| Stabist |
Hi!
I just need a final confirmation - during to some different reasons (mainly short of $$ :bawling: ) I've reached a decision to scale down a bit the first of my two A5 projects ...
So - I have soma "spare" transformers with 18V secondaries - that just calls the conversion into A3 :D
I have heatsinks that are capable of dissipate 200W/stereo (but 300W would be probably allready too much for them - and as said no $$$ to buy new at the moment) - again - A3.
And I have allready soldered PCB for A5 - but it's a "flexy PCB" - so the number of input fets can be variated :D
I have enough PSU capacitors :D (at least something I'm not short off :D)
So now I have two choices:
- direct scaled down A5 with 12 output transistors and just adjust necessary resistors because of lower PSU voltage
- make a sort of balanced input Aleph3 with 8 transistors, etc ...
I find second option better - I guess there is no need for those 2 additional pairs of transistors considering lower PSU rails ...
And if I'm correct - it's just important to match appropriate values for bias setting and voila - my "crosser" will be born :D |
|
|
| MikeW |
Build an A-30. Use all your transistors and crank up the bias.:cool:
I built one with 8 outputs per side and a 21-0-21 800va transformer with a C-R-C-R to drop the voltage to 26. I can't get it away from my son. |
|
|
| Stabist |
| Hmm - is there big difference (I know - I will check it out in schematic) between A5 and A30 - because as said - A5 PCB is allready assembled together ... |
|
|
| Nelson Pass |
| quote: | Originally posted by Stabist
Hmm - is there big difference (I know - I will check it out in schematic) between A5 and A30 - because as said - A5 PCB is allready assembled together ... |
Not so big that you can't tack one on to the other's board. |
|
|
| Stabist |
Hmm - what I like most at A3 is 60W/4R - because my speakers have 5R ... I see A30 has "only" 40W ...
Well - what needs to be done is a little more deep look into things (now you've give me another option :D) so ... |
|
|
| Stabist |
Thanks mr. Nelson - I've made a short comparison of A3, A30 and A5 schematic - in the topology itself there are really just few differences - and Ok, I would have (Ok that I would do anyway) to change few resistor values ...
Now I have these Q:
- how come A3 had such high capacitance bank in PSU??
- What limits A30 power in 4R comparing to A3? |
|
|
| MikeW |
| It depends how high you bias it. Wuffwaffs has an excel speadsheet program for Aleph power. |
|
|
| Stabist |
Just to check it out - biases for original A3, A30 and A5 were??
I have in mind somethin like - A3 around 2,5A, A5 2,1A and A30 ?? |
|
|
| MikeW |
| Bias at 2.5 amps with 24 volts at the MOSFETs will give you 45 watts @5 ohms and 50 watts@ 4ohms.:D |
|
|
| Stabist |
Yep :D
And 2,7A will give me magic 60W :D But also (according to that sheet) around 135W dissiptaion/channel :( Which could allready be a little overdosed for my two heatsinks ...
BTW - another Q - if I use 4 heatsinks (like original Aleph) - BUT I mount transistors only on two (and so avoid long wirings) - but I DO connect all four heatsinks with massive Al "corner blocks" - will it spread the heat pretty proportional or will I get two HOT heatsinks and two medium warm ?? |
|
|
| Stabist |
Now I don't understand it anymore at all ... :(
I've done some simulations with wuffwaff's excell sheet:
- if I take rail 34V, bias 2,5A and AC current gain 50% - i get 60W/8R BUT only 50W/4R (and the climax at 6R with 75W) ... Hmm - does that mean that in "real life" AC current is somewhere around 62-63% or what's wrong with my calculation??
And also - A3 - 25V rails, 2,5A bias and 50% - 30W/8R, again 50W/4R and climax at 4,5R with 58W.
So from this point of view - it's almost same for my speakers wheter i would made A3 or A5 ... ?!?!? Hmmm ... |
|
|
| Brian Donaldson |
| Just replace R19 and R21 (on A2 scem) with pots and adjust ac current gain and bias until you find the right balance of heat, and current gain to suit your needs, then measure and replace with fixed resistors. |
|
|
| wuffwaff |
Hi Primoz,
you´re right, to reach the specs at the specified bias it is neccessary to dial in a bit more ac-current-gain or a bit more bias.
William |
|
|
| Stabist |
| AO does this mean that also original Aleph has ac current gain around 60% |
|
|
| uli |
AFAIK the Aleph 2 has a current gain of 69%
Uli
:nod: :nod: :nod: |
|
|
| wuffwaff |
Hi Primoz,
either that or maybe the power into 4 ohms was lower. Just look at some old tests where power was measured.
William |
|
|
| uli |
Primoz,
the formula to calculate the ac current gain should be:
gain = (R1/R2)x(R3/R4)
where:
R1 = output sense resistor (A30: 0R47 / 6)
R2 = sorce resistor of aleph current source (A30: 0R47 / 3)
R3 = senseresistor of source of aleph current source (A30: 1k)
R4 = feedbackresistor of outpt sense resistors (A30: 820R)
=> (0,07833/0,15666)x(1000/820) = 0,5 x 1,2195 = 0,6097
=> ac current gain is ~ 61% with Aleph 30.
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
COOL!
Now I'm learning something :D
THANKS GUYS!!!
That will help me a lot in "constructing" my AlephCrosser :D |
|
|
| uli |
oh, I forgot:
maximum output current from the current source is about:
Imax = (I idle)/(1-X) where X is the ac current gain.
=> max. power is (Imax * Imax * R load)/2 assuming the psu voltage is at least +- ((Imax * R load) + 3) Volts.
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
| Btw - another one - why had A3 so big PSU capacitance - I mean - A5 or A30 have only oarund 60% of A3 capacitors value in PSU ... |
|
|
| uli |
In fact is the stored energy
in A5 60 Joules (35˛ x 0,1F)/2
in A3 55 Joules (25˛ x 0,176F)/2
in A30 37,5 Joules (25˛ x 0,12F)/2
Maybe the capacity of A30 is enough cause this amp is not as potent as A3 concerning current capacity.
The A5 and A30 are maybe designed rather for 8 Ohms load than for 4 Ohms.
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
I think there are some interesting project infront of me :D
first this A3/A5 crosser and after that, when I'll save enough $$ to complete the "ultimate Aleph project" I have in mind an Integrated Aleph combining A5+ and AP 1.7
In somewhere in between time to finnaly finalize MiniA/BoZ combination ... |
|
|
| wuffwaff |
Hi Uli,
is the stored energy of a power supply really important? To use it you must reduce the voltage to zero wich is a bit difficult when listening to music.
I think it is more important how much energy you can take out of the power supply for a given voltage drop wich would mean that capacitor size is the important faktor and not the voltage.
William |
|
|
| Stabist |
Hi!
Today I obviously had too much time left so I've build few chassis 3D models ...
The thing is - I have 4pcs of Fischer's SK 56 heatsink - each 145mm high ...
So my first "idea" (more like my interpretation of original mr. Nelson idea) quite some time ago was this: |
|
|
| Stabist |
But unfortunatelly I can not squeeze everything inside such chassis - I mean my toridal transfomer is pretty big, and 4pcs of 68000uF/50V Sprague capacitors also needs space - so to somehow squeeze everything in - there would have to be a lot of pretty long wires passing very close to transformer, etc etc .. And I don't like that to much ... - I have some bad experiences with Zen and internal wiring layout ...
So I've thought it out to have some kind of two floor amp - with PSU in gorund floor and circuit in upper (Btw - mr. Nelson - I still think original Aleph's are one of best looking amps ever) |
|
|
| Stabist |
And I must say that would look pretty awsome for my taste! So some time ago I was almost sure to make this kind of a chassis ...
But ... Then I've figure it out - if in first case there was almost no space - in this version there would be a lot of unused space - so why not add also preamp in same chassis ...
BUT - than again - how to deal with buttons etc enough aestetically not to spoil clear lines too much ... Hmmm ...
So I've decided to give another try - this time more classical - but because two 300mm heatsinks would be use per each side - that would mean more than 600mm in depth :eek: overall ... Hmm - and only 150mm height :bawling: That would look a bit strange ...
So I've decided to make a kind of golden ratio proportions - so the depth around 600mm, width 370mm and height 220mm ...
And that's what it came out: |
|
|
| Stabist |
| I've simulated 2 versions - I think it's best to let the picture show what a major difference is ... |
|
|
| Stabist |
And a photo with both versions that are still "in play" before I make finnal decision - it just depends whether will I do Integrated Aleph (probably mix of A5' and AP 1.7 (maybe BZLS)) or power amp A5+ (the plus stands for a little bit higher biased version) ...
OH, BTW - both models are in ~1:1 scale ... |
|
|
| uli |
| quote: | Originally posted by wuffwaff
is the stored energy of a power supply really important? ...capacitor size is the important faktor and not the voltage.
|
Hi William,
of course you are right ;)
I just tried to explain the psu variations in Aleph amps.
(PS: Mine are loaded with 8x33mF/32V per monoblock)
Uli
:nod: :nod: :nod: |
|
|
| wuffwaff |
Primoz,
just make shure that the air can enter the heatsinks from below. Personally I like the "smooth" one better. Maybe you could just make the sides closed and leave the bottom open?
William |
|
|
| Stabist |
Yes of course - in all cases I would taken care somehow for free air flow ...
Will see what will come out at the end ...
But I must say - after throwing some measurments and to the paper and calculated a bit how much $$ will I spend on Aluminium etc - I got pretty high number ... But than again - it's "the ultimate project" and I intedn to live happily for quite some years with him :D |
|
|
| uli |
Primoz,
I know nothing about your mechanical skills, but what about diy the heatsinks????
http://sound.westhost.com/heatsinks.htm
unfortunately the famous lotusblossom site seems to be offline, there was a complete plan for diy heatsinks
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
| Infact - I must do some compromise between costs and final result - and I've calculated that DiY heatsinks are very expensive - then it's cheaper for me to buy new heatsinks that will have "right" height and all ... |
|
|
| Stabist |
Hmm - tonight I've decided to try figure out with "real" parts ...
First thing was - I've decided to try it out with not so mirrored interior layout as I wanted it till now - and the result is I might squeeze everything into basic a little original A3 look-a-like chassis - so on this picture you can stereo PCB (but I will have to cut it to halfes) and PSU parts ... And a kind of internal layout that would work it out I think: |
|
|
| Stabist |
And view from up - the idea is:
PCB cut to half - so each channel is mounted on one heatsink - in the photo that would be "North" & "West" ones - but transistors would be mounted 3 per heatsink - so I would spread them equally - so there will be some wiring job ...
I have some thought about trannie - probably it would be a good idea to shield it with some "magnetic material" ... |
|
|
| Stabist |
| Or maybe it would be better to lay down capacitors - but then the trannie gets very very close to heatsink - but I gues that doesn't hurt him at all considering some people mount trannie directly to heatsink ... ... |
|
|
| Stabist |
| Or maybe try it simetrical - and then to mount PCB plates to opposite heatsinks (left and right) - and again - mount transistors on all 4 sides ... |
|
|
| Stabist |
Ofcourse the rear heatsink will have to sacrifice few fins to get place for connectors ...
But for 230V, fuse connector and for power switch I had in mind to place them to bottom - as near as possible to trannie ...
btw - my first plan was to make it with centralized trannie ... But I guess it's better to have capacitors closer than in this case: |
|
|
| uli |
I like the last one best...;)
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
Maybe with little mod the last one becomes more attractive to me also ;) - and the mod is - cut PCB to half, capacitor closer to trannie - and again - each channel on opposite heatsink and then transistors 3pcs/heatsink ...
BTW - what about mounting 6 transistors on left and 6 on right heatsink - would front and rear had some major effect to cooling or would I get local overheating?? I mean - each heatsink is rated around 0,32K/W ...
Probably it would be smart thing to put e.g. copper bar to side heatsinks???
Oh btw - the amp I'm talking about is no more mix between A3&A5 - this is "fully" A5 - trannie is 900VA/4x25V, capacitors are 68.000uF/50V ... I have some MUR860 left - so I intend to use them ... And that's pretty much it ... |
|
|
| Stabist |
| Would it look strange to put input and output connectors on lid? So I avoid cutting few fins - the problem is to find anyone that does that over here ... All "proffesionals" are to bussy to do it for me :((( |
|
|
| uli |
why dont you put them on the top cover?
I would put 3 Fets per heatsink to distribute heat evenly.
Uli
:nod: :nod: :nod: |
|
|
| Stabist |
| quote: | Originally posted by uli
why dont you put them on the top cover?
I would put 3 Fets per heatsink to distribute heat evenly.
Uli
:nod: :nod: :nod: |
You refer to connectors I gues - yes I got idea to split the top - let's say to 1/4 at back and 3/4 - and 1/4 would be fixed part with connectors on;
I also think it is better to have transistors on all 4 sides - but what about long wires (OK, it is also truth that transistors are on 4 PCBs - so it's only one set of wires/3 transistors ...
Oh btw - that's how it looks to squeze capacitors as close as possible to trannie and to mount two PCB on left and right heatsink ... I think I like this most :D |
|
|
| MikeW |
Make it narrow and deep. Put 2 heatsinks on each side. Made a back panel for the inputs and outputs. Put a nice piece of wood on the front.:cool:
It would keep the transformer away from the inputs.:smash: |
|
|
| Stabist |
MikeW - I know that's "a simple" option - but I've built a cardboard model the other day - and it's just to big ... Gosh - it was more deep than the A5 and Zen v4 that are currently at my place ...
And also - I intend to build chassis by myself - it's quite easier for me to work with 300x300mm Al plate than with 600 x 300 ...
But nevertheles - never say never - I might change my mind ;) |
|
|
| Dane |
What if you mount the toroid vertically on the chassis front?
Will the 4 caps fit between it and the rear chassis? |
|
|
| Stabist |
| quote: | Originally posted by Dane
What if you mount the toroid vertically on the chassis front?
Will the 4 caps fit between it and the rear chassis? |
No - unfortunately heatsinks are 145mm high and trannie is 155+ ... But if it woulb be possible - then caps would squeze in, yes ... |
|
|
|
