Really Warm Gainlcone
At the moment my LM3886 chips are mounted on a piece of 3mm 50mm wide Angle Alloy. But the alloy is getting really really warm, to the point where I can't leave my hand on the alloy. Is this normal? Do I need better heatsinking? Or could this be oscilating?
Specs:
2x18vac 300va Transformer,
Carlosfm Snubberized PSU MkII (2.2k -> 10,000uf - > 100nF - > 1R + 100nF).
LM3886 with 1500uF 35v Electro's at psu pins.
Original BrianGT kit values (220Ohm, 22k, 22k, 680ohm) plus 10k mute.
So, any ideas? Sound is great, I would never complain of the quality. And whilst the amp was getting hot its only started getting this hot tonight since doing the better psu instead of just 6x1500uF (3x per rail)
At the moment my LM3886 chips are mounted on a piece of 3mm 50mm wide Angle Alloy. But the alloy is getting really really warm, to the point where I can't leave my hand on the alloy. Is this normal? Do I need better heatsinking? Or could this be oscilating?
Specs:
2x18vac 300va Transformer,
Carlosfm Snubberized PSU MkII (2.2k -> 10,000uf - > 100nF - > 1R + 100nF).
LM3886 with 1500uF 35v Electro's at psu pins.
Original BrianGT kit values (220Ohm, 22k, 22k, 680ohm) plus 10k mute.
So, any ideas? Sound is great, I would never complain of the quality. And whilst the amp was getting hot its only started getting this hot tonight since doing the better psu instead of just 6x1500uF (3x per rail)
Re: Re: Re: Really Warm Gainlcone
On the more difficult/demanding musical transients the PSU voltage sags, with lower capacitance.
As the PSU voltage drops, the chip has less heat to dissipate.
Are you sure you have similar heatsinks?
3mm is very thin, specially without fins.
And 50mm wide is two inches, have you noticed that?
Unless I didn't understand and those 50mm is just the size of the L shaped bar and he is not specifying the width.
pinkmouse said:
On the more difficult/demanding musical transients the PSU voltage sags, with lower capacitance.
As the PSU voltage drops, the chip has less heat to dissipate.
Are you sure you have similar heatsinks?
3mm is very thin, specially without fins.
And 50mm wide is two inches, have you noticed that?
Unless I didn't understand and those 50mm is just the size of the L shaped bar and he is not specifying the width.
Oscillations...
All my clones, single, paralelled & bridged have had ocillation-problems when first tested...
Even the BG kit ones. And all done "by the book".
So, now I always make an output inductor, 10-12 rounds of 0,8mm trafo-wire around an "AAA" battery, paralell with a 8-10 ohm 3W resistor. This have cured all my osc. problems.
Arne K
All my clones, single, paralelled & bridged have had ocillation-problems when first tested...
Even the BG kit ones. And all done "by the book".
So, now I always make an output inductor, 10-12 rounds of 0,8mm trafo-wire around an "AAA" battery, paralell with a 8-10 ohm 3W resistor. This have cured all my osc. problems.
Arne K
rampage101 said:
Without fins on your heatsink the fan will not be very effective...
Anyway, the idea would be like the one I suggested Brian for the led on the LM3886 Snubberized PSU when we were making the layout. Check it out.
A single diode in one of the secondaries of your trafo, a small cap and an 8 or 9V reg (7808/9). Don't run the fan at 12V, it makes too much noise. You can try a resistor instead of the reg, but the fan may have problems in starting up.
Hi rampage101,
I'm surprised they are getting hot. I use 40 x 40 x 3mm al angle all the time with my gainclones and haven't had any heat problems. I've only used LM1875 and LM3875s though.
The picture below shows my test GCs with small lengths of al angle - 180mm. They get warm but not hot.
On my real ones I used larger pieces.
http://www.briangt.com/gallery/nigc-erskine
regards
I'm surprised they are getting hot. I use 40 x 40 x 3mm al angle all the time with my gainclones and haven't had any heat problems. I've only used LM1875 and LM3875s though.
The picture below shows my test GCs with small lengths of al angle - 180mm. They get warm but not hot.
On my real ones I used larger pieces.
http://www.briangt.com/gallery/nigc-erskine
regards
Attachments
When I was playing around with Speaker Workshop, I decided to invert the signal (by swapping the +ve and -ve on the output of the amp) unfortunately I was feeding this back through the pc soundcard, kindof a virtual short, but not quite..... This resulted in a very hot heatsink (and in fact the entire chassis), and I have comparitively a BIG heatsink, it was too hot to touch. This is with only a 1000uF / rail PSU so if something is not quite right then these chips can get mightly hot even withg a big heatsink!
maybe check you haven't got a near short on output.
as someone else asked, is it getting hot when no music is playing?? If it is then it's probably oscilating.
Tony.
maybe check you haven't got a near short on output.
as someone else asked, is it getting hot when no music is playing?? If it is then it's probably oscilating.
Tony.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Still havn't tested it in regards to hot with no input, but its getting damn hot listening at moderate->high levels. Thats what it looks like with my pre on top.
rampage101, it's common amongst many Gainclone builders that they have no clue about the cooling requirements. It often comes as a suprise that you actually need heatsinks.
http://www.national.com/appinfo/audio/files/Overture_Design_Guide13.xls
http://www.national.com/appinfo/audio/files/Using_Overture_Design_Guide.pdf
Those two documents can give you the answer how many degrees/watt you'll need. When you have got an answer you can check out a heatsink manufacturer how such a heatsink would look like. The hard thing is to convert it to what you have or what you plan to build. Essential for cooling is a big surface and a possibility for the air to go upwards if you won't use a fan.
http://www.national.com/appinfo/audio/files/Overture_Design_Guide13.xls
http://www.national.com/appinfo/audio/files/Using_Overture_Design_Guide.pdf
Those two documents can give you the answer how many degrees/watt you'll need. When you have got an answer you can check out a heatsink manufacturer how such a heatsink would look like. The hard thing is to convert it to what you have or what you plan to build. Essential for cooling is a big surface and a possibility for the air to go upwards if you won't use a fan.
peranders said:
Hi Peranders,
I hope your initial comment wasn't referring to me.
As I said in my previous post, the size al angle I use for test rigs has proven adequate for me. They do get warm. In the final versions I use about 4 times as much and it is mounted in a metal case. (overkill in my opinion).Thanks for the links to Nationals design guides I have forgotten about them and it was interesting to revisit them.
We must remember what these chips are designed for. Car radios, portable radio, mini hifi system, small compliances with minimal area or volume for what we DIYers consider proper heatsinks.
A couple more points for rampage101.
I would add a metal base that connect the front and back angles together.
Have you used heat transfer compound?
Are the screws tightened sufficiently?
I have heard platted cat5 speaker cable is capacitive. Could this be a problem.
On a safety issue, is that fuse holder rated at 240VAC? Most of the ones I have seen are rated for 12 or 24 VDC. I would add some heatshrink on the 240VAC solder joints.
regards
Greg
It's good to remember that it's a good thing not to have the IC near it's shut down temperature becasue it will reduce it's lifetime. Meaning the LM3886/3875 should not run burning hot in steady state.
Beppe, if you must transfer lot's of heat it's essential to have heat compound between the IC and the heatsink.
Greg, I was making a general observation about the cooling.
Beppe, if you must transfer lot's of heat it's essential to have heat compound between the IC and the heatsink.
Greg, I was making a general observation about the cooling.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.