Output banana posts making contact to
PCB pads via compression using threads
and a nut (and some Green gunk) ?
A real No-No, PCB materials have plastic
flow under compression, soldered inserts
needed.
Output inductor held together with TyWraps ?
This is presented as Quality construction
and Design ?
PCB pads via compression using threads
and a nut (and some Green gunk) ?
A real No-No, PCB materials have plastic
flow under compression, soldered inserts
needed.
Output inductor held together with TyWraps ?
This is presented as Quality construction
and Design ?
Bonsai , i agree on the whole of your post but this point can hardly
be followed considering the existing devices caracteristics.
The Hitachi laterals 2SJ50/2SK135 or their TOP3 siblings 2SJ162/2SK1058
are rated 100W/7A and i used three pairs of the formers in a 2 X 100W/8R
amp that was used for stage monitor and often with 4R loads at full power
in overheated ambiances even it had no cooling fan and heatspreaders were
somewhat undersized for such a load , yet it did survive more than a decade
of such treatments before i retired it for home hifi use , still fully functionnal
after 26 years.
Such laterals fets are extremely rugged and obviously can withstand
more aggressive conditions than equaly power sized bipolars , the 20A/60W
requirement being inadequate since there s no such devices that have Id max
that high , rather the rule 40W output/100W TDP (in fact 200W since
we talk of a pair) should prove reliable enough for laterals.
Last edited:
obviously you are confused also ...
Perreaux uses 60.000 mfd for a maximum power of 2*350W I use also 60.000 mfd for my P3A amplifier but this is a personal over design idea .
I will not continue to argue on this matter since practice has proven you wrong most of manufacturers go with more or less with the profussion logic even the pro ones
No completely replace them with radial vertical capacitors higher voltage aand temperature since the pcb can also feature that
Smart kit capacitors was already 10 year old stock at the time
that is correct but we talk about double threads and nuts of 8mm diameter screw facing a dual layer gold plated and extremely thick pcb ...you will not get more contact than that ever
As about the coil when you have a coil of 30mm tube diameter , 2.5 mm of thickness , and 16 turns in 3 rows the coil will actually support it shelf without any problems first because of the thickness the straps is just an extra precaution
Yes, I am also using tie wraps to secure the output inductor in my e-Amp - works very well.
Maybe I underate the mosfets a bit.
I need to watch out because Bob will be here any minute to set me straight as well.
;-)
Bonsai, that is some target specification.
I agree on many of the details you state.
I want to look at a couple where I am not quite in agreement or maybe I should say, I would like to add conditions.
a.) the 1/3 Maximum Power for 1hour and Ts<=60ºC is an almighty target to reach. I would suggest this really only applies to a PA amplifier that must be expected to run all night keeping the paying audience happy irrespective of how hot the Ta becomes. For a "Hi End" amplifier reproducing (all normal forms of) music, the average levels of output power are so much lower than 1/3 of full power that this target can easily be reduced. I'd suggest that full power for 10seconds and 1/3 full power for 1minute and 1/8 full power all day are targets that will ensure reliable operation for many years/decades to come. Not select, but meet all three targets with maximum mains voltage applied.
b.) Transformer rating @ 3 times total maximum power into specified load. In my opinion this is unnecessarily heavy and expensive for ClassAB, where the bias Pq ~1% to 5% of total maximum output power. If Pq were instead ~10% to 30% then a bigger transformer may be required to run cool enough to ensure reliable operation. If the amplifier were ClassA and the Pq is now ~200% then really big transformers are necessary. For these I agree completely with Nelson Pass where he specifies in the range 6 times to 10 times total maximum power output.
For "Hi End" I would suggest that 2 times is adequate for good sound performance and cool running if Pq < 20% (4pair for 200W/ch using ~+-60Vdc supply rails and biased to ~80mA/pair using Re=0r33)
If one were building a 9pr Re=0r15 then we are in a different ball park, not "just" Hi End.
I agree on many of the details you state.
I want to look at a couple where I am not quite in agreement or maybe I should say, I would like to add conditions.
a.) the 1/3 Maximum Power for 1hour and Ts<=60ºC is an almighty target to reach. I would suggest this really only applies to a PA amplifier that must be expected to run all night keeping the paying audience happy irrespective of how hot the Ta becomes. For a "Hi End" amplifier reproducing (all normal forms of) music, the average levels of output power are so much lower than 1/3 of full power that this target can easily be reduced. I'd suggest that full power for 10seconds and 1/3 full power for 1minute and 1/8 full power all day are targets that will ensure reliable operation for many years/decades to come. Not select, but meet all three targets with maximum mains voltage applied.
b.) Transformer rating @ 3 times total maximum power into specified load. In my opinion this is unnecessarily heavy and expensive for ClassAB, where the bias Pq ~1% to 5% of total maximum output power. If Pq were instead ~10% to 30% then a bigger transformer may be required to run cool enough to ensure reliable operation. If the amplifier were ClassA and the Pq is now ~200% then really big transformers are necessary. For these I agree completely with Nelson Pass where he specifies in the range 6 times to 10 times total maximum power output.
For "Hi End" I would suggest that 2 times is adequate for good sound performance and cool running if Pq < 20% (4pair for 200W/ch using ~+-60Vdc supply rails and biased to ~80mA/pair using Re=0r33)
If one were building a 9pr Re=0r15 then we are in a different ball park, not "just" Hi End.
Last edited:
Is it possible you have misunderstood Bonsai?
50mF/rail/ch equates to a total of 200mF for a dual mono amplifier.
I'm using 40000uF per rail per channel in my latest design. This generates a lot of harmonics on the mains and high peak charging currents in the secondaries and associated connections to the rectifiers and cap bank which have to be managed through careful layout and attention to wiring etc. I think once you get much above this, you really have to consider whether more capacitance will bring added benefit in a class AB design - and the output stage needs to be sized to handle what the PSU can deliver as well otherwise whats the point?
Noise is also an issue, but using smart techniques to filter out noise to the front end are probably more effective.
Noise is also an issue, but using smart techniques to filter out noise to the front end are probably more effective.
Andrew, what I meant was 20 milli Farad (mF) per rail per 100W up to a max of 50mF per rail per amplifer. So using this calc a 200W hi end amp would have 80mF per Channel (i.e. both rails) and total 160mF. Per my earlier post, much above this I think its diminshing returns, and thats why I suggested the 50mF per rail per channel as an upper limit.
A lot of people may diagree with this, saying its too high, but PSU quality is critical IMO. I am sure there are engineering justifications for 4.7mF per 100W, per rail per amp . . . but I think those are not really high end designs. My Marantz PM7000 has 12mF per rail and is rated at 105W per channel. Its a very mid-fi amp.
A lot of people may diagree with this, saying its too high, but PSU quality is critical IMO. I am sure there are engineering justifications for 4.7mF per 100W, per rail per amp . . . but I think those are not really high end designs. My Marantz PM7000 has 12mF per rail and is rated at 105W per channel. Its a very mid-fi amp.
Last edited:
I guess you discount new switchmode for audio technology then? I recommend you watch this space, some of these things are starting to do extremely well. I just bought my first pair to try and it took some convincing let me tell you. fully regulated 500W using what is basically a class A/AB amplifier as the output stage, 400g it weighs including 2 auxiliary supplies. dynamic behavior and noise matches linear regulated and beats the significant linear unregulated supply it was tested against, no hum at all, its absent from the measurements done by a third party on a high grade agilent analyzer.
no sag? haha the thing doesnt know the meaning of the word, if anything you might say its too stiff and could perhaps cause resonance with some drivers. it was this dynamic performance that took it over the line for me
thin film caps once they get large enough to be more useful could be the next big thing, check them out. extremely linear
finally some people that actually listen to the technology they build are starting to design these switchers, i'm actually quite excited to try it out. is it high end? I guess we'll see.
extremely little capacitance
no sag? haha the thing doesnt know the meaning of the word, if anything you might say its too stiff and could perhaps cause resonance with some drivers. it was this dynamic performance that took it over the line for me
thin film caps once they get large enough to be more useful could be the next big thing, check them out. extremely linear
finally some people that actually listen to the technology they build are starting to design these switchers, i'm actually quite excited to try it out. is it high end? I guess we'll see.
extremely little capacitance
Last edited:
Might be more accurate to state the max allowable ripple level on the rails, and at which load, for "high end" amps.
Or explain why PS capacitance should go up linearily with power instead of the sqrt of output power.
(device Vce at least 125% higher than max supply rail voltage)
No he originally said uf not mf ...
As to a 1/3 power conditioning , amplifers are traditionally bench tested this way , they are checked after running for a 1/3 of their rated output for 1 hr, before performance testing, this was done by all of the major mags...
So you have not done any subjective analysis regarding added capacitance in the psu ...?
My understanding is that inductors in the psu really benefits tubes and constant current class-a amplifiers and not A/B amplifers ...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.