Choosing a power transformer

[jacco vermeulen]

Why times 1.34 for 30deg C ?

The 30C is taken from your calculation example.
55C heatsink operating temperature, 25C ambient.
Post #62

(To get an accurate number, one has to iterate a couple of times. The found Rth number leads to a different delta T, which in turn leads to a differing Rth. It converges to a stable number within a couple of repeated calculations)

 

[alibear]

Alan0354, it appears that you find the calculations confusing ( that is not a criticism ). Why not do as Jacco suggests and fit a power resistor, or resistors, on to the heat sink, supply with current that you require and actually see or should I say feel the results. Leave running for an hour or so to reach thermal equilibrium
Alan

 

[infinia]

I found if you can get any increase in air flow pays huge dividends in cooling over adding more metal. still air is bad.
heck just painting flat black gives 5-10% over shiny in still air.
The bigger the delta T the more efficiency the heat sink is working. creates its own air movement I reckon> chimney effect and all that science stuff.

the best cooling solution ive seen is crown amps changing their topology to eliminate the OPT insulator thereby running the heat sinks closer to die temps +hotter heat sink = more eff.

 

[Alan0354]

Alan0354, it appears that you find the calculations confusing ( that is not a criticism ). Why not do as Jacco suggests and fit a power resistor, or resistors, on to the heat sink, supply with current that you require and actually see or should I say feel the results. Leave running for an hour or so to reach thermal equilibrium
Alan

Thanks

Ha ha, I have been shopping for a chassis for a long time. The chassis is about the best and biggest I can find on ebay. It is the second most expensive at $280 total. that's all I can get, it has the second biggest heatsink already. it's more like what you see is what you get. This is it unless I make my own.

That said, I think the most efficient way is for me to put on the amp. Measure(feel) the heat, adjust the idle current so I can touch the fins.

Thanks

 

[Alan0354]

The 30C is taken from your calculation example.
55C heatsink operating temperature, 25C ambient.
Post #62

(To get an accurate number, one has to iterate a couple of times. The found Rth number leads to a different delta T, which in turn leads to a differing Rth. It converges to a stable number within a couple of repeated calculations)

Thanks for your explanation, as the last post, this chassis is close to having the biggest heatsink I can find short of make one myself. so it's like take it or leave it. At this point, the best way is to build it, observe the heat and adjust the bias to get the acceptable range. All the learning and calculation pretty much boil down to this.

Thanks

 

[Alan0354]

So, after a big detour on the heatsink, back to the transformer.

So should I get the 22-22V, 20-20V or 18-18V 500VA transformer. It's the choice with highest Class AB power with lowest Class A power to Highest Class A power but not much head room.

It's the matter now buying the transformer and make the best out of the chassis.

I still have a 30-30 625VA sitting here. But I think that's too high voltage. I shouldn't even have bought it.

 

[alibear]

Alan0354 , totally confused now, seems your back to square one. You must make up your mind what you need / want. Either a low power class A or a higher power A/B. Your choice Simples
Alan

 

[Alan0354]

Yep, that's pretty much it. I can't decide. 18-18 should be plenty for my living room with the JMLab. But what if I buy a pair of static electric speaker like Martin Logan or Magnapan? they are known to be inefficient.

With 22-22 I can only use 200mA each pair, only 8W Class A. But I get 72W Class AB.

 

[AndrewT]

Since you already have the 30-0-30Vac transformer and most of what you need to asseble the amplifier, I suggest you build a low bias ClassAB version with relatively cold heatsink and TEST it thoroughly.

Then increase the bias in safe steps and retest.

When you reach a comfortable limiting device temperature, stop testing and listen for a while.
Then decide whether you need to buy a new transformer.

 

[jacco vermeulen]

I can't decide.

You have a great looking amp case.

For class A use, it would be sub-standard though.
The 28-rib heatsinks look something in the order of 11''-12'' wide, with a length of 6'', but the height of the ribs is on the short side to be really efficient.
Best you can expect for class A use is 0.45 C/W , 65W dissipation at a heatsink temperature which is still acceptable.
(provided you do not reside in Alaska, I'd still not underestimate a constant idle heat dissipation level of 130W)

For class AB, you have a perfect amp case.
And you already have a 30Vac/625VA transformer for a class AB power amplifier, capable of driving next to every loudspeaker.

Your 5-pair output stage, biased at 25mV across a 0R22 emitter resistor, would have a quiescent current level of ~0.57A.
In a 4 ohm impedance, that's around 2.5W continuous class A output, good for 94dB sound pressure on your Jacques Mahul loudspeakers.
In your living room, that's plenty.

With rails at ~40Vdc and the output stage biased to 0.57A quiescent current, your heatsinks would be at an acceptable ~45C (25C ambient)

Another merit of multi-parallel output stages : All the advantages of class AB, plus a lot of class A output.
(0R15 emitter resistors require a whole lot of effort, 0R1 is close to a dead short and asking for trouble. 0R22 is common sense)

 

[Alan0354]

Thanks Andrew, Alibear and Jacco for all the help. I think I am going to buy the 22-22. This will give me a little buffer in Class A region than the 30-30 I have. I am glad I raise this heat issue and never thought it's that complicated. So I am going to do the calculation so I would not pass 63W for each side and make the best. I don't mind buying the extra transformer. I sunk in a lot of money, I want the perfect amp.

Thanks