Wondering if I can get anymore bias current without affecting long term reliability and stability. About 1.2A bias current for this class A amp using SC5200's. Heatsink is is extremely hot, can only stand touching it for about 5 seconds, but I have be playing music for about a hour straight and the amp hasn't died.
Do I need to stop at 1.2A or is there a way to tell if I can dial up the bias current more without killing the output stage?
Do I need to stop at 1.2A or is there a way to tell if I can dial up the bias current more without killing the output stage?
Attachments
Nice.
How high does your scope go? With long untwisted wires, and lots of enclosed loop areas, and no bypass or decoupling caps, maybe the high temps are partly due to high-frequency ringing and/or oscillation.
Anyway, I would not go much higher without better cooling. (You could at least stand the heatsink up, to get some more help from convection.)
How high does your scope go? With long untwisted wires, and lots of enclosed loop areas, and no bypass or decoupling caps, maybe the high temps are partly due to high-frequency ringing and/or oscillation.
Anyway, I would not go much higher without better cooling. (You could at least stand the heatsink up, to get some more help from convection.)
I'd say that touching a heatsink for 5 seconds on an amp thats been running for 60 minutes is OK. It's class A, its reached thermal equilibriam and there should be no transients to push dissipation in the junctions higher.
Remember worst case for pure class A is no signal. The louder you turn it up the cooler it runs
Remember worst case for pure class A is no signal. The louder you turn it up the cooler it runs
Hi,
the standards and norms may give a hint.
For audio-video the DIN EN6065 may apply.
It sets for a maximum surrounding temperature Ta of 35°C and 45°C for tropical climate.
max temperature rise Tc in K above Ta for touchable parts:
Knobs, Handles etc, metallic: +30K (normal use) +65K (defunct)
Knobs, Handles etc, nonmetallic: +50K (normal use) +65K (defunct)
casing, metallic: +40K (normal use) +65K (defunct)
casing, nonmetallic: +60K (normal use) +65K (defunct)
values of TC need to be reduced by 10k for tropical climate.
For nontouchable parts basically Ta depends on the specific parameters of the part.
I would´nt want any touchable part above 60°C.
jauu
Calvin
the standards and norms may give a hint.
For audio-video the DIN EN6065 may apply.
It sets for a maximum surrounding temperature Ta of 35°C and 45°C for tropical climate.
max temperature rise Tc in K above Ta for touchable parts:
Knobs, Handles etc, metallic: +30K (normal use) +65K (defunct)
Knobs, Handles etc, nonmetallic: +50K (normal use) +65K (defunct)
casing, metallic: +40K (normal use) +65K (defunct)
casing, nonmetallic: +60K (normal use) +65K (defunct)
values of TC need to be reduced by 10k for tropical climate.
For nontouchable parts basically Ta depends on the specific parameters of the part.
I would´nt want any touchable part above 60°C.
jauu
Calvin
Heating my Livingroom
I have 5x Yamaha MX-1000 amps. They are Class A & 330 RMS @ Low distortion.
When I lived in Germany I used them to Heat my livingroom in the winter. The heatsinks were hot enough, that it was UNcomfortable to keep your hand touching them. The Heatsinks wouldn't actually burn you, but Almost.
Back to your Ques. You can push your heatsinks UNTIL, the sound degrades To YOUR Ears.
OR get Better Cooling for those Heat Heatsinks.
MLStrand56
I have 5x Yamaha MX-1000 amps. They are Class A & 330 RMS @ Low distortion.
When I lived in Germany I used them to Heat my livingroom in the winter. The heatsinks were hot enough, that it was UNcomfortable to keep your hand touching them. The Heatsinks wouldn't actually burn you, but Almost.
Back to your Ques. You can push your heatsinks UNTIL, the sound degrades To YOUR Ears.
OR get Better Cooling for those Heat Heatsinks.
MLStrand56
Last edited:
I've heard of a 5 second rule, that if you can't keep your fingers on it for 5 seconds it's too hot. But this is DIY. Wouldn't you like a greater helping of reliability than you need for a 90 day warranty? Make the sink big enough to hold onto indefinitely, under any conditions.
Actually, if you were testing that sink flat on the bench and it barely passes the 5 second test, it will probably be OK with the sink and fins mounted vertically in free air.
What really matters is case temperature. I usually do the finger test on the transistor plastic case, right over the die.
Actually, if you were testing that sink flat on the bench and it barely passes the 5 second test, it will probably be OK with the sink and fins mounted vertically in free air.
What really matters is case temperature. I usually do the finger test on the transistor plastic case, right over the die.
Last edited:
Calculate die temperature.
If die temperature is kept under 87.5C, long term reliability is in the clear for the C5200.
(87.5C => (150-25)/2 + 25C ambient)
The hot water out of your home taps is generally set to about 60 Degrees (Not always). You can hold your hand under the hottest tap water for about 5 secs.
The same rule is applied to using the human body to measure the heatsink temp. If you can hold on for 5-10 seconds the temperature is likely to be about 60-65 Degrees. This is considered good for long term reliability.
Tj (the die temperature) can be considerably more than this. Tcs (the case temperature marginally more). Most guys would aim to keep Tcs below about 70 Degrees - (80 Degrees is OK but long term reliabilty then comes into play).
The same rule is applied to using the human body to measure the heatsink temp. If you can hold on for 5-10 seconds the temperature is likely to be about 60-65 Degrees. This is considered good for long term reliability.
Tj (the die temperature) can be considerably more than this. Tcs (the case temperature marginally more). Most guys would aim to keep Tcs below about 70 Degrees - (80 Degrees is OK but long term reliabilty then comes into play).
Last edited:
Linear Audio Vol 3, "Design Considerations for a Class A Amplifier Enclosure"
http://www.linearaudio.net/images/stories/Didden LA V3 PK lr.pdf
Free download.
Patrick
http://www.linearaudio.net/images/stories/Didden LA V3 PK lr.pdf
Free download.
Patrick
60° hot heatsink is a sure way to degrade lifetime for ALL components in an Amp.
Regardless wich class.
I would never let the heatsink be the item to save money on when building a Amo.
If there is too little space for a good enough heatsink, the enclosure is too small.
I am avoiding using fans unless I am building amp for PA/stage-use.
Also I try not to use the enclosures other metal area as a part of the heatsink, especially in Class A-Amps. General component degrade is the reason.
Keeping enclosure in a pretty low temp is only obtained by designing the heatsink big enough to do the job all the time. I always try to keep any touchable surface on a temperature lower than 50°C
Regardless wich class.
I would never let the heatsink be the item to save money on when building a Amo.
If there is too little space for a good enough heatsink, the enclosure is too small.
I am avoiding using fans unless I am building amp for PA/stage-use.
Also I try not to use the enclosures other metal area as a part of the heatsink, especially in Class A-Amps. General component degrade is the reason.
Keeping enclosure in a pretty low temp is only obtained by designing the heatsink big enough to do the job all the time. I always try to keep any touchable surface on a temperature lower than 50°C
> 60° hot heatsink is a sure way to degrade lifetime for ALL components in an Amp.
> Regardless wich class.
That I do not agree, and I have explained that in detail in the article.
Above all if you just look at heat sink temperature, and ignore the thermal path all the way to the semiconductor substrate, you may well come to the wrong conclusions.
It is in the end the junction temperature that is of critical importance, and the Toshiba reliability handbook has illustrated this clearly and scientifically.
I do agree that lower operating temperatures is always beneficial.
But the benefits have diminishing returns at some point.
So my personal design guideline is still <=100°C junction temperature.
And this is much more meaningful than defining heat sink temperature.
You can start discussing where the heat sink temperature should be measured ....
But that is only my personal opinion,
Patrick
> Regardless wich class.
That I do not agree, and I have explained that in detail in the article.
Above all if you just look at heat sink temperature, and ignore the thermal path all the way to the semiconductor substrate, you may well come to the wrong conclusions.
It is in the end the junction temperature that is of critical importance, and the Toshiba reliability handbook has illustrated this clearly and scientifically.
I do agree that lower operating temperatures is always beneficial.
But the benefits have diminishing returns at some point.
So my personal design guideline is still <=100°C junction temperature.
And this is much more meaningful than defining heat sink temperature.
You can start discussing where the heat sink temperature should be measured ....
But that is only my personal opinion,
Patrick
Last edited:
Hi,
How hot you can run a heatsink is not just to do with heatsink temperature.
How hot you should run an external heatsink is to do with safety and burns.
What really matters is the difference between device die temperature and
the heatsink temperature, the former is always more, its just like ohms law,
the temperature difference (voltage) creates the thermal flow (current)
through the devices and any insulation used total thermal resistance.
i.e. For 4 output devices you can run the heatsink hotter than for 2, as
the thermal resistances are halved, so are the temperature differences.
rgds, sreten.
Doubling output devices for class A when they are basically not
needed in terms of power handling, is a very good idea thermally.
How hot you can run a heatsink is not just to do with heatsink temperature.
How hot you should run an external heatsink is to do with safety and burns.
What really matters is the difference between device die temperature and
the heatsink temperature, the former is always more, its just like ohms law,
the temperature difference (voltage) creates the thermal flow (current)
through the devices and any insulation used total thermal resistance.
i.e. For 4 output devices you can run the heatsink hotter than for 2, as
the thermal resistances are halved, so are the temperature differences.
rgds, sreten.
Doubling output devices for class A when they are basically not
needed in terms of power handling, is a very good idea thermally.
Last edited:
The device datasheets will usually quote thermal resistance numbers for various mounting methods, from which it is trivially to calculate how far above heatsink temperature the die is (Don't forget to add the resistance of any insulating washer).
For a non accessable heatsink, how hot it gets is mostly irellevant, it is die temerature that matters, and data on that will also be in the datasheet.
I tend to feel that touch tests are a very poor way to figure this as if the thermal resistance from the die is very low (A big device on a copper heat spreader as the RF power guys do, say) then a heatsink hot enough to cause second degree burns is not necassarily a problem, where a device with a poor thermal path may run too hot even with the heatsink only slightly above ambient.
Do the math, you might be surprised by what modern sand will run at with good reliability, electromigration is not the MBTF killer at high temperature that it once was.
And yes, more devices will allow the heatsink to run hotter for a given die temperature, which given the price of heatsinks Vs the price of sand can argue for quite extensive parallelism if it means the total heatsink area can be smaller.
Of course if the heatsink is external then touch temperatures come into it, but for internal heatsinks do the math and don't be afraid to run hot (Just not too close to the main capacitors).
73 Dan.
For a non accessable heatsink, how hot it gets is mostly irellevant, it is die temerature that matters, and data on that will also be in the datasheet.
I tend to feel that touch tests are a very poor way to figure this as if the thermal resistance from the die is very low (A big device on a copper heat spreader as the RF power guys do, say) then a heatsink hot enough to cause second degree burns is not necassarily a problem, where a device with a poor thermal path may run too hot even with the heatsink only slightly above ambient.
Do the math, you might be surprised by what modern sand will run at with good reliability, electromigration is not the MBTF killer at high temperature that it once was.
And yes, more devices will allow the heatsink to run hotter for a given die temperature, which given the price of heatsinks Vs the price of sand can argue for quite extensive parallelism if it means the total heatsink area can be smaller.
Of course if the heatsink is external then touch temperatures come into it, but for internal heatsinks do the math and don't be afraid to run hot (Just not too close to the main capacitors).
73 Dan.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.