Im slowly building a pair of 6 deep crippled bridged f4 in 5u 500mm. Temperature is a real problem, i would go for lower rail voltage and crank the currentImpedance of your loudspeakers is almost flat 5 Ohms, above 100 Hz. I would go for more current.
So, 18 V transformers.
There is also a problem that, even at modest bias and higher voltage, 4U/300 mm deep case will be very hot.
Naaaah. Just put it in the freezer with extention cables through pre-drilled holes.
^ That thermal coefficient was likely calculated at a much higher operating temperature than you'd likely want. I personally de-rate them around 25% for the temps / dissipations we use typically. That's a WAG.
I was looking for an old .pdf I recall Modushop published for recommendations. I can't seem to find it. ZM and many others also have a ton of anecdotal experience and might weigh in with some good rules of thumb. Perhaps they'll chime in.
tl;dr - My guess is that a 4U 400 would be the likely minimum recommendation for 100W dissipation per side without some sort of forced cooling. I think your actual temps would be higher. How much?
I hope the smaller one will work for you, but I'm just being cautious.
I was looking for an old .pdf I recall Modushop published for recommendations. I can't seem to find it. ZM and many others also have a ton of anecdotal experience and might weigh in with some good rules of thumb. Perhaps they'll chime in.
tl;dr - My guess is that a 4U 400 would be the likely minimum recommendation for 100W dissipation per side without some sort of forced cooling. I think your actual temps would be higher. How much?
I hope the smaller one will work for you, but I'm just being cautious.
Your calculation is slightly off.
2x18V will result with +/- 22.5V rails. At 0.4A bias, one channel would dissipate 18W only. As said, more bias equals better sound. You are good up to 1A bias, which results with 45W dissipation per heatsink or with 15° Celsius temp. rise.
^ Agreed if using the proper terms. I think they were using 0A4 meaning the current through each of the source resistors vs. total Iq or bias.
I had to learn that also, so I'm just guessing. Also have the luxury of having a similar conversation in the BA-3 thread (same output stage).
If not, apologies.
I had to learn that also, so I'm just guessing. Also have the luxury of having a similar conversation in the BA-3 thread (same output stage).
If not, apologies.
I think ~0.6A each will get you about 50W class A into an 8 ohm load with +/-32V rails.
Note that since the F4 is push-pull, if you run it at lower bias currents it will just transition out of class A at high output levels.
Also, at +/-32V you should consider the dissipation and voltage limits of the input jfets.
Note that since the F4 is push-pull, if you run it at lower bias currents it will just transition out of class A at high output levels.
Also, at +/-32V you should consider the dissipation and voltage limits of the input jfets.
Thank you.
If I need only 25W in Class-A and 50W for PP, how many current each ?
I used TR cascode for input pairs.