Aleph-X: High-Power Version

[Blitz]

Just to avoid any confusion on the VAC ratings above: I am looking in the example only on the transformer-part (AC), William has made some assumptions on voltage drop through Diodes, filtering, transformer regulation to show how much V AC is needed to get 20 V DC. I personally would simulate this with PSU-Designer, as this is the must accurate and simple way to come up with reasonable AC ratings. And you need as well some (4V*10 A perhaps)additional VAC to incorporate the losses.

Best Regards

[Blitz]

...and than multiply the whole rating by two (I forgot the current source), so a 480 VAC-Transformer would be the minimum.

As said earlier my personal favorite is this combination ( I have a 6 ohm speaker):

voltage 19,7 volts actual rail voltage at the fets
bias 6,3 amps total bias for one channel
ac current gain 50%
number of fets 12 total number of fets for one channel
Peak current 6,3 amps maximum output current
power 8 Ohms 78,3 watts
power 6 Ohms 104,4 watts
power 5 Ohms 99,2 watts
power 4 Ohms 79,4 watts
power 2 Ohms 39,7 watts

Dissipation 248,22 watts
Dissipation per fet 20,685 watts
Transformer secondaries 15,15 volts for 1.3
16,42 volts for 1.2
17,91 volts for 1.1
19,70 volts for 1

Transformer wattage 496,44 watts for factor 2
744,66 watts for factor 3
Source-Resistor 0,48

which resuts (per channel)

- for pi-filter:
min. (19,7V+4V)*6,3A*2 = 298 VAC
max. (based on current spikes in normal operation):
(19,7+4V)*25A*2=1185 VAC

- for choke-input:
min. (19,7V+4V)*6,3A*2 = 298 VAC
max. (19,7V+4V)*7,5A*2 = 355 VAC
calculated with 2 chokes (15mH) per channel

Best Regards

[Blitz]

OK, guys, I have just realized that the Version 1.0 has been downloaded about 100 times now, but no one has any comments ? Well than let's have a look together at least at the current source. I changed it now to the version hifizen is suggesting in his beta-version 0.9: So, It is basically Fred's version with connection to the negative rail. Clearly, this is nothing new, but should serve me / you as a wrap-up how the circuit would look like. I guess the resistor values need to be adjusted as we have changed the rail voltage quite a bit. Is there anyone out there with a simulator who give us the right values (wold be great to have it for 22,7V and 19,75V)?

THanks & Best REgards

[nania]

Blitz

If you feel there are mistakes in the spreadsheet, than shows us where exactly.

The buildout formula I described is an attempt to simplify in a few easily comprehensible steps how to buildout ANY SIZE version of the Aleph-X. I haven't seen the Williams spreadsheet and don't know the formulas used within it so I won't comment on its merit at this time but I suspect there is an error because the power value described under a 2ohm load seems a bit low. There will be some falloff (collapse) as the speaker load falls below the 2.5ohm level but I doubt that it will be as steep as the Williams model presents it. I should remind you that my example was to build out a version that would be stiff to 2.5ohms not 2ohms. It would be interesting to see what the spreadsheet model says the power output would be under a 2.5ohm load which is the example you should have provided if you wanted to highlight a disparity or attempt to invalidate my method.

I am not sure what your message / posting is about.

Well, my post was an attempt at trying to move the elephants before the mice. Many people want to build bigger than what is practical and the result is usually not as good as something planned to a purpose. I tried to provide some useful perspective since once you build your amp, you will have to live with it and pay the electric bill it generates. If others on this forum are like me, they have their amps on for the better part of the day so this is issue is not insignificant IMHO

I believe your VA calcs on the transformer are wrong as well

You are entitled to your opinion. I will tell you that any time the transformer is stressed the possibility of the music signal integrity degenerating increases. I will also tell you that the use of inductance in the power supply is not a panacea and I much prefer a simpler CRC power supply with the capacitance size and quality reducing the ripple but again, this is only my opinion and the readers may form their own as they read through our respective posts.

[Ian Macmillan]

Since there seem to be very few comments forthcoming on your high power circuit, I thought I would chip in with something. I suspect that the lack of opinions has more to do with the time of year than any lack of interest.

I've only had a quick look at your circuit but one thing strikes me as indeed it did for Grey's circuit, and that is the inclusion of V1, R11 and V3, R33. I appreciate the need for some form of differential current adjustment but these parts can easily cause more problems than they solve.

Others, including Grey, have mentioned that these parts are not strictly necessary, but I would go further and say they are best removed. If you model this current source in isolation, you will see that the current is not as constant as one would like with variations in voltage across the current source (as opposed to at the intended control point). This has to do with the variation in current in (and hence in the voltage drop across) V1, R11, etc. as the voltage across the entire current source varies. The net effect of this is to make the amplifier more sensitive to DC at the output and hence less easy to trim for zero voltage.

In my own prototype of Grey's circuit, I have also omitted the over current protection parts associated with Q4.

I haven't yet tried a high power version although it is my intention to do so. Hence I cannot yet comment on what really matters other than through simulation.

Ian.

[Blitz]

Ian,

Great to have you here and thanks for your comments. Do you see any issues with the current source itself ? HAve you the possibility to simulate the right resistor values for it ?


Nania,

Please don't get me wrong: I appreciate your input, but I truely like to learn from your calcs vs. the calcs other made. You find Williams Spreadsheet in this threat ( page one or two) as an attachment. So, please have a look at it and suggest your corrections.

Your essay on Watts is correct, but it is like someone comes to Mercedes-Dealer and wants a Mercedes, but while being there someone else tells him that in most situations a VW Polo will be equally fast and that he should get a Polo. So, this threat is meant for people who know what they want. Or for all the Aleph 2 Owners / Builders who want a X-Version now. Besides thjis I am with you: I run a triode amp with 25 W SE (VV520V2) and it is awesome ( could have only some more bass control though....).

On the transfomers we are more or less on the same page: If you use a CRC_Filter and don't want any limitations on sonics, go for a 1200VA. Technically less would do so...I missed that comment on your threat. By the way: What issue do you have with choke-input ? With a nice LCRC you have a better ripple rejection than with similar (R)CRC-Filter, need much less VA for the same sonics ( see above), have much less stress ( 7,5 A peaks to 25A peaks) on the diodes etc.

Best Regards

[nania]

Blitz

...it is like someone comes to Mercedes-Dealer and wants a Mercedes, but while being there someone else tells him that in most situations a VW Polo will be equally fast and that he should get a Polo.

I don't know that this is a representative analogy. Saying that building a smaller amp which sounds better than a bigger one is comparing a Polo to Mercedes might be insulting to the Tri-Star if the quality of our aural experience is the goal. You will concede the point that louder is not always better, yes? So why put your resources into a bigger amp if it doesn't help you enjoy the music more? The only reason would be to match it better with a future speaker or listening environment and if you are making that kind of a change, you will most likely improve those parameters rather than worsen them.

What issue do you have with choke-input ?

Outside of the ways it can create more problems than it can solve if implemented poorly (with the wrong value, wrong diameter wire in the coil, etc.) there are the issues related to its physical footprint in the case, its cost and sonically its potential for current limiting/wavering effects and the consequence that has on bass response. To me, these outweigh the 10mV additional noise reduction you may get using a coil.

[Cobra2]

Please go onwards with the calculations,

-but remember that (I think) most of the hundreds of PCB-buyers will use a filterless PSU, atleast for a start,
(and might fine-tune the PSU when everything is in working order/or as a future upgrade/tweak).

What do we then come up with?

Arne K
NORWAY

(Need hot amps...it's freezing outside...)

[Blitz]

OK, Nania I see that we don't come together on choke-input (still I don't get the point with the current though as the delivery should depend creatly on the impedance of your PSU, mainly determined by your last capacitor and real good chokes have not more resistance than the resistor you use anyhow). I guess it won't help that most of the Cellp-Amps by MArk LEvinson use them, and Nelson and many others besides me like chokes as well as stated several times in other threats.

For everybody else interested in this subject, one of the greatest transformer winders building any choke you can think of (his standard are 6 chamber-chokes to avoid resonances, and than he has three even better options on top of that to offer) you will find here:

http://www.ae-europe.nl/ (dutch side, bit very friendly guy talking german and english as well)

For the tube freaks: You will get there OPTs on or surpassing the level of TAngo / Tamura even with amorphous core for a fraction of the price of Tango.

Best Regards

[Blitz]

Cobra,

Sorry, our threats where crossing each other. Yes, I am looking for some input o the component values as well.

And I promise this was my last posting on PSUs. BUT BEsides heatsinks the transformer migh be t h e most expensive part in this amp, right ? As the voltage ratings are significantly different between the different filters used ( and you use always one, a simple C is like a pi-Filter from the transformer's point of view...so up to 1200VA needed for 100W), you should know where you want to go to. Fine-Tuning in that respect is only possible if you ordered a special transformer with all the taps for these different scenarios. Very unlikly for most of us.

And now back to the circuit: Any thoughts ?