How do you go from 25 A AC to 13 A DC?
What's the calculation?
1kVa 20V rms and 25A rms per winding
Last edited:
A transformer feeding into a capacitor input filter must be de-rated.
The transformer manufacturer usually tells you how to de-rate and by how much.
A typical divisor is 1.8, or the maximum continuous DC output current is ~55.5% of the rated current (= maximum continuous AC current)
This takes account of the extra heating effect of the pulsed current in the windings where the copper loss is proportional to I²R
The AC current into a resistor load creates an amount of heat that determines the maximum output.
A pulsed capacitor input current that flows for only 5% to 15% of the waveform period and takes currents that are typically 3 to 10times the rated current dissipate massively more internal heat unless they are de-rated for that duty.
The transformer manufacturer usually tells you how to de-rate and by how much.
A typical divisor is 1.8, or the maximum continuous DC output current is ~55.5% of the rated current (= maximum continuous AC current)
This takes account of the extra heating effect of the pulsed current in the windings where the copper loss is proportional to I²R
The AC current into a resistor load creates an amount of heat that determines the maximum output.
A pulsed capacitor input current that flows for only 5% to 15% of the waveform period and takes currents that are typically 3 to 10times the rated current dissipate massively more internal heat unless they are de-rated for that duty.
Just so I got this right
1000VA with 18V secondaries is good for a continuous 27.8 AC rms
27.8 x 0.555 = 15.4A DC
10/15.4 = 0.65 or 65% of maximum rating.
1000VA with 18V secondaries is good for a continuous 27.8 AC rms
27.8 x 0.555 = 15.4A DC
10/15.4 = 0.65 or 65% of maximum rating.
Last edited:
Andrew's explanation in post #544 beautifully shows why one of the bias limits when building these amps is 'no more bias current (in watts) than 50% of your transformer VA.'
Yes.
So the question I would be asking myself is:
Will it run cool, or warm, or hot?
Will the Ta inside the ClassA chassis make it run warmer?
To ensure a long reliable life I would set the continuous DC current to ~25% of the AC rating, i.e. 7Adc. That will ensure it runs cool and only the high Ta of the ClassA raises the cool temp to warm.
To help with this I put the smoothing capacitors directly above the ventilation holes in the floor of a very tall chassis.
The transformer hung from a plate strung across the middle of that tall chassis and was surrounded by the still cool air coming up through and passed the capacitors.
Last edited:
I'm all for big transformers and agree in principle but if you don't have room for such a large transformer I wouldn't have an issue using anything up to 35% of the transformer rms current rating.
Thanks for your explanation about how you made the calculations.
Last edited:
Pa (and experience ) gave us rule of thumb - xformer VA =2 to 3 times dissipation
my sweet spot is 250VA Donut per usual FW channel , ranging from 60 to 100W of heat
however , not being too much involved in last few pages of thread , I would say that (post #523) 3 pairs per side is wimpy ..... except if using IRFP150 and even in that case I wouldn't burden them with more than 35W each
considering price of case and PSU for big chunky amp , mosfets are peanuts
one of these days I'm going to make Babelfish XS150 , and use it as phono preamp
my sweet spot is 250VA Donut per usual FW channel , ranging from 60 to 100W of heat
however , not being too much involved in last few pages of thread , I would say that (post #523) 3 pairs per side is wimpy ..... except if using IRFP150 and even in that case I wouldn't burden them with more than 35W each
considering price of case and PSU for big chunky amp , mosfets are peanuts
one of these days I'm going to make Babelfish XS150 , and use it as phono preamp
1) Yes I am aware of this but Andrew disagrees.
Andrew says VA rating of 2 x heat dissipation is too small.
2) I agree, and advised the same.
Andrew T does not disagree.
I said I would limit the continuous bias current to 7A.
That equates to a dissipation of about 350W.And gave my reasons why I would do that.
And fits into the range of 2 to 3 suggested by Zen.
Sorry Andrew you do DISAGREE.
23V DC rails (from 18V secondaries) with 10A draw is 460W at the heatsink.
460 x 2 = 920VA
You said 1000VA with 18V AC secondaries and 10A DC draw is too small.
I said "Yes."
and followed with what I would do to ensure long term transformer reliability.
Last edited:
You should be a politician.
I was pretty sure you were responding to the fact my maths was correct.
The statement below sounds like a diplomatic way of saying you don't feel the operating conditions are acceptable. It certainly doesn't sound like you are endorsing it.
Anyway, thanks for sharing your explanation behind your calculations. I like your conservative approach.
Transformer ratings, transformer regulation, core saturation etc is like black magic to me.
everyone has different attitudes to conservatism, (or Factors of Safety).
Much of my audio equipment is decades old.
I expect it to last "forever". Reliability for me means it should never fail.
You and everyone elses attitude to reliability will be different.
You make your own assessments and build accordingly.
Much of my audio equipment is decades old.
I expect it to last "forever". Reliability for me means it should never fail.
You and everyone elses attitude to reliability will be different.
You make your own assessments and build accordingly.
BA-3b P3 pots
a question regarding the P3 pot settings:
with two P3 pots on the balanced pcb, how do I go about changing the sound quality i.e. the harmonic content of the distortion? changing both pots in the same direction or changing one is enough or ....? just confused. sorry if this was already answered elsewhere; I missed that part.
also: will fiddling with P3(s) throw off the settings of P1s and P2s? as it took me a while to get those done with properly faked thermal conditions roughly equivalent to those with a closed amp housing.
Thanks.
a question regarding the P3 pot settings:
with two P3 pots on the balanced pcb, how do I go about changing the sound quality i.e. the harmonic content of the distortion? changing both pots in the same direction or changing one is enough or ....? just confused. sorry if this was already answered elsewhere; I missed that part.
also: will fiddling with P3(s) throw off the settings of P1s and P2s? as it took me a while to get those done with properly faked thermal conditions roughly equivalent to those with a closed amp housing.
Thanks.
On a balanced, if you don't have a signal generator, distortion analyzer (or good sound card FFT) and scope, etc... don't touch them.
If you have all the required widgets, I'd get them set as identically as possible, and then just stop. But others might want to actually play with P3 as designed, so you could merely set each half to similar distortion readings and see what you like.
A little, but not too much. The bias will not change that much as the circuit's total drain resistance isn't changing with P3, it's just getting divided unevenly.
I have everything, it's just that digital signal processing feels too much like work, so it is not my favorite at home;plus making cables/connections...Well I guess at some point I will get to it.
I am thinking for input I should probably get a transformer to go from signal gen unbalanced to balanced; then on output take a measurement on + to gnd, and another one on - to gnd and keep checking FFT on both as I turn the P3 on either side? or maybe one FFT of + to - would do a better job? I am fed up with adding the ground connection to split up the two halfs.
Where I was really going with my question was if the 2nd order was being cancelled in this circuit or the P3s could still add some. I was hoping Papa would offer some reassurance but he jumped it. I imagine it is possible he may have not spent much time on the balanced version so it is left to DIY 😉.
I am thinking for input I should probably get a transformer to go from signal gen unbalanced to balanced; then on output take a measurement on + to gnd, and another one on - to gnd and keep checking FFT on both as I turn the P3 on either side? or maybe one FFT of + to - would do a better job? I am fed up with adding the ground connection to split up the two halfs.
Where I was really going with my question was if the 2nd order was being cancelled in this circuit or the P3s could still add some. I was hoping Papa would offer some reassurance but he jumped it. I imagine it is possible he may have not spent much time on the balanced version so it is left to DIY 😉.
Last edited: