Thanks for clarifying. I had my definition wrong.
at 3A bias for 4 pairs of devices at 50V rails my heat dissipation is :
3A/4pair = 0.75A per device
50*0.75 = 37.5W or 75W per sink (with 2 devices per sink)
I think these sinks at 10" long are 0.24degC/W so only a 18C rise.
Looks like I'm good.
Also looks like perhaps having the diodes on it may also be ok.
My voltage drop across 0.5ohm to the diode will be :
V=IR => 0.75(0.5) = 0.375V
Which if I remember correctly, NP in the F5T article says that one should keep the voltage drop around 0.3-0.4V which is just before the diodes start conducting.
Just to circle back, with this configuration at 3A bias I should get
Class A to a peak of 6A at 8 ohm load and I^2R=P
I should be running 6^2*8=288Wpeak or 144W average Class A at 8ohm.
Last edited:
Hi Lo_Tse, in addition to Sangram’s replay, I want to add:
I found the diodes’ area of operation, between not conducting and hard on (5A++), relatively small, and highly temperature dependent. Therefore I added them on the main heat sink, which temperature can be controlled accurately with the bias current.
I would also not trust and base my design on the MUR3020 datasheet I/V graph. This is a typical performance graph. Some companies publish very typical graphs (meaning not typical at all, but rather highly optimistic…). If you have the possibility, characterize your diodes at different temperatures, with and without the parallel resistor.
-- xLoff
The duty cycle is 50%.
The average is roughly half the peak.
Combine those two and the worst case becomes 2W per diode over the whole cycle.
The peak value lasts for how long? Is it a 10us transient that repeats 100times every 400us? or a 0.1ms transient that repeats 15times every 0.55ms? or ...
How often does the amplifier actually deliver those maximum transients? Once every 3minutes, or once a night?
These are not PA amplifiers run at near 100% all the time.
They are domestic listening duty where average power is likely to be <<10% of maximum capability.
The diodes should need only tiny heatsinks, or you have chosen the wrong amplifier.
Actually its the 10.080" profile, and 10" long. So lots of cooling
Okei. Then you'r all good
the 0.24C/W applies at the manufacturer's deltaT.
The deltaT used by most manufacturers is from 70C to 80C (not degrees C).
At a predicted deltaT of 18C the sink will not achieve 0.24C/W.
You need to find the manufacturer's data for de-rating the sink for other values of deltaT.
Expect the De-rating Factor (DF) to exceed 1.3 and maybe as high as 1.6 for an 18C/W difference between Ta and Ts.
Actual temperature rise will be = 75 * 0.24 * DF = ~27C (for DF=1.5) (note that this DF formula only works when used in reverse, i.e. use the actual sink temperature difference to find the DF and then use the formula to determine the dissipation capability in watts.)
giving:
Ta = 28°C
Ts~55°C
Tc~73°C
Last edited:
Andrew. this sinks can do the job. i use this sinks (only 6"long), they are at 20C above ambient with 45W dissipation. and then only about half the sinks are effektive do to distance from output Devices to edges.
this sinks are 0.80C/W pr 3inches according to heatsinkUSA. but in reality. they do much better.
this sinks are 0.80C/W pr 3inches according to heatsinkUSA. but in reality. they do much better.
Last edited:
well. actualy its not far from reality. it is however hopeless to modell anything from heatsinkUSA's data.
according to heatsinkUSA's data, my sinks should be 0.6C/W (0.8C/W pr 3", mine is 6". 25% more effektive)
but in reality as it is now (only about half the sink are working effectively) 20c/45W=0.44C/W.
my best guess is that this sinks are about 0.3C/W at 25c ambient.
according to heatsinkUSA's data, my sinks should be 0.6C/W (0.8C/W pr 3", mine is 6". 25% more effektive)
but in reality as it is now (only about half the sink are working effectively) 20c/45W=0.44C/W.
my best guess is that this sinks are about 0.3C/W at 25c ambient.
Last edited:
As I mentioned before, the main sink is at a temperature at least 20-25oC higher than ambient already, why attached the diode to it. I thought the whole idea is to keep the diode cool??
Regards,
the room ambient temperature is meaningless in this senario.
You can not mount the diodes outside the amp chassis anyhow. The chassis ambient is another story. And that ambient is much closer to the sink temperature.
and the important thing is not to keep them cool. but keep them at a even temperature.
Last edited:
Hmm in the F5 that I built I found that the heatsink temp is generally 5-8C higher than the ambient interior case temperature. And that is with the transformer inside the case generating its own heat.
For this build the transformer and most of the caps are in a separate case, so I suspect my ambient interior temperature to be perhaps lower.
So from what AndrewT says, that the diodes would probably only need 2W dissipation for the whole lifecycle in this application, perhaps then the little heatsinks with a 6.3C/W rating may be enough? And perhaps better than mounting it on the main sinks if the inside ambient is say 5-10C lower than the main heatsink temp?
Not true, the room ambient can have much effect on the chasis ambient. Imagine placing the amp in a very cold room, heat loss thru the sink surface beccomes fast such that the inside temp will drop too. That's the whole idea of placing heatsink outside of the chasis.
My other experiences taught me that most electronic components (solid state ones) like low operating temperature and hate high temp - they worked harder and aged faster (e.g. caps). Aging faster meaning shorter life time. I prefer keeping them as cold as possible.
One do not have to worry about whether the ambient temperature of the diode is even (I suspect you mean the temp remain relatively stable) or not , as long as the diode did not turn on, right? To turn on the diode, a threshold has to be reached and thermally speaking, the threshold is relatively high based on the data sheet. By the way, once the amp is warm up under working conditions, it should be at thermal equilibrium, both inside and outside - the temperature should be relatively stable at this point and no worry of uneven temp.
I still believe the inside chasis temperature should be substantially lower than the sink temeparture. If not, the sink is insufficient for the application. The sink should cool the output MOSFET and prevent the inside temperature of the amp become too hot. I am sure someone remembered the first Class A amp sold by Musical Fidelity - the A1 I think. It pumped out barely 30-40 watts class A (cannot rememberd exactly) and sounded great. The amp used the whole top cover (and the chasis) of the amp as the heat sink. Unfortuantely, the sink was too small and got so hot that one can literally cook a egg on it. It got so hot that you got burn if you touched it for 2-3 seconds. As expecetd, the amp typically live a very short life.
Regards,
in that senario, the sink temperature drops accordingly too. it will stil be 5-8c difference between sinks and chassis temperature. and stil 25c between room temperature and sink temerature.
what do you think will happen if one N-ch diode starts conducting, while the others don't?
remember, the sinks radiates heat both inside and outside of the chassis. and transformers can get 15-20c above ambient ad a closed chassis with som wents, and you got a hot invironment. very very far from room ambient.
what do you think will happen if one N-ch diode starts conducting, while the others don't?
remember, the sinks radiates heat both inside and outside of the chassis. and transformers can get 15-20c above ambient ad a closed chassis with som wents, and you got a hot invironment. very very far from room ambient.
Last edited: