[precsmo]

Hi to all gurus, I'm using the left, right and ground channels with AD843JN in my Pimeta, it seems works fine and it sounds great! But after 2 to 3 minutes, these opamps tend to get really hot! please advice what to do. Here are my settings:

Thread (test and seems works fine):
-Tuned to 28V (to Pimeta +/-14V), I've tried lower it down to 24V, but no difference

Pimeta:
-left, right and ground channels all with AD843

-gain set to default, Gain 11, as in TangentSoft's diagram where:
-R3 is 10k, R5 is 3.32K, R6 is 1M and R4 is 10k

-all 4 power rails capacitors of 1200µF 35V, I ordered it wrongly instead the 1000µF 35V (should I change it back to 1000µF as recommended default maximum in TangentSoft)?

-all buffers are 2 stacks, though the R11 for left and right channel buffers I used were 100Ω, the ground channel was 200Ω as I miss ordered it and will replace it soon, does this affect anything?

-volume control I used 50k ALPS RK27

-the rest I used as recommended default in TangentSoft

My another experience, I find Gain 11 is more than enough for my HD580, should I reduced the gain, is the gain directly affects the opamps from heating up? I really enjoy the sound, but I can't feel safe to run the amp before I can stop the Opamps keep heating up badly, please help! Thanks a thousand in advanced! I tried to search everywhere, but can find much about AD843 settings.

 

[dsavitsk]

These opamps do run hotter than some, so it depends what you mean by "really hot". None of the things you mention (caps, gain, etc.) would contribute to excessive heat. It is possible they are oscillating, though unless they are too hot to touch, you are probably okay.

 

[firefox360]

Well usually a low gain can cause the op-amps to be unstable which makes it oscillate, which in turn makes it heat up. Although with a gain of 11, that doesn't look to be the source of problems for you.

The power capacitors are okay at 1200uF. You can actually go as high as you want, and doing so can actually improve sound quality.

The resistors should be the same on all the buffers. Whether they affect the operations of the op-amps or not, I'm not entirely sure.

The volume control seems alright for the purpose.

One question I have is that when the op-amps do heat up, do you notice any change in sound quality? Massive distortion perhaps? And whats the DC offset on the outputs?

 

[Filburt]

The difference in resistor value does make a difference. You're running your ground channel buffer at lower bandwidth than your L/R channel buffers. You may be getting oscillation.

Gain actually can make op-amps more stable, but I'd personally recommend a lower gain as you can get a lower noise floor and lower distortion. Maybe try something around gain 4 to 6.

 

[mono]

Quote:

Originally Posted by precsmo /img/forum/go_quote.gif Hi to all gurus, I'm using the left, right and ground channels with AD843JN in my Pimeta, it seems works fine and it sounds great! But after 2 to 3 minutes, these opamps tend to get really hot! please advice what to do. Here are my settings:

How hot is "really hot"? They are expected to run pretty warm, I even thought about putting heatsinks on mine because I "thought" after a certain point they sounded worse warmed up, but I found another option mentioned next...

Quote:

Thread (test and seems works fine): -Tuned to 28V (to Pimeta +/-14V), I've tried lower it down to 24V, but no difference

I think it made a difference but the difference was small enough you didn't perceive it. I run AD843 at lower voltage, 20V (+-10V). How much class-A bias current are you using? I forget what the default is on Pimeta, try less.

Quote:

Pimeta: -left, right and ground channels all with AD843 -gain set to default, Gain 11, as in TangentSoft's diagram where: -R3 is 10k, R5 is 3.32K, R6 is 1M and R4 is 10k -all 4 power rails capacitors of 1200µF 35V, I ordered it wrongly instead the 1000µF 35V (should I change it back to 1000µF as recommended default maximum in TangentSoft)?

I think Tangent left the gain suggestion at 11 merely to ensure that if unchanged, it would be loud enough for any practical use. Most people will probably choose slightly less than half that, depending on the cans.

The capacitors are fine, larger may decrease ripple some and have a minimal gain (all else equal) but will also take slightly longer to initially charge and drain, may produce more of a thump at turn on or off.

Quote:

-all buffers are 2 stacks, though the R11 for left and right channel buffers I used were 100Ω, the ground channel was 200Ω as I miss ordered it and will replace it soon, does this affect anything?

It may degrade sound slightly, should be changed, but shouldn't be the cause of the AD843 heating up.

We dont' know how hot though, you might measure current. To make it easier, remove the buffers and just measure with opamps installed, and of course factor a fraction of a mA for the railsplitter and (whatever you'd set it to use) some for the LED.

Quote:

My another experience, I find Gain 11 is more than enough for my HD580, should I reduced the gain, is the gain directly affects the opamps from heating up? I really enjoy the sound, but I can't feel safe to run the amp before I can stop the Opamps keep heating up badly, please help! Thanks a thousand in advanced! I tried to search everywhere, but can find much about AD843 settings.

Yes, reduce the gain to get the most out of the opamp bandwidth and have full travel adjustment with the POT. Since we don't know how loud you want it, the final gain is up to you. I randomly suggest 5 for those cans or lower if you never need it to get very loud.

The opamp is rated to 70C operating and 125C storage temp. I expect the 70C means to stay within it's other spec'd parameters, not "let the magic smoke out, hot". Point being, if you can leave your finger on for more than 5 seconds without pain, I would't worry about it so far as a need to turn the amp off.

 

[tomb]

The AD843's will pull a heckuva lot of quiescent current. So, they may run hotter than many other opamps. (current * voltage has to be rejected somehow). One way to tell if they're oscillating (without a scope) is your bass response. A lack of bass and a general lack of feeling of power is an indication of oscillation.

I also agree with Filburt. While the bandwidth setting resistor on the BUF634 usually has a trivial affect on a circuit, in this case it may be critical. If you have the symptoms described above, that might be the first thing to try - you could always put more 200's on the L and R buffers - it won't affect the sound quality that much until you get the 100 ohm for all three. Or, you could unsolder the stack on the Ground channel. That would make the bandwidths equivalent. The consensus is (including Tangent) that stacking the BUF's on the ground channel is a waste of time.

If you don't have the symptoms in the 1st paragraph, then maybe they're just running hot, period.

IMHO, gain of 11 is about right for an HD580, but then I don't like rotating my volume knob all the way, anyway. Tangent built in a little extra because the crossfeed eats up some gain, and he was taking that possibility into account.

 

[precsmo]

Thank you so much for all the detailed explanations and advices! Everything is so detailed and well explained here I probably don't have to ask more questions. My parts are all coming and will see what will new results by then.

by the way, about hot, I felt it quiet hot, I think I can last my finder for 5 seconds, thought, it sounds good! So, I'm much less worry now and again, thanks for the time writing me the advices, that's how I find DIYing headphone amp so fun!