| MEGA-amp |
Hey everybody,
I finally decided what to build after reading countless posts and studying, an Aleph 2. And I highly recommend reading Teach Yourself Electricity and Electronics by Stan Gibilisco, very helpful.
I am purchasing 4 pcs 10"X18" 1/2" baseplate w/4" fins for alpeh 2 monoblocks. I plan on mounting KRISTIJAN's FET boards squared and centered on the sink, 1 on each sink of course. I'm not great with formulas, so I am wondering if this size heatsink is sufficient? I am assuming it is.
Also planning on purchasing Plitron's 1500VA 38v sec transformers for each block. I will be using a Pi filter per KRISTIJAN's PS schematic. What value should I use for the coil? I've seen people using 3.9mh coils with that secondary voltage, just wondering if i'm the ballpark.
Many thanks,
John |
|
|
| moe29 |
heatsinks sound big enough.
FETS mounted in the middle is ideal.
1500 VA toroids are fine... so would 1000 VA, up to you. i think the
original had 750VA? (i think i remember reading that somewhere)
38 volt secondaries are fine.
i used 2mH coils in my Aleph 2... 4mH would be fine too. Kristijan's
PSU shematic is a good one to follow. Check out mcmelectronics.com
for them.
sounds like you've done your homework! Good Luck, you will not
be disappointed :) |
|
|
| zygibajt |
I think 2mH should be ok.
If you haven't used the coil 38VAC might be a bit too high,but with coil should be allright.
1500VA should be just fine.
I know (from good source) that original Aleph 4 uses 2KVA transformer,so I think Aleph 2 monoblock uses at least 1,5KVA,and I wouldn't be supriced if it's more.
Good Luck with your project!
Bartek |
|
|
| moe29 |
Aleph 4 is a stereo Aleph 2 in one chassis, thus 2KVA transformer.
1K VA is plenty for an Aleph 2. More is overkill, but that's up to
you and your wallet. ;) |
|
|
| zygibajt |
| quote: | | Aleph 4 is a stereo Aleph 2 in one chassis |
Not exactly. |
|
|
| grataku |
| This is an old subject but with the different AC freq the low pass should be a little bit different betwen US and EU. In practice 4mH in EU is equivalent to 2mH in the US. |
|
|
| moe29 |
Quote from the service manual:
The Aleph 4 is a stereo 100 watt audio power amplifier which
operates in single-ended class A mode. |
|
|
| zygibajt |
I meant that Aleph 4 is not exactly two Aleph 2 in one chassis.
The power requirements might be a little bit diffrent becouse of the small diffrances in schematics.
Bartek |
|
|
| GRollins |
True, the Aleph 4 and the Aleph 2 were both 100W amps, but the Aleph 2 ran higher bias. All things being equal, more bias is good, so the nod goes to the Aleph 2.
Grey |
|
|
| Brian Donaldson |
And the power supply voltage is totally non critical. I've ran mine all the way from +- 25V to +- 65v with very little sound difference (except power output) I had to tweak the bias back (I have a pot)a bit to keep temp down at 65v.
The amps sounded more open with the higher rails an lower bias. Best compromise was around 4.5A and +- 50V
I strongly considered putting my 22A variac in an enclosure and using it as a soft start and heat adjustment for non critical listening so I could choose not to put almost 1 kw of heat into my balmy Houston livingroom. |
|
|
| Zapped |
Grey seems to be saying that a high bias will yield the best sound (something I have seen asserted elsewhere in these threads a few times), but Brian says the best sounding option for his Aleph was higher rail voltage and lower bias. How to decide?
Would it be reasonable to just pick a rail voltage within some range, and then bias as high as you can without overheating the sinks? Or is there a need to have a certain cross product between the two values? Or meet some other rule of thumb or design requirement?
By the way, how come the Aleph X designs in the threads seem to demand a lower rail voltage than the regular Alephs, for a roughly given power output ? (or is that just my mistake impression?)
Hope these don't sound like silly questions.
Larry Wright
Seattle area |
|
|
| Brian Donaldson |
The AlephX is a bridged pair of Alephs with crossed feedback so the will have 40% less voltage and 40% more current for the same output levels and impeadances.
The sound wasn't better with less bias, but we thought they sounded more open and airy with the voltage raised. But I reduced the bias a little to keep the temp down. You can also (I did) add a pot to tweak the AC curent gain to prevent overcurrent clipping on the positive swing without upping the bias.
Pick a voltage that gives you the calculated wattage you want, Make sure your heat sinks and # of fets are up to it and twiddle the knobs untill you like it. Then pull the pots, measure, and replace with fixed resistors. If you want and have adequate sinks, add more output fets. The circuit is beautifully simple and scaleable and very forgiving of different voltages.
Thanks again Nelson. This amp makes me wonder why anyone would make an amp with 30 transisitors (excluding output devises). |
|
|
| GRollins |
Hmmm... current to the exclusion of voltage? Not quite what I had in mind. When I said 'All things being equal' I meant that the rail voltage would be held constant. There is a slight difference in the rails of the Aleph 2 and Aleph 4, but I seem to recall that the bias current is--percentage-wise--more tilted in favor of the 2. Ideally, you'd have both high voltage and high current, though that idea seems to stick in the craw of some folks who have bought heavily into the current uber alles mindset. The reason we don't see both is a function of money and heat dissipation.
Given that the rail voltage is set pretty much by the choice of the transformer (assuming no regulator or voltage dropping resistors), then we as DIY folks are pretty much at the mercy of what transformers are available at the time that we happen to build a given circuit.
The bias current, however, is set by the circuit. Given that it's much cheaper and easier to buy a 50 cent resistor than a 50 dollar transformer, we have more options as regards to bias current.
(Yes, I'm splitting hairs by arbitrarily deciding that the transformer isn't part of the circuit. Clearly, it is. Still, for the sake of the question at hand, let's draw an imaginary line down the page dividing the power supply from the active circuitry.)
In the final analysis you'll choose a transformer. That will determine what your rail will be. Then you'll run the bias up as high as you dare given your heatsinks and personal tendencies towards risk-aversion.
The Aleph-X is a bridged amplifier. Bridged amps are basically two smaller amplifiers run back-to-back. Assuming sufficient bias current, the power output of the amplifier will be four times that of one of the two smaller amps. I started by putting together the Mini-A, which was a 10W amplifier. Then I built two of those in tandem, so the resulting power output would be on the order of 40W or so. The grounding resistors at the outputs eat a bit, so the actual power comes in somewhere in the high 30's, depending on the actual value of the resistors. The upshot is that it's the difference between a 'normal' amp and a bridged one.
Grey
EDIT: Incidentally, I played with the bias on my first pair of Aleph 2s. Lower bias leaned out the sound, making it more "hi-fi" and subjectively tilting the spectrum towards the high end. Raising the bias made it warmer sounding; more tube-like. When the dust settled, I ended up running the bias about 10% higher than a stock pair of Aleph 2s. Given that I water cool the amps, I've got a lot of lattitude in heat dissipation. |
|
|
| kilowattski |
Mega-Amp,
Where did you find your heatsinks and could I be so nosy as to ask how much they cost? |
|
|
| MEGA-amp |
R-theta quoted 400US for the sinks, and 120US or so for ground shipping. A bit pricey, but I think it may be worth it.
JH |
|
|
| Brian Donaldson |
Grey,
I think I may have been misunderstood. My Alephs are biased a bit higher than stock. I added 2 fets per rail and dropped the source resistors to .68ohm. So without dropping the bias back a bit they ran around 6A at idle. I started with some surplus transformers that hummed and didn't give but 35v rails. I couldn't justify cranking the bias as hot as my sinks would allow. That would mean close to 7 amps. Disgusting overkill for 60 watts RMS out. To be sure not to make the same compromise again, I got 96v ct transformer with the thought of regulating or CRCRC or something. I wanted a few volts to burn. But to get them up and going, connected all of the caps together straight to the freds and varied the voltage through a variac.
A friend and I spent an evening twistng pots and listening, using only a thermometer and ears as test equipment. The Alephs were running midbass and tweeters only so we might not have noticed any leaning out. And I also never ran the bias below about 2.5 Amps. (actually I did have a little meter across a source resistor to know where the bias was for A/B comparisons.
Like you said. The transformer is something you hope to buy once. I was playing around with voltage and bias to find the relationship that best suited my needs. I can always use a larger or smaller resistor in CRCRC PSU to loose more or less of the voltage to afford less or more bias and stay within my heatsinks constraints. (altho these resistors are more like $8.00 each)
I shot my mouth off without thinking about the AlephX PSUs. They are bridged therfore require roughly half the voltage and double the current for the same wattage out. I was thinking of the relationship of halfing the load impeadance requireing 40% less voltage and 40% more current for the same wattage. |
|
|
| Brian Donaldson |
John (mega)
$400.00 isn't that bad for new custom ordered sinks. Mine were close to $500.00 from M&MMetals. And that was full stick of extrusion that I had to cut to length. |
|
|
|
