A question, from ignorance.

[Poindexter]

What is the 'normal' operating temperature for one of these amps? I have a pair of LM1875, power supplies +-25v, on a heatsink that looks like a railroad track (it's actually an overhead door track, about 40mm (base), 25mm('track'), 40mm (height of web) that's about 90mm long, masses about 90gm. After a few minutes operation, signal or no, that heatsink is almost too hot to touch. Output offset is in the 'hood of 10mV, so significant power is not being dissipated there.

Have I simply underestimated the amount of heatsink and thermal transfer involved in one of these wee beasties?

Sounds bloody mahvelous, by the way. I'm diggin' it, if it don't burn the house down. I'm listening to Dori Caymmi now, waiting to see if it's going to go into thermal shutdown. Heatsink is now too hot to touch more than briefly.

Mahalo,

Poinz

[BWRX]

You just need a bigger/better heatsink. You can operate the chips at temperatures too hot to touch (that's probably at least 60C) but it is not good for their longevity. A good rule of thumb is to use heat sinks that at least allow you to keep your hand on them indefinitely during normal operation. Most of us would opt to use heat sinks that only get reasonably warm during normal operation.

[Poindexter]

Thank you, Brian. Testing, it seems everything else is right there; gain is right on, hum+noise is down below a tenth of a millivolt, bandwidth is ridiculous (can't measure the -1dB point at the bottom, my generator only goes down to 7Hz, at the top -1dB at 120KHz), sound is better than I thought a sand amp could be, very smooth even with fresh caps.



I've swiss cheesed the heatsink, since it lies across the available airflow, and just ordered a lower voltage trans from Digi-Key, since the PS is a little high at ±26V at the chip pins.

Report soon,

Poinz

[Nordic]

Lovely woodwork, what did you use as "varnish"?

[woolly]

If DC output is not the cause, it could be high freq oscillation. Current may be low at HF but power produced (sound and heat) is proportional to frequency so will be high at HF for even low current.
Have i understood correctly that you have up to 120KHz? Try restricting high limit to say 30KHz. A small value cap (~0.1uF) from + to - will do the job. This will cancel any super high frequency. I have done this and high freq audio is not affected.

My heat sink is not huge and only gets warm when hammering out loud music for long time. A bigger sink is always good, so try that.

Add a fan? Something i am considering.

[BWRX]

That enclosure is absolutely stunning! I love how the classy wood frame is accented by the industrial aluminum tops, bottoms, and knob. Excellent craftsmanship. It's like a modernized gaincard enclosure the way you have two separate halves with continuous front and back panels.

Got any more pics?

[Poindexter]

Thanx, gents. The chassis is Jatoba (sometimes called by the dealers 'Brazillian Cherry'; dunno why, it's nothing like Cherry), and is finished with polyurethane varnish, about 8 coats (the Jatoba is very large pored, takes a lot of coats to fill up). I am a forty year guitarmaker, jeweler, airplane maker, so I have a lot of time in the woodshed; but I am very much gratified by your appreciation. The varnish, by the way, is Minwax aerosol, my fave for small stuff an aerosol can do. It's light-bodied, very clear (very little ambering), and dries pretty fast and very hard. It's buffable with the nice compound I get from the auto finish guy.

However, I just fired up to test with music, and although the sound is if anything a little better, it's still heating up like crazy. Fifteen minutes, I can't comfortably hold my finger on the sink.

On the possibility, I'm going to test for oscillation, thank the gent above very much. I can test out to the single digits of MHz at least, and see if there's any friskiness. Anybody cured this behavior before?

Onward,

Poinz

[BWRX]

You could potentially use the bottom or top slab of aluminum as the heatsink for each chip. It's hard to say if it would be much better though, as I can't really visualize what your current "overhead door track" heat sinks look like. A picture is worth a 1000 words

[Poindexter]

Okay; at present, all chassis plates are at safety (green wire out of the wall) ground. I am loathe to change this, since with this chip, the plate is at pin 3, which is Vee, or -26 volts. Scary, sorta.



I only had time to take one shot for you, so it's not huge and byootiful and stuff. You will percieve that the heatsink is not utterly wimpy (the length dimension, from left to right, is 3 1/2"). I have seen builds in the photo section that used smaller. The caps you see are B G 100uF 50V, for scale. Orange is PS+, blue is PS-, brown is circuit ground, green is safety (earth) ground, yellow is +speaker out, blue/white and orange/white are signal in.

Tomorrow morning, when I have a glass (or two) less wine in me gullet, I'm going to just hook the thing up running (with 8R resistors on the posts, thankew) to my lovely Millivac AC millivoltmeter, and just see if there's any signal on the output with the input grounded. If I don't get joy, it's orrf to NS support, and whimper a bit.

I don't get to build anything straightaway, these days. Vishnu is sticking his thumb in my ribs, just so I know I ain't cool.

I ain't cool, but,

Thanks anyway,

Poinz

[AndrewT]

Hi,
that aluminium channel cannot cool one chipamp and you look as though you have two chips on it.

You need a sink that can dissipate at least 4times as much heat.
It also needs to be oriented to allow the airflow to pass over the fins.
You box is so nice I suspect you're not interested in a vertical sink so consider whether a big, slow fan (>=100mm diameter) could do your cooling for you.

But if passive cooled or fan cooled, there must be a way for the cold air to get in and a route for the hot air to get out.