paulb said:
Experimentation in electronics is dangerous by itself :
- Touching the the tip of the soldering iron by accident can cause sever burnts
- Capacitors reversed by accident usually explode and can hit you
- Etching your own PCBs mens dealing with corrossive and toxic chemical products that could also produce toxic gases
- Building high-power amplifiers means dealing with high DC voltages that are more deadly than AC since when DC current passes throug the skin it gets almost glued to the metal part you are touching
- Power electronics is dangerous since components have to dissipate heat and this means heating and potential burnts when touching it [even can put the room in fire by accident]
- High power class-G amplifiers can use 280V DC peak to peak, this is almost like mains but in DC [more dangerous]
So I see no point in participating further in a forum whose philosophy seems to state that using 450VDC or designig off-line SMPS with isolation transformers are more dangerous practices than lots of other practices that are usual to the DIY world and not less dangerous [like he ones I mentioned]
Experimenting with more than 60V AC or 100V DC adds the danger of being shocked but this is not a reason to leave experimentation, it's just a reason to be more careful and take some clever additional precautions
I even think that the fact of having some forum moderators evaluating your risks and the risks of other people and deciding what is good and bad for everybody is offensive by nature, it's like if they were attempting to control your life as superior beings
People must have full information on risks and then be fully free to decide to take them or not, I see deciding for other people as a very bad and non-democratic practice
Off-line SMPS and PFC are turning into the standards, for non-indusrial applicaions of more than 200W PFC will be mandatory in Europe in the future and 50Hz rectification for more than 200W will be forbidden so people will have to change its mind and use PFC and 400..450V DC anyway, high VA 50Hz transformers may even get hard to get here in the next 10 years since it will be unlawful to load them with a bridge rectifier and some capacitors
I send a big kiss for everybody capable of understanding my point of view
Bye
Oh, I forgot to ask a simple question :
Why 100V to 800V DC on vacuum tubes are fine but 120V to 450V DC in solid state circuits is 'promoting unsafe practices'?
Following this rule, Vacuum Tube forun should be deleted as it promotes unsafe practices
Isn't it ridiculous? Doesn't it look like discriminatory practices?
Touching the wrong part of any vacuum tube amplifier or preamplifier can kill you
Why 100V to 800V DC on vacuum tubes are fine but 120V to 450V DC in solid state circuits is 'promoting unsafe practices'?
Following this rule, Vacuum Tube forun should be deleted as it promotes unsafe practices
Isn't it ridiculous? Doesn't it look like discriminatory practices?
Touching the wrong part of any vacuum tube amplifier or preamplifier can kill you
"The capacitance numbers suggested here for the A40 do not necessarily apply to all Class A electrical amps as I understand it. The amount of filter capacitance is highly dependent on the amount of current bias used in the amp. The Aleph-X built with 250 milliohm source resistors would need at least 100mF per channel to tame the ripple. Look here for more information on this topic. For corroboration, ask Nelson Pass how much capacitance is included in the AX series of amps?"
Its really quite simple. If you are building a singled ended class A amplifier like the Zen family, they have very low power supply rejection ratios. CLC filters, capacitance multipliers, and full regulation are seen with these types of circuits.
Push-pull amplifiers with even modest amounts of feedback have high power supply rejection ratios. The A40 falls in this category.
K-amps is asking about a 750W amplifier. Me thinks it will not be single ended class A.
3,300µF is a common value for a capacitor, the next value up is usually 3,900µF. For 60hz and an 8 ohm loudspeaker 3,300µF gives better than 90% regulation. For 50hz you would want the 3,900µF cap. Bigger than this really is beyond the point of diminishing returns in my mind.
For stereo that would be 6,800µF (rounding to closest standard values). For a stereo amp rated for 4 ohms that would be 13,500µF (guess what size caps are in a Crown DC300A?).
A Crest 6000 has 120,000µF. That sounds like a lot, doesn't it? It is a stereo amplifier, so that is 60,000µF per channel. It has two ± voltage tiers, so that is four 15,000µF caps per channel. It is rated to drive 2 ohms. It is rated to run on 50hz. By our rule of thumb it should have 3,900µF for 8 ohms, 7,800µF for 4 ohms, or 15,600µF for 2 ohms. Per rail, per channel. It has four voltage rails and two channels, so it needs 15,600µF X 4 X 2 = 124,800µF of filter capacitance.
The bare minimum by our rule of thumb.
If you were building a stereo amplifier to run a 2 ohm load from a shared (mono) supply with a single ±V rails then a pair of caps in the range of 27,000µF~33,000µF are in order.
Its really quite simple. If you are building a singled ended class A amplifier like the Zen family, they have very low power supply rejection ratios. CLC filters, capacitance multipliers, and full regulation are seen with these types of circuits.
Push-pull amplifiers with even modest amounts of feedback have high power supply rejection ratios. The A40 falls in this category.
K-amps is asking about a 750W amplifier. Me thinks it will not be single ended class A.
3,300µF is a common value for a capacitor, the next value up is usually 3,900µF. For 60hz and an 8 ohm loudspeaker 3,300µF gives better than 90% regulation. For 50hz you would want the 3,900µF cap. Bigger than this really is beyond the point of diminishing returns in my mind.
For stereo that would be 6,800µF (rounding to closest standard values). For a stereo amp rated for 4 ohms that would be 13,500µF (guess what size caps are in a Crown DC300A?).
A Crest 6000 has 120,000µF. That sounds like a lot, doesn't it? It is a stereo amplifier, so that is 60,000µF per channel. It has two ± voltage tiers, so that is four 15,000µF caps per channel. It is rated to drive 2 ohms. It is rated to run on 50hz. By our rule of thumb it should have 3,900µF for 8 ohms, 7,800µF for 4 ohms, or 15,600µF for 2 ohms. Per rail, per channel. It has four voltage rails and two channels, so it needs 15,600µF X 4 X 2 = 124,800µF of filter capacitance.
The bare minimum by our rule of thumb.
If you were building a stereo amplifier to run a 2 ohm load from a shared (mono) supply with a single ±V rails then a pair of caps in the range of 27,000µF~33,000µF are in order.
EVA,
I can understand you point fully. And while no one shold stifle the experimentation and progress of anyone else, you should just mention the risks involved and that should be that.
As long as you are not deliberately making people do unsafe things, you are ok... we afte rall deal with electricity all the time.
I can understand you point fully. And while no one shold stifle the experimentation and progress of anyone else, you should just mention the risks involved and that should be that.
As long as you are not deliberately making people do unsafe things, you are ok... we afte rall deal with electricity all the time.
DJK,
So let me get this straight: Assuming I am bulding a Push-Pull design,
The amount of class-A operation would depend on:
1) The Rail voltages used
2) The bias voltage drop across each emitter resistor in the output (or idle bias)
3) Number of devices used.
So it means that if I have 90 volt rails and 0.132v across each 0.22ohm resistor (0.6amps bias), I can get 200 watts class-A using 12 devices (6 pairs)
1) IS THIS ACCURATE?
2) Also does it mean that keeping the bias voltage constant and adding more devices will give me more class-A output?
I guess I am still fuzzy about how much bias I need for a 200watt class-A amp. In another thread a guy told me that 6 pairs with 0.132v across the emmitter resistor should do it. This equates to about 315 watts dissipation.
However reading the A40 article led me to believe that for every watt of class-A needed, 2.5 watts needs to be dissipated as heat.
Using that and using a linear extrapolation, I need to dissipate 500 watts to get 200 watts class-A and not 315 watts.
3) Which is correct? 315 or 500?
4) Lastly When we generally talk about the bias (in amps) needed for a given class-A output, do we mean the bias per complementary pair of output devices or the combined total of all output devices?
thanks in advance!!!!
So let me get this straight: Assuming I am bulding a Push-Pull design,
The amount of class-A operation would depend on:
1) The Rail voltages used
2) The bias voltage drop across each emitter resistor in the output (or idle bias)
3) Number of devices used.
So it means that if I have 90 volt rails and 0.132v across each 0.22ohm resistor (0.6amps bias), I can get 200 watts class-A using 12 devices (6 pairs)
1) IS THIS ACCURATE?
2) Also does it mean that keeping the bias voltage constant and adding more devices will give me more class-A output?
I guess I am still fuzzy about how much bias I need for a 200watt class-A amp. In another thread a guy told me that 6 pairs with 0.132v across the emmitter resistor should do it. This equates to about 315 watts dissipation.
However reading the A40 article led me to believe that for every watt of class-A needed, 2.5 watts needs to be dissipated as heat.
Using that and using a linear extrapolation, I need to dissipate 500 watts to get 200 watts class-A and not 315 watts.
3) Which is correct? 315 or 500?
4) Lastly When we generally talk about the bias (in amps) needed for a given class-A output, do we mean the bias per complementary pair of output devices or the combined total of all output devices?
thanks in advance!!!!
look at the A-75 article, that's the one that made it clear for me - theoretical push pull efficiency approaches 50%.
P=I*I*R => Irms=SQRT(200/8)=5A,
Ipeak=1.414*Irms=7.07
Push pull can deliver twice its idle current in class A=> bias required=3.535 A
Power dissipated per rail = Idle current * rail voltage= 318w/ rail
Total dissipation= 636 W per channel for 200 Wrms into 8 ohms, 100 Wrms into 4 (with your 90 volt rails to get 200 class A watts into 4 ohms, you'd need to increase the bias to 5 A, dissipating 900 watts.)
Lower rail voltage would reduce dissipation - 200Wrms peak swing is around 57 volts at 8 ohms, add 5 volts or so for the drop in the output devices and your dissipation drops to 438W, but you lose the ability to exceed 200 W in class AB.
P=I*I*R => Irms=SQRT(200/8)=5A,
Ipeak=1.414*Irms=7.07
Push pull can deliver twice its idle current in class A=> bias required=3.535 A
Power dissipated per rail = Idle current * rail voltage= 318w/ rail
Total dissipation= 636 W per channel for 200 Wrms into 8 ohms, 100 Wrms into 4 (with your 90 volt rails to get 200 class A watts into 4 ohms, you'd need to increase the bias to 5 A, dissipating 900 watts.)
Lower rail voltage would reduce dissipation - 200Wrms peak swing is around 57 volts at 8 ohms, add 5 volts or so for the drop in the output devices and your dissipation drops to 438W, but you lose the ability to exceed 200 W in class AB.
K-amps
When I talk about bias I use the following conventions:
bias: the amount of Class A current bias through each device (includes signal and source devices)
amp bias: the amount of Class A current bias through each channel
total bias: the amount of amp bias times the number of channels
In the case of the original Aleph-X, an amp bias of 5A carried by the four devices would equal 1.25A of bias at full power. I have noted that there is usually alot of confusion defining "bias" so if anyone has a better way to define what we mean, please post it here so all the membership can benefit.
When I talk about bias I use the following conventions:
bias: the amount of Class A current bias through each device (includes signal and source devices)
amp bias: the amount of Class A current bias through each channel
total bias: the amount of amp bias times the number of channels
In the case of the original Aleph-X, an amp bias of 5A carried by the four devices would equal 1.25A of bias at full power. I have noted that there is usually alot of confusion defining "bias" so if anyone has a better way to define what we mean, please post it here so all the membership can benefit.
A push-pull output stage biased into class A can approach 50% efficency. This assumes a fully regulated supply and no losses in the output stages.
In reality it will never happen.
Ever figure out the losses in a fully regulated linear supply?
Ever measure how close your amplifier can drive to the rail?
Losses of 10V~13V are common.
Ballpark, I figure a 100W class A amplifier while theoretically needing only 200W will actually pull closer to 285W (200W X 1.414). That allows for power supply ripple, emitter resistor losses, CE voltage drops, etc.
In reality it will never happen.
Ever figure out the losses in a fully regulated linear supply?
Ever measure how close your amplifier can drive to the rail?
Losses of 10V~13V are common.
Ballpark, I figure a 100W class A amplifier while theoretically needing only 200W will actually pull closer to 285W (200W X 1.414). That allows for power supply ripple, emitter resistor losses, CE voltage drops, etc.
1) So by that virtue will a 200 watt class-A amp pull 570watts?
Also, I have this unanswered question:
Assuming I have my idling bias pot set to give me 0.6amps across each PNP (or NPN) device. Lets assume that I an using 6 pairs, that would make it 3.6 amps.
Now if I keep paralelling more output devices, will I get more class-A output into the same 8 ohm load? (assuming the voltage drop across the emitter resistors stays constant.
Thanks!
Also, I have this unanswered question:
Assuming I have my idling bias pot set to give me 0.6amps across each PNP (or NPN) device. Lets assume that I an using 6 pairs, that would make it 3.6 amps.
Now if I keep paralelling more output devices, will I get more class-A output into the same 8 ohm load? (assuming the voltage drop across the emitter resistors stays constant.
Thanks!
K-amps
Yes, that is exactly how it works. The amp bias can be increased by adding devices when the desired bias is greater than the power dissipation capability of the device. Keep in mind that there is sonic detriment on the high frequencies when you start paralleling devices because you increase the capacitance of the devices on the signal.
Yes, that is exactly how it works. The amp bias can be increased by adding devices when the desired bias is greater than the power dissipation capability of the device. Keep in mind that there is sonic detriment on the high frequencies when you start paralleling devices because you increase the capacitance of the devices on the signal.
K-amps
No, as I understand it, the capacitance in the device gate attenuates the high frequency signals no matter what. I think 1nF is about as high as any audiophile will want to go but I'm not the last word on this. It really depends on where you want to cut off your high end.
AudioFreak
Will you kindly edit out the faulty link in this post? Thanks.
No, as I understand it, the capacitance in the device gate attenuates the high frequency signals no matter what. I think 1nF is about as high as any audiophile will want to go but I'm not the last word on this. It really depends on where you want to cut off your high end.
AudioFreak
Will you kindly edit out the faulty link in this post? Thanks.
There Is A Bigger Picture Than 200W............
Millwood, not as ridiculous as you may think.
"Off-line SMPS and PFC are turning into the standards, for non-indusrial applicaions of more than 200W PFC will be mandatory in Europe in the future and 50Hz rectification for more than 200W will be forbidden so people will have to change its mind and use PFC and 400..450V DC anyway, high VA 50Hz transformers may even get hard to get here in the next 10 years since it will be unlawful to load them with a bridge rectifier and some capacitors"
Audio amplifiers are just about the last bastion of domestic appliances using large transformers.
Just about EVERYTHING else in the average modern home is running from a SMPS.
The ultimate in efficiency audio device that I can think of is modern DVD players with SMPS and Digital Amplifiers in-built.
SMPS and PFC have the advantages in lower power consumption (higher efficiency) which equals lower energy costs for the consumer, and for the power company lower transmission losses, better voltage regulation and less carbon dioxide emissions into the environment.
Eric.
Millwood, not as ridiculous as you may think.
"Off-line SMPS and PFC are turning into the standards, for non-indusrial applicaions of more than 200W PFC will be mandatory in Europe in the future and 50Hz rectification for more than 200W will be forbidden so people will have to change its mind and use PFC and 400..450V DC anyway, high VA 50Hz transformers may even get hard to get here in the next 10 years since it will be unlawful to load them with a bridge rectifier and some capacitors"
Audio amplifiers are just about the last bastion of domestic appliances using large transformers.
Just about EVERYTHING else in the average modern home is running from a SMPS.
The ultimate in efficiency audio device that I can think of is modern DVD players with SMPS and Digital Amplifiers in-built.
SMPS and PFC have the advantages in lower power consumption (higher efficiency) which equals lower energy costs for the consumer, and for the power company lower transmission losses, better voltage regulation and less carbon dioxide emissions into the environment.
Eric.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.