wow, this thread is busy again! Ok, let's see here...
CRC filters:
These aren't really sensetive to what particular component values you use... you can tailor the resistance and capacitance to suit the parts you have on hand, or meet your performance objectives in terms of ripple and power loss. You can parallel as many caps as you want to achieve the necessary value (within reason of course). Actually, you may achieve better performance with multiple paralleled caps instead of single large caps. If you want to do some comparisons, I recommend you download the PSU Designer II software from www.duncanamps.com. It's a really handy little tool when you just want some quick results.
Magnetic fields can be addressed either by using an iron core inductor to confine the magnetic flux to the core, or by putting air-core coils in a separate shielded enclosure. Iron core inductors are quite large and expensive. You do see them a lot in tube power supplies, where the currents are much lower. Of course, you can always just ignore the magnetic fields and hope they don't interfere with your circuit... maybe experiment with different positions and orientations for the coils to see what gives the least interaction with the circuit.
Nelson:
Do you find the thermal modulation of the thermistors a problem? Or do they have a positive effect due to their slightly reduced impedance while the circuit is drawing higher currents? I'm just wondering how they sound compared to straight fixed resistance in a CRC supply...
CRC filters:
These aren't really sensetive to what particular component values you use... you can tailor the resistance and capacitance to suit the parts you have on hand, or meet your performance objectives in terms of ripple and power loss. You can parallel as many caps as you want to achieve the necessary value (within reason of course). Actually, you may achieve better performance with multiple paralleled caps instead of single large caps. If you want to do some comparisons, I recommend you download the PSU Designer II software from www.duncanamps.com. It's a really handy little tool when you just want some quick results.
Magnetic fields can be addressed either by using an iron core inductor to confine the magnetic flux to the core, or by putting air-core coils in a separate shielded enclosure. Iron core inductors are quite large and expensive. You do see them a lot in tube power supplies, where the currents are much lower. Of course, you can always just ignore the magnetic fields and hope they don't interfere with your circuit... maybe experiment with different positions and orientations for the coils to see what gives the least interaction with the circuit.
Nelson:
Do you find the thermal modulation of the thermistors a problem? Or do they have a positive effect due to their slightly reduced impedance while the circuit is drawing higher currents? I'm just wondering how they sound compared to straight fixed resistance in a CRC supply...
hifiZen
Thanks for the URL, I'll be playing with it today. When you say positioning of the circuits are key in avoiding the flux influence, do you have a minimum distance formula or do you prefer to just use perpendicularity. Distance would be a problem with so many caps crowding the box. Also, did you consider what value inductor would have provided 19mV of ripple? No doubt the series resistance would also be reduced with the smaller coil.
Thanks for the URL, I'll be playing with it today. When you say positioning of the circuits are key in avoiding the flux influence, do you have a minimum distance formula or do you prefer to just use perpendicularity. Distance would be a problem with so many caps crowding the box. Also, did you consider what value inductor would have provided 19mV of ripple? No doubt the series resistance would also be reduced with the smaller coil.
My favorite values for coils is about 2 mH. As far as spacing
to avoid hum pickup from the coils, it is pretty negligible
compared to what your transformer will put out, but right
angle orientation and a few inches is enough. If you can't
hear hum, then there you are.
As to Thermistor modulation, the time constant on the
devices is pretty slow and with Class A designs they have
an constant average current, so it doesn't seem to be an issue.
All we're really looking to do is filter out the ripple with this
circuit, not provide true regulation.
to avoid hum pickup from the coils, it is pretty negligible
compared to what your transformer will put out, but right
angle orientation and a few inches is enough. If you can't
hear hum, then there you are.
As to Thermistor modulation, the time constant on the
devices is pretty slow and with Class A designs they have
an constant average current, so it doesn't seem to be an issue.
All we're really looking to do is filter out the ripple with this
circuit, not provide true regulation.
Well, there you have it... i guess magnetic flux output from these coils isn't much of a problem. Now that I think about it some more, the flux will consist mostly of large steady-state component, with only a small fluctuating ripple flux. Obviously, I don't play much with CLC supplies using air core inductors... just tube amp supplies, which use iron core chokes and hence contain the flux to the core.
The coil I simulated was the 2.2mH 14awg air core inductor available from www.zalytron.com. This coil has a series resistance of 0.328ohm, and so is a little more lossy than a 0.25ohm resistor. At $12.40 each, it's about the cheapest I could find, and within the range available from Zalytron, it looked to be a pretty good cost / performance tradeoff. I only tried about 2 other coil values, but I don't think I'd bother trying to find a coil that would match the resistor's ripple at 19mV, since I'd rather have resistance in there than inductance which can oscillate with the capacitors. Lower values of inductance will increase the resonant frequency, and are likely to have less series resistance so that this oscillatory behaviour will have less damping. In all honesty, I haven't done any serious calculations or simulations of this LC resonant behaviour (nothing more than just a quick resonant frequency calculation). I'd rather just avoid this characteristic in the first place... one less thing to think about, and less complex circuit behaviour.
The coil I simulated was the 2.2mH 14awg air core inductor available from www.zalytron.com. This coil has a series resistance of 0.328ohm, and so is a little more lossy than a 0.25ohm resistor. At $12.40 each, it's about the cheapest I could find, and within the range available from Zalytron, it looked to be a pretty good cost / performance tradeoff. I only tried about 2 other coil values, but I don't think I'd bother trying to find a coil that would match the resistor's ripple at 19mV, since I'd rather have resistance in there than inductance which can oscillate with the capacitors. Lower values of inductance will increase the resonant frequency, and are likely to have less series resistance so that this oscillatory behaviour will have less damping. In all honesty, I haven't done any serious calculations or simulations of this LC resonant behaviour (nothing more than just a quick resonant frequency calculation). I'd rather just avoid this characteristic in the first place... one less thing to think about, and less complex circuit behaviour.
Hey guys!
The Aleph X prototype seems suitable to work,so far the last schematic I catched seems OK.Great job!But what is the power/8 ohms of this amp?And which is the gain ( x or dB)?
Is the P 240 suitable for the output mos.and which is the draw current of the amp?
I think we'll need big sinks out there
Cheers
Anael
The Aleph X prototype seems suitable to work,so far the last schematic I catched seems OK.Great job!But what is the power/8 ohms of this amp?And which is the gain ( x or dB)?
Is the P 240 suitable for the output mos.and which is the draw current of the amp?
I think we'll need big sinks out there
Cheers
Anael
Powerrating Aleph-X
My finished Aleph-X channels with the +/-12V rails deliver
over 20 watts in 8 Ohm and 40 watts in 4 Ohm.
Measurement done with a large dummy load.(resistive).
Draw on the powersupply is near 5 amperes all the time!
I don't realy care about these figures. To me they sound
wonderfull. They are better than my Aleph 3 clone. That's
for sure. Actually, they are better than anything I build/
bought before.
Regards
My finished Aleph-X channels with the +/-12V rails deliver
over 20 watts in 8 Ohm and 40 watts in 4 Ohm.
Measurement done with a large dummy load.(resistive).
Draw on the powersupply is near 5 amperes all the time!
I don't realy care about these figures. To me they sound
wonderfull. They are better than my Aleph 3 clone. That's
for sure. Actually, they are better than anything I build/
bought before.
Regards
People have to remember that an Aleph-X draws twice the
current, the same as the peak output current instead of
1/2 of the peak output current.
current, the same as the peak output current instead of
1/2 of the peak output current.
You want to figure on a 2+ volt loss off each rail's minimum
voltage, and so here we are peaking at about 9 volts
on each half of the amplifier.
If the amp were completely efficient and had a pure +/-12
volts to work with, the peak voltage would be 24 volts
across the load, giving 72 watts peak, and 36 watts rms.
voltage, and so here we are peaking at about 9 volts
on each half of the amplifier.
If the amp were completely efficient and had a pure +/-12
volts to work with, the peak voltage would be 24 volts
across the load, giving 72 watts peak, and 36 watts rms.
The Saint
It depends on how you build it out. A 128W rms 8 ohm version will use 25.6V rails and have a total of 16A of bias current. If you look earlier into this thread GRollins explains how to build it out to your required size.
Nelson Pass
I have just two questions about source resistors in the Aleph-X. GRollins advised us to use 0.5V/current load on each mosfet to calculate the source resistor.
Why does he use this 0.5V value to make the calcuation when the rails are 12V? Is it because the maximum input signal is only this value?
How does GRollins calculate the source resistor power rating? It would appear that the source resistor would need to be the same as the power dissipated by each mosfet but according to the circuit this doesn't seem to be the case.
It depends on how you build it out. A 128W rms 8 ohm version will use 25.6V rails and have a total of 16A of bias current. If you look earlier into this thread GRollins explains how to build it out to your required size.
Nelson Pass
I have just two questions about source resistors in the Aleph-X. GRollins advised us to use 0.5V/current load on each mosfet to calculate the source resistor.
Why does he use this 0.5V value to make the calcuation when the rails are 12V? Is it because the maximum input signal is only this value?
How does GRollins calculate the source resistor power rating? It would appear that the source resistor would need to be the same as the power dissipated by each mosfet but according to the circuit this doesn't seem to be the case.
skaara
Exactly! Just like in the XA-200 or any other of the larger amps. Nelson parallels mosfets to keep the power dissipation per device to a reasonable amount so they don't fry up. The problem you run into with paralleling mosfets in the increasing capacitance. That's why you should choose your mosfets carefully. The reason you match closely is so each parallel device is carrying its fair share of the power burden.
Exactly! Just like in the XA-200 or any other of the larger amps. Nelson parallels mosfets to keep the power dissipation per device to a reasonable amount so they don't fry up. The problem you run into with paralleling mosfets in the increasing capacitance. That's why you should choose your mosfets carefully. The reason you match closely is so each parallel device is carrying its fair share of the power burden.
I know about capacitance, and matching, but I read somewhere that there is a problem with parlelling mosfets in aleph x, but I guess it wasnt true.
Paralleling multiple output devices is one thing that we have yet to completely iron all the bugs out of in this design, it's mostly a matter of increasing the drive current and decreasing the front end impedance but this amp also seems to show stability issues under some situations so we have approached it on a case by case basis thus far.
Has anyone thought of putting a zobel network across the two O/Ps of the Aleph X. to help with stability problems..
Values of a 10 Ohm 5 watt and 100nf ~250 AC rated Cap works well for me.
In my 1kw design it was very important that this was in place.
Regards
Anthony Holton
Aussieamplifiers
Values of a 10 Ohm 5 watt and 100nf ~250 AC rated Cap works well for me.
In my 1kw design it was very important that this was in place.
Regards
Anthony Holton
Aussieamplifiers
The Saint
The issues of stability were addressed earlier in this thread. GRollins stated the circuit was rock solid and HarryHaller provided a solution to address unruly oscillations.
The issues of stability were addressed earlier in this thread. GRollins stated the circuit was rock solid and HarryHaller provided a solution to address unruly oscillations.