Don't get too excited about transformer VA specs, as the
power factor comes in to make simple calculations vague
anyway. Just see to it that the transformer has about twice
what you think you need. The easiest way to throw away some
voltage is to make the loss before the rectifier bridges with
a thermistor or coil, but save some for the CRC.
Now a CRCRCRC is nicely over the top 😎 , but really, a CRC
or CLC will actually do the job. CLC's work great, but I like the
CRC's because 1) they are cheaper 2) they have no resonance
3) no mechanical noise and 4) no magnetic field to be picked up.
On the other hand, they are slightly more efficient.
power factor comes in to make simple calculations vague
anyway. Just see to it that the transformer has about twice
what you think you need. The easiest way to throw away some
voltage is to make the loss before the rectifier bridges with
a thermistor or coil, but save some for the CRC.
Now a CRCRCRC is nicely over the top 😎 , but really, a CRC
or CLC will actually do the job. CLC's work great, but I like the
CRC's because 1) they are cheaper 2) they have no resonance
3) no mechanical noise and 4) no magnetic field to be picked up.
On the other hand, they are slightly more efficient.
CRCRCRC
Terry, looking at your prior power supplies, they look like CRC, not CRCRCRC like you say they are. Looks like you got a board that does a CRC x 2.
Terry, looking at your prior power supplies, they look like CRC, not CRCRCRC like you say they are. Looks like you got a board that does a CRC x 2.
Re: CRCRCRC
You're right. I've never used a CRCRCRC. I just used that as an example of how to maybe drop the voltage so I could find out if the VA would drop also. I'm trying to find a way to power this amp without having to have a custom transformer. I bought a pair of 22-0-22 400vA. I'm trying to see what I can do to make them work. My problem is that when folks answer my questions I usually understand them at first and then they start using terms I'm ignorant to and they lose me, or they give rather a vague answer that makes sense to someone who already has a grasp on the subject but still doesn't give me a simple enough explanation for me to get it.
An example is when Nelson said "Just see to it that the transformer has about twice
what you think you need. "
My problem is that I don't have a clue how much I need so "twice" means nothing to me. I guess that's why the thread suggests that this amp is not good for us beginners. I understood someone to say that the 16-0-16 300vA transformer I asked about was too small. It calcs out to 9.375A. Yet today someone suggested to use a 20-0-20 400va which only calcs out to be 10A. I feel like I'm chasing my tail. I haven't heard anyone say use a transformer with X amount of voltage and X amount of va and it will work well. I'm not trying to build a Lamborghini. I just want a good sounding, reliable amp. And mostly I want to learn.
I’m going to keep plugging along here but boy, if it’s this tough just trying to get the PSU together, I wonder what it’s going to be like to build the rest
Blessings, Terry
lgreen said:
You're right. I've never used a CRCRCRC. I just used that as an example of how to maybe drop the voltage so I could find out if the VA would drop also. I'm trying to find a way to power this amp without having to have a custom transformer. I bought a pair of 22-0-22 400vA. I'm trying to see what I can do to make them work. My problem is that when folks answer my questions I usually understand them at first and then they start using terms I'm ignorant to and they lose me, or they give rather a vague answer that makes sense to someone who already has a grasp on the subject but still doesn't give me a simple enough explanation for me to get it.
An example is when Nelson said "Just see to it that the transformer has about twice
what you think you need. "
My problem is that I don't have a clue how much I need so "twice" means nothing to me. I guess that's why the thread suggests that this amp is not good for us beginners. I understood someone to say that the 16-0-16 300vA transformer I asked about was too small. It calcs out to 9.375A. Yet today someone suggested to use a 20-0-20 400va which only calcs out to be 10A. I feel like I'm chasing my tail. I haven't heard anyone say use a transformer with X amount of voltage and X amount of va and it will work well. I'm not trying to build a Lamborghini. I just want a good sounding, reliable amp. And mostly I want to learn.
I’m going to keep plugging along here but boy, if it’s this tough just trying to get the PSU together, I wonder what it’s going to be like to build the rest
Blessings, Terry
I'm still not clear on why this is so difficult. See my previous post.
Okay, let's try to shotgun approach.
The caveat here is that I haven't gone back and read the last ten pages of this thread. Not having done so, perhaps I missed something. My limitation is that this PC is on its last legs and I don't stay online very long.
If I read some of these posts correctly, it seems that you're trying to reduce the VA of a transformer. You can't. It's built to have a certain VA. Now, if it's the rail voltage you're trying to reduce, that can be done. If it's the current draw you're trying to reduce, you redesign the circuit.
The product of the rail voltage at the circuit, times the current it draws gives you the power dissipation in watts, aka Pd, or Power Dissipation.
On the subject of rail voltage reduction via CRC filtration, the objective is to use smallish resistors. Depending on the current draw of the circuit you might want values on the order of half an Ohm or so. Certainly no more than one or two Ohms.
Make sure to use power resistors. You're going to be getting hot. 20W ceramic resistors are cheap. Buy a handful and use them in series, parallel...whatever works.
Getting in a twist over the exact rail voltage isn't necessary. The circuit will not break. You can run the same circuit at 18V or at 25V. The only thing you need to keep an eye on is heatsinking. Trying to split hairs between 24 and 25V rails is going to make you crazy. Build the amp. Put the thing on your kitchen table and measure the pertinent voltages and currents. Then hook it up and listen to it. While listening, pull out the notes you made on the voltages and currents in the circuit and contemplate them calmly.
While it's laudable to try to understand what's going on in microscopic detail, it's also possible to over-analyze things and paralyze yourself with fear. Emotion clouds the mind. When the mind is clouded, it cannot understand. This is the exact opposite of what you're trying to do.
Or, as Charlie Papazian is fond of saying: Relax, don't worry. Have a homebrew.
There is a line from Shakespeare:
"There is nothing either good or bad, but thinking makes it so." (Hamlet)
I have heard this line reworded to say,"Nothing is hard, but thinking makes it so." I like this phrase, even though it's not the precise quote, because it's so true.
Making allowances for the changes in the use of the English language since Bill set quill to parchment, it translates as meaning that you've decided that it's hard, so it is. The solution is to quit telling yourself how hard it is and open yourself to how easy it is.
Grey
P.S.: There's nothing about class A amps that makes them any more difficult than class AB, B, C, D, etc. Barring a project that specifies a particular transformer in the parts list, you're still going to have to sit down and make some decisions as to what part to buy when it comes time to assemble the power supply.
Okay, let's try to shotgun approach.
The caveat here is that I haven't gone back and read the last ten pages of this thread. Not having done so, perhaps I missed something. My limitation is that this PC is on its last legs and I don't stay online very long.
If I read some of these posts correctly, it seems that you're trying to reduce the VA of a transformer. You can't. It's built to have a certain VA. Now, if it's the rail voltage you're trying to reduce, that can be done. If it's the current draw you're trying to reduce, you redesign the circuit.
The product of the rail voltage at the circuit, times the current it draws gives you the power dissipation in watts, aka Pd, or Power Dissipation.
On the subject of rail voltage reduction via CRC filtration, the objective is to use smallish resistors. Depending on the current draw of the circuit you might want values on the order of half an Ohm or so. Certainly no more than one or two Ohms.
Make sure to use power resistors. You're going to be getting hot. 20W ceramic resistors are cheap. Buy a handful and use them in series, parallel...whatever works.
Getting in a twist over the exact rail voltage isn't necessary. The circuit will not break. You can run the same circuit at 18V or at 25V. The only thing you need to keep an eye on is heatsinking. Trying to split hairs between 24 and 25V rails is going to make you crazy. Build the amp. Put the thing on your kitchen table and measure the pertinent voltages and currents. Then hook it up and listen to it. While listening, pull out the notes you made on the voltages and currents in the circuit and contemplate them calmly.
While it's laudable to try to understand what's going on in microscopic detail, it's also possible to over-analyze things and paralyze yourself with fear. Emotion clouds the mind. When the mind is clouded, it cannot understand. This is the exact opposite of what you're trying to do.
Or, as Charlie Papazian is fond of saying: Relax, don't worry. Have a homebrew.
There is a line from Shakespeare:
"There is nothing either good or bad, but thinking makes it so." (Hamlet)
I have heard this line reworded to say,"Nothing is hard, but thinking makes it so." I like this phrase, even though it's not the precise quote, because it's so true.
Making allowances for the changes in the use of the English language since Bill set quill to parchment, it translates as meaning that you've decided that it's hard, so it is. The solution is to quit telling yourself how hard it is and open yourself to how easy it is.
Grey
P.S.: There's nothing about class A amps that makes them any more difficult than class AB, B, C, D, etc. Barring a project that specifies a particular transformer in the parts list, you're still going to have to sit down and make some decisions as to what part to buy when it comes time to assemble the power supply.
I Grey,
Thanks for taking the time to write such a thoughtful response.
I'm not really stuck on reducing the voltage of a transformer so much as finding the right transformer to begin with. Why this voltage reduction thing got started is because I suggested a couple of transformers that I might like to use and for the most part I was told that they were not good choices. Seems that the choice that most feel would work is something like 18-0-18VAC 800va. I can't find one of those. So I thought I'd ask some questions so I could better understand how transformers work and why this particular amp seems to need way more vA than anything else I've built. Since I couldn't find any transformers with that low of voltage and that high of a vA rating, I thought I'd look at the possibility of using a higher voltage/vA rated transformer and reduce the voltage so I would have enough to make this amp work into a 4ohm load.
You say the vA is fixed in a transformer, yet through the answers I've gotten in this thread it seems that if you lower the voltage you also lower the vA since it is a result of the volt * amps. If I'm wrong about this then I'm still really screwed up about the whole thing.
I don't want to split hairs about anything. You say the voltage doesn't matter and it won't break. Does this mean I can run it at +/-30V rails? Am I better off with a pair of 16-016VAC 300vA transformers or a pair of 22-0-22VAC 400vA transformers? My plan was to get the 22-0-22 and reduce the voltage because it was a 400vA, but if the vA is going to reduce as the voltage reduces then I may not gain anything and will just waste energy.
The only reason I'm going through all of this is because it seems that I have to know what power I'm using before I can work out the rest of the BOM.
Blessings, Terry
Thanks for taking the time to write such a thoughtful response.
I'm not really stuck on reducing the voltage of a transformer so much as finding the right transformer to begin with. Why this voltage reduction thing got started is because I suggested a couple of transformers that I might like to use and for the most part I was told that they were not good choices. Seems that the choice that most feel would work is something like 18-0-18VAC 800va. I can't find one of those. So I thought I'd ask some questions so I could better understand how transformers work and why this particular amp seems to need way more vA than anything else I've built. Since I couldn't find any transformers with that low of voltage and that high of a vA rating, I thought I'd look at the possibility of using a higher voltage/vA rated transformer and reduce the voltage so I would have enough to make this amp work into a 4ohm load.
You say the vA is fixed in a transformer, yet through the answers I've gotten in this thread it seems that if you lower the voltage you also lower the vA since it is a result of the volt * amps. If I'm wrong about this then I'm still really screwed up about the whole thing.
I don't want to split hairs about anything. You say the voltage doesn't matter and it won't break. Does this mean I can run it at +/-30V rails? Am I better off with a pair of 16-016VAC 300vA transformers or a pair of 22-0-22VAC 400vA transformers? My plan was to get the 22-0-22 and reduce the voltage because it was a 400vA, but if the vA is going to reduce as the voltage reduces then I may not gain anything and will just waste energy.
The only reason I'm going through all of this is because it seems that I have to know what power I'm using before I can work out the rest of the BOM.
Blessings, Terry
Hi Terry,
When I built my mini aleph I had a couple of small toroids that had 21v secondaries. That was a bit high so I simply un-wound several turns off of the secondaries until I found ~15v..just right for me. If you remove 20% or so from your 22-0-22 400vA trannys I bet you'de be somewhere in your ballpark. It's a bit risky because I'm sure things could get mussed up somehow, but it has worked for me.
Another way I have heard of reducing v is to create windings in the opposite direction that the tranny is currently wound. Seems to me you may lose a few VA (increased resistance on the secondaries??) but you wouldn't be tearing the toroids down. I've never tried it but I'de bet a buck it would work fine...
Take it sorta with a grain of salt, I'm still sort of an amp newbee.
When I built my mini aleph I had a couple of small toroids that had 21v secondaries. That was a bit high so I simply un-wound several turns off of the secondaries until I found ~15v..just right for me. If you remove 20% or so from your 22-0-22 400vA trannys I bet you'de be somewhere in your ballpark. It's a bit risky because I'm sure things could get mussed up somehow, but it has worked for me.
Another way I have heard of reducing v is to create windings in the opposite direction that the tranny is currently wound. Seems to me you may lose a few VA (increased resistance on the secondaries??) but you wouldn't be tearing the toroids down. I've never tried it but I'de bet a buck it would work fine...
Take it sorta with a grain of salt, I'm still sort of an amp newbee.
I've posted this multiple times, but here we go again.
Grey's recipe for Alephs and Aleph-Xs
Serves...well, who cares how many it serves...just sit some people down and listen.
The following recipe assumes monoblock construction. If you want to build a stereo amp, double the VA of the transformer.
STEP ONE: Decide on a target RMS wattage into 8 Ohms. I can't help you with this. It's your decision. Some things just can't be simplified any further. Life's a bummer. Get over it.
Just for giggles, let's say 50W.
STEP TWO: Determine what voltage and current this represents. This is the IR=E part. You know...higher math.
One moment, please, whilst I consult the calculator gods...
Okay, looks like that's 20Vrms and 2.5A.
STEP THREEa:
If building an Aleph, your target rail voltage will be 1.6 times this voltage or +-32V.
STEP THREEb:
If building an Aleph-X, your target rail voltage is +-16V, or about .8 times the output voltage.
STEP FOURa:
If building an Aleph, the output stage will be biased for a minimum of the target 2.5A output.
STEP FOURb:
If building an Aleph-X, the amp as a whole will be biased for a minimum of twice the desired current, or 5A.
N.B.: If you look at the numbers, you'll notice that the Pd of the amp circuit is the same whether you're building an Aleph or an Aleph-X. It's either 2*32V*2.5A=160W or 2*16V*5A=160W. This makes sense because both circuits have the same efficiency, i.e. roughly 1/3 of the Pd is available as output wattage. (If you'll consult my earlier post, you'll see that I said to multiply the desired output wattage * 3 to get the Pd. Now I've gone and derived it in a somewhat more complicated manner, but the original post still stands.)
STEP FIVE: Multiply the circuit Pd * 2 to get the necessary VA for the transformer. This is because you don't want to run the transformer at the limit. Since we're expecting Pd=160W, the answer here is 320VA.
STEP SIXa:
For an Aleph, multiply the rail by .75. This comes out to 24Vac. (This is one rail...you'll need a 24-0-24V transformer.)
STEP SIXb:
For an Aleph-X, multiply the rail by .75. The magic answer here is 12Vac, so you'll need a 12-0-12V transformer.
STEP SEVEN: Now, add a volt or two for luck. When in doubt, round up to the next larger size, be it voltage, current, or VA.
Man, I don't know how I can make it any simpler. You can complicate this in all sorts of ways, but I'll be damned if I can think of a way to make it simpler, short of choosing your target wattage for you. I've posted this recipe before (albeit sometimes with slightly different numbers), but people insist on complicating things. 800VA for an 18-0-18V transformer? Nah. Not unless you want to be really, really conservative. Or unless you want to run a 4 Ohm load, in which case you'll need to double the VA (to 640VA) and increase the bias in the circuit to handle the load. If you want a stereo amp and 4 Ohm loads, you're going to have to multiply by 4, but you're going to have a hard time finding a low voltage transformer with that much current.
I've pressed my luck running this PC this long. I need to get off and let the poor thing rest.
Run the numbers. If you want to muddy the waters by throwing in Vgs and all that, it's fine by me, but you'll find that you come up with numbers that are startlingly close to mine. True, you'll have a deeper understanding of where these numbers came from, but this way will get you to a functioning amplifier faster.
Good luck.
Grey
Grey's recipe for Alephs and Aleph-Xs
Serves...well, who cares how many it serves...just sit some people down and listen.
The following recipe assumes monoblock construction. If you want to build a stereo amp, double the VA of the transformer.
STEP ONE: Decide on a target RMS wattage into 8 Ohms. I can't help you with this. It's your decision. Some things just can't be simplified any further. Life's a bummer. Get over it.
Just for giggles, let's say 50W.
STEP TWO: Determine what voltage and current this represents. This is the IR=E part. You know...higher math.
One moment, please, whilst I consult the calculator gods...
Okay, looks like that's 20Vrms and 2.5A.
STEP THREEa:
If building an Aleph, your target rail voltage will be 1.6 times this voltage or +-32V.
STEP THREEb:
If building an Aleph-X, your target rail voltage is +-16V, or about .8 times the output voltage.
STEP FOURa:
If building an Aleph, the output stage will be biased for a minimum of the target 2.5A output.
STEP FOURb:
If building an Aleph-X, the amp as a whole will be biased for a minimum of twice the desired current, or 5A.
N.B.: If you look at the numbers, you'll notice that the Pd of the amp circuit is the same whether you're building an Aleph or an Aleph-X. It's either 2*32V*2.5A=160W or 2*16V*5A=160W. This makes sense because both circuits have the same efficiency, i.e. roughly 1/3 of the Pd is available as output wattage. (If you'll consult my earlier post, you'll see that I said to multiply the desired output wattage * 3 to get the Pd. Now I've gone and derived it in a somewhat more complicated manner, but the original post still stands.)
STEP FIVE: Multiply the circuit Pd * 2 to get the necessary VA for the transformer. This is because you don't want to run the transformer at the limit. Since we're expecting Pd=160W, the answer here is 320VA.
STEP SIXa:
For an Aleph, multiply the rail by .75. This comes out to 24Vac. (This is one rail...you'll need a 24-0-24V transformer.)
STEP SIXb:
For an Aleph-X, multiply the rail by .75. The magic answer here is 12Vac, so you'll need a 12-0-12V transformer.
STEP SEVEN: Now, add a volt or two for luck. When in doubt, round up to the next larger size, be it voltage, current, or VA.
Man, I don't know how I can make it any simpler. You can complicate this in all sorts of ways, but I'll be damned if I can think of a way to make it simpler, short of choosing your target wattage for you. I've posted this recipe before (albeit sometimes with slightly different numbers), but people insist on complicating things. 800VA for an 18-0-18V transformer? Nah. Not unless you want to be really, really conservative. Or unless you want to run a 4 Ohm load, in which case you'll need to double the VA (to 640VA) and increase the bias in the circuit to handle the load. If you want a stereo amp and 4 Ohm loads, you're going to have to multiply by 4, but you're going to have a hard time finding a low voltage transformer with that much current.
I've pressed my luck running this PC this long. I need to get off and let the poor thing rest.
Run the numbers. If you want to muddy the waters by throwing in Vgs and all that, it's fine by me, but you'll find that you come up with numbers that are startlingly close to mine. True, you'll have a deeper understanding of where these numbers came from, but this way will get you to a functioning amplifier faster.
Good luck.
Grey
Hi Grey,
Thanks for explaining once more. That is very helpful.
I am plannig on using multiple output devices. On the BOM it says to match Q2 and Q11. May I assume that when adding multiple outputs that they parallel with each of the original 4?
Does this mean that if using a total of 12 outputs per channel that all three @ Q2 and all 3 @ Q11 need to be matched?
Thanks, Terry
Thanks for explaining once more. That is very helpful.
I am plannig on using multiple output devices. On the BOM it says to match Q2 and Q11. May I assume that when adding multiple outputs that they parallel with each of the original 4?
Does this mean that if using a total of 12 outputs per channel that all three @ Q2 and all 3 @ Q11 need to be matched?
Thanks, Terry
I am considering an AX with 14-15v rail and 4.5-5A bias per channel. What is a good source for a 500-600W toroidal transfomer with 12+12 secondaries? Plitron and Avel Lindberg have the voltage only at lower power ratings.
Look at the schematic and divide it into 4 quadrants. Left and right Aleph current sources (the top MOSFETs). Left and right outputs (the bottom MOSFETs). Strictly speaking, you need to match the MOSFETs in each quadrant. It's a good idea to match all the devices across the top and all the devices across the bottom--or at least use two matched groups with similar numbers. It will make the DC offset easier to deal with.
Transformers? Try surplus places.
Why are you trying for so much current at such a low voltage? If you're not trying to drive a really, really low impedance, you can get by with less bias. If you're trying to build a stereo amp, consider two transformers, even if you mount them in the same chassis.
When all else fails, you can use two transformers in series, one for V+ and one for V-. Another alternative is to parallel two identical transformers for more current.
Or you can use four transformers, two in parallel for the V+, wired in series with two transformers in parallel for V-...
Grey
Transformers? Try surplus places.
Why are you trying for so much current at such a low voltage? If you're not trying to drive a really, really low impedance, you can get by with less bias. If you're trying to build a stereo amp, consider two transformers, even if you mount them in the same chassis.
When all else fails, you can use two transformers in series, one for V+ and one for V-. Another alternative is to parallel two identical transformers for more current.
Or you can use four transformers, two in parallel for the V+, wired in series with two transformers in parallel for V-...
Grey
I am trying to get about 50W into 4 ohms, which requires about 2.3 to 2.5 A per side (4.5-5A per channel). I can certainly use two transformers but I am not sure about paralelling the two secondaries in a transformer as I have no idea how close their voltages are.
GRollins said:
huh, I wonder what the chances are that I can buy 12 fets and be able to make matched quadrants out of them?
Probably two, slim and none. 🙁
I wonder if we could get enough folks together to buy a big amount so that we could make a few matched sets?
Maybe I'm a Johnny come lately.
Thanks for the help.
Blessings, Terry
I bought a bunch of IRF644s--like 100 of them--and found them to be pretty scattered. On the other hand, I bought, I think, 20 IRFP044, and they were all pretty tightly grouped.
I don't know whether it's a quality control problem or something that varies by device. It may be that the IRF644s are normally all over the map, whereas IRFP044s are more consistent just by nature.
Most people seem to get by buying about 50% more than they need. I buy double, or to the next price break, depending on my mood and whether I think I'll be able to use the device in other projects later on. It also gives me a larger pool to match from. This is all to the good.
As long as you don't subject them to static, MOSFETs have virtually infinite shelf life. Stick them on the shelf. You never know when you'll need spares or suddenly get seized with the desire to build something late on a Saturday night when there's no hope of buying anything faintly like electronic parts. Happens to me all the time.
Grey
I don't know whether it's a quality control problem or something that varies by device. It may be that the IRF644s are normally all over the map, whereas IRFP044s are more consistent just by nature.
Most people seem to get by buying about 50% more than they need. I buy double, or to the next price break, depending on my mood and whether I think I'll be able to use the device in other projects later on. It also gives me a larger pool to match from. This is all to the good.
As long as you don't subject them to static, MOSFETs have virtually infinite shelf life. Stick them on the shelf. You never know when you'll need spares or suddenly get seized with the desire to build something late on a Saturday night when there's no hope of buying anything faintly like electronic parts. Happens to me all the time.
Grey
Ctong,
you maybe interested in this:
http://www.diyaudio.com/forums/showthread.php?s=&postid=732081#post732081
you maybe interested in this:
http://www.diyaudio.com/forums/showthread.php?s=&postid=732081#post732081
I bought 50 fairchild IRFP244s and had was able to getting 8 sets of 6 well within .01V. If you buy in standard package lots (like tubes of 25, check the mfr. site for this info) you're more likely to get devices from the same wafer, which should give you good matches.
Remember, 4 quadrant matching is ideal, but there have been many AXs successfully built looking at just matching across the top and bottom.
Remember, 4 quadrant matching is ideal, but there have been many AXs successfully built looking at just matching across the top and bottom.
Hi,
I matched the bottom 6 and two groups of three for the current sources. Bought 40 fets and easily got the 24 fets needed.
William
I matched the bottom 6 and two groups of three for the current sources. Bought 40 fets and easily got the 24 fets needed.
William
Has anyone tried using the MTW32N20E/D in place of the IRFP204's? they seem to have similar specs.
Thanks, Terry
Thanks, Terry
still4given said:
The capacitance seems a bit high on them, but at the same time
they dissipate more and they have higher current capacity.
I would try to get away with using as few as possible - I think
you could idle them at say 50-60 watts if you give them enough
sinking.
😎
Nelson Pass said:
Hi Nelson,
Thanks for looking this over. Are you saying that these would not be limited to the 20-30 watts that the 204's and 004's are?
I am planning to use a 4 X 12v 1.2KvA transformer. According to the AXE-1.2 chart I could go as high as 8A bias with only 4 FET's per channel. Am I reading this right?
Thanks, Terry