you are prepared to buy different bigger sinks when you experiment with your temperatures.

Try for the chip being cool to the fingers when no signal is present
Try for warm heatsink when a normal signal is present. The chip will be hotter than warm.

Or just read the National datasheet and double the suggested size.
 
Sparky OR said:
I don't really understand how to calculate the amount of heat sink I'd need. I picked up a heat sink from a computer processor for $1, split into two sections of fins, each half is about 1x 1.3 x 2.5". The data sheet specs 1.4dC/W. Looking in a Mouser catalog, I see heatsinks rated at 3.5dC/W, 6dC/W etc. How do I use the spec to figure out how much I need. Thought about just putting the kits together on a breadboard and see how hot they get.
Any tips on this?
Thanks:smash:

I use computer CPU heatsink(pentium 4) with LM3886 or LM3875 and even Lm1875. I bought(*) the heatsink at computer swap meet for US$2.00 each, cheap and effective and compact, if you like you can activate the fan. I use the non insulated version and bolt directly to the heatsink with artic silver 5 cooling paste with no insulation which I prefer.

With the 1st 2 chipamps and powered from +/- 34Vdc, the chipamps are all very cool, <36deg centigrade(I measured not from chart), after hours of listening. If you drive them hard you may see an increased in temp., if you like activate the fan powered lower than 12vdc, eg 9Vdc so the speed is reduced also the noise level, the chipamp runs happily. Don't worry too much.

for LM1875 max supply voltage is +/- 30 volt from spec sheet.

I have at least 8 units and can make many projects.
 
IIRC, someone (Kurt Chang?) on one of the other forums, built an 1875 chipamp using Radio Shack transformers, etc., and laid it out on a board with a thin guage aluminum "L" section from the hardware store for a heat sink. Evidently it stayed cool enough as to not present any promblems at all.

Best Regards,
TerryO
 
TerryO said:
IIRC, someone (Kurt Chang?) on one of the other forums, built an 1875 chipamp using Radio Shack transformers, etc., and laid it out on a board with a thin guage aluminum "L" section from the hardware store for a heat sink. Evidently it stayed cool enough as to not present any promblems at all.

Best Regards,
TerryO


a simple heatsink is ineffective if
1. LM is powered from high DC supply like 25V or higher.

2. if the speaker impedance is low, you are pushing the chipamp hard.

both above will increase chipamp temp. A simple heatsink is therefore ineffective.

The cpu heatsink is effective(with fins)in dissipating heat, cheap and compact. I use the fan when I drive low impedance woofers(2 in parallel) eg 4 ohm.
 
ttan98 said:



a simple heatsink is ineffective if
1. LM is powered from high DC supply like 25V or higher.

2. if the speaker impedance is low, you are pushing the chipamp hard.

both above will increase chipamp temp. A simple heatsink is therefore ineffective.

The cpu heatsink is effective(with fins)in dissipating heat, cheap and compact. I use the fan when I drive low impedance woofers(2 in parallel) eg 4 ohm.


I agree, although I believe that Kurt was driving HE horns and that while he used the chipamp for all kinds of uses, his preference was bi or triamping. I suppose that if the conditions you discribe are the intended usage, then perhaps a 3875 or 3886 might be a better choice.

Best Regards,
TerryO
 
split into two sections of fins, each half is about 1x 1.3 x 2.5".

hi there,
thats will not suficient.
pls check my last thread, i was using 2 of pentium heatsink for my stereo lm1875 ( the fail portable project ).
its a hot kind of chips, you must get suspicious if the heatsink stay cool.
the heatsink is coated, scratch by smooth sandstone paper, have some good paste, and a real thin micas, then you will find more actual heat.
the portable amp still running well till today ( with fully casing ), but i limit the pot to use 25K only, thats good and loud enogh to drive ineficient 4' or lighter 6,5'. but i use 3 times biger heatsink + thermostated blower fan for my biger project ( stereo pair, 10A, 34VDC, for 12' threeway ).
i don't know hows the heat if you use the full timer blowing fan, the sound is annoying, but i will try the ttan98 idea ( 9V for fan ).

i think the lm1875 need an ideal heat to perform the best sound.......... or just my ears adaptation??
i dont know, hope someone capable to explain this.

Brgds.
Eka
 
eketehe said:



i don't know hows the heat if you use the full timer blowing fan, the sound is annoying, but i will try the ttan98 idea ( 9V for fan ).

i think the lm1875 need an ideal heat to perform the best sound.......... or just my ears adaptation??
i dont know, hope someone capable to explain this.

Brgds.
Eka

"the sound is annoying, but i will try the ttan98 idea ( 9V for fan ). "

At 9volt may be still loud for you, try a few volts lower, say 6volts, fan with lower speed and less noise is much better than no fan at all. If you feel like it design a sensor, to detect when a temp exceeds a specific temp, then turn the fan on and stops when it drops below a specified temp. A few designs on the net you can view and easy to build.

"i think the lm1875 need an ideal heat to perform the best sound.......... or just my ears adaptation?? "

This maybe true to some extend, what is the ideal temp., most probably when the amplifier temp stabilises and stop rising, and warm to touch and not hot to touch. The ideal temp, I don't know maybe there is none.
 
Ideal temp. . . no idea. But, I do have some input on voltages.

From collected empiric data:
The chip isn't appreciated as much on 20vdc rails and nearby values.

Its more appreciated from 11 to 16vdc or (skip 18, 19, 20, 21,* 22, 23, 24) and its also appreciated at from 25 to 32**vdc.

*I highlighted the worst in bold. As it is the mathmatical average of averages, then its probably quite measurable.

**Operation at or above 30vdc is restricted to 16 ohm speaker applications or else thermal runaway may occur.

An especially quiet lineup of fans is made by CoolerMaster. This can be run from a simple power resistor.

An inexpensive bimetallic spring thermostat can be found at the hardware store, in the form of an attic fan controller.

An outdated bimetallic spring type air-conditioning thermostat can be inserted between the power resistor and the fan.
 
Retired diyAudio Moderator
Joined 2002
I helped my friend build our first LM1875 amp last night, and we used small heatsinks for to-220 devices. A single one was getting a bit warm, but putting 3 together seemed to suffice until we can get a real chassis design. We were able to turn it up pretty loud with the small heatsinks.

--
Brian
 
Hi,

I am building my first amp based on LM3875. I used a perf board for the layout and the amp seems to perform decent when tested. Since I am trying to fit the amp in a square 5" x 5" x 5" box made out of acrylic, I do not have room for large heatsinks, so I took 2.5" x 1.5" heatsinks and put one thin cpu fan in front of them (12V, 0.3A), running from a 9V battery. I now tried to attach it to the amp rectifier board without success. As Peter suggested in a separate thread, I used a linear pot to drop the voltage from 27.1V DC from the rectifier board down to 12V and than substituted the pot with a 1M resistor. Now when I connect fan wires through the resistor, the fan doesn't move (even though Multimeter reads around 12V), yet when i touch these cables against the rectifier board (for a fraction of a second) the fan starts to work (extremely fast). Would you know, why I am not able to get the fan to work?

These are some sample picture of my project:
http://profile.imageshack.us/user/vladp/images/detail/#141/dsc4321aa0.jpg
http://profile.imageshack.us/user/vladp/images/detail/#299/dsc4324rm9.jpg
http://profile.imageshack.us/user/vladp/images/detail/#299/dsc4323qe1.jpg
http://profile.imageshack.us/user/vladp/images/detail/#299/dsc4322br0.jpg

Thanks for your help!
 
It is not enough to make a resistor divider from 27 Volt downto 12 Volt.
If this divider has got too high resistance values, then not enough current can flow.
1 Mohm sounds like a very high value to use.


27 Volt downto 12V. difference is 15 Volt
You should be able to use one <=15k resistor and one 12k resistor
or something like that.
Across one is 15 Volt and the other 12 Volt
 
sharp fins said:
As Peter suggested in a separate thread, I used a linear pot to drop the voltage from 27.1V DC from the rectifier board down to 12V and than substituted the pot with a 1M resistor.
That must have been a misunderstanding. What use is it to adjust first with a pot and then whatever the outcome replace the pot with a 1 M resistor?

Use Ohm's law. You have 27 V and need 12 V. So you must convert 15 V times the current into heat. The current is given with 0,3 A, so you need a resistor with 15 V / 0,3 A = 50 Ohm. Practical value will be 47 Ohm. In that resistor you will turn ~15 V * 0,3 A = 4,5 W into heat. Use at least a 10 W resistor for that, or significantly bigger, if you put it in the amplifier case.

Fans need more current to start than to keep running. More current means more voltage drop across the resistor -> less current for the fan. Could be that the fan would run perfectly once it is started, but that it will no start on its own. That is why fan regulators often come with a start-up boost, and that is also, why you cannot get your fan to work other than touching the rectifier board (= 27 V).
 
Thank you, lineup! Thank you, pacificblue!

Well, the use of the pot made sense to me, because you can adjust it up and down until you get the fan speed you are happy with, measure the resistance at that state and put a static resistor based on the value measured.

It was silly of me not to realize that I was dropping the current too.. I will test all of the above.

Thanks again!
 
Hi Sharp,
use your pot to measure the voltage at the fan motor when the speed/noise seems about right.
It will probably be quite a bit less than 12V.
Now you can apply Pacific'c formula (ohm's law) to find the series resistor that will drop sufficient voltage to run your fan motor at slow speed.
 
Hey

I've got a two lm3875's on Peter's boards attached to a single 10cm by 8cm by 1.5cm (ish) chunk of aluminum, which is in turn coupled to the small aluminum enclosure using thermal goo and some screws.

The amp gets warm and the chips are definitely hot to the touch after running...

I freely admit to "eyeballing" the heatsink value... I'm not going to kill my amp am I? (It sounds excellent and performs over extended periods w/o issue)

Here's a snap:
An externally hosted image should be here but it was not working when we last tested it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.