still4given said:
Terry,
those transformers are not suitable for an AX - the voltage is high enough to make yourself a 200+W amp, but the idle dissipation on such a beast will be over 600W.
The 16 volt transformers you mentioned earlier will do for a moderately biased AX - not really enough for your target of 4 ohms unless you use a pair per channel. They ought to be OK for about 50W/8ohms, though.
Look for the AXE-1-2 spreadsheet - it will help you plan your transformer, heat sink and output stage requirements. You'll need at least 8 fets to make 50W into 4 ohms on huge heat sinks - probably 4 times what you used on the Krell.
Hi Bob,
I guess I don't understand Class A well enough. I'm not sure that I need to have 50W Class A into 4 ohms. I just want to be able to hook up a 4 ohm load to it. My nearfields monitors are 4 ohm and I would like the option of useing it for those if I want. Why is it that the AX needs so much less voltage and so much more heatsink than the Krell? I'm not getting the picture some how.
I kind of got into this AX amp a little late I guess. A friend gave me the boards and I'm trying to get up to speed. If my 25-0-25 trannies won't work then I would prefer to buy the best ones for the job instead of just buying what's easily available. Can you suggest a size for me to use?
Thanks, Terry
I guess I don't understand Class A well enough. I'm not sure that I need to have 50W Class A into 4 ohms. I just want to be able to hook up a 4 ohm load to it. My nearfields monitors are 4 ohm and I would like the option of useing it for those if I want. Why is it that the AX needs so much less voltage and so much more heatsink than the Krell? I'm not getting the picture some how.
I kind of got into this AX amp a little late I guess. A friend gave me the boards and I'm trying to get up to speed. If my 25-0-25 trannies won't work then I would prefer to buy the best ones for the job instead of just buying what's easily available. Can you suggest a size for me to use?
Thanks, Terry
the AX is single ended - so it is not as efficient as the Krell - it typically idles at 3 times rated power output (bias dependent). The reason fo rthe low voltage is that the AX uses differential output stages - like a bridged amp sharing the input section.
You'll be able to use a 4 ohm speaker, but unlike your Krell, the power available will not double unless you are very heavily biased. For nearfield monitor use, it will probably be suffiicent.
transformer recommenations depend on your power/impedance goal. plug the numbers into the spreadsheet and see what you need. The 16V transformers will be good for about 3.5A total bias - giving you about 75W into 8 ohms and 37 into 4. Figure you'll get rails 1.2x the AC voltage after a CRC filter and use a Va rating at least 2x your calculated dissipation - >3x would be mechanically quieter.
With 4 transistors in that setup (19V rails, 3.5A bias) you'll need heat sinks rated at .1C/W - with 8 you can get away with .3C/W
You'll be able to use a 4 ohm speaker, but unlike your Krell, the power available will not double unless you are very heavily biased. For nearfield monitor use, it will probably be suffiicent.
transformer recommenations depend on your power/impedance goal. plug the numbers into the spreadsheet and see what you need. The 16V transformers will be good for about 3.5A total bias - giving you about 75W into 8 ohms and 37 into 4. Figure you'll get rails 1.2x the AC voltage after a CRC filter and use a Va rating at least 2x your calculated dissipation - >3x would be mechanically quieter.
With 4 transistors in that setup (19V rails, 3.5A bias) you'll need heat sinks rated at .1C/W - with 8 you can get away with .3C/W
BobEllis said:
Hi Bob,
Thanks, That makes things clearer.
Couple of questions,
1) How did you come up with the figure of 3.5A bias? Is that based on the Va of the transformers? I see some guys claiming 6A bias. Is this better?
2) You say the Aleph is less efficient that the Krell. How do they compare sound wise?
I don't really have a Power/impedance goal per say. My goal is to make a good sounding reliable amp. I guess I would prefer that I didn't need $300 worth of heatsinks.
75W into 8 ohms sounds plenty to me.
Blessings, Terry
1. yes, I plugged numbers into the spreadsheet until I got a recommendation for at least a 300 VA transformer. Higher bias gets you more power into low impedance loads, and better sound. How high to go depends on your budget and tolerance for heat in your listening room.
2. I have no idea how they compare. I am sure that you'll find people who say that they prefer the AX to the Krell and vice versa. See the various Aleph and AX tweaking threads for ways to alter the character of your AX.
A pair of Conrad's MF350-150 (or whatever the largest size is) should handle a channel of AX in this configuration, but won't give you much room to increase the bias. Someone recently mentioned that 4 pieces cost around $150 delivered.
2. I have no idea how they compare. I am sure that you'll find people who say that they prefer the AX to the Krell and vice versa. See the various Aleph and AX tweaking threads for ways to alter the character of your AX.
A pair of Conrad's MF350-150 (or whatever the largest size is) should handle a channel of AX in this configuration, but won't give you much room to increase the bias. Someone recently mentioned that 4 pieces cost around $150 delivered.
Hi Bob, Is there any detriment to sound using more output devices? I could use say 12 per channel if necessary. Is it better to go up in vA on the transformers?
If you take a transformer that has say, 800vA and you unwind the secondary to a lower voltage does the vA go down too? Just trying to cover the angles.
Thanks, Terry
If you take a transformer that has say, 800vA and you unwind the secondary to a lower voltage does the vA go down too? Just trying to cover the angles.
Thanks, Terry
more output devices means more gate-source capacitance and potentially limited top end - perhaps giving a sound described as "darker." That said, three devices per quadrant ought to be fine.
One of the factors that goes into determining the VA rating of a transformer is its current capacity. Assuming that the secondary windings are not significantly oversized, the maximum current is pretty well fixed. Therefore, assume that the VA rating is reduced in direct proportion to the reduction of voltage.
Using a higher rated transformer and 12 output devices would give you more room to experiment with higher bias. Is that "better" for you?
One of the factors that goes into determining the VA rating of a transformer is its current capacity. Assuming that the secondary windings are not significantly oversized, the maximum current is pretty well fixed. Therefore, assume that the VA rating is reduced in direct proportion to the reduction of voltage.
Using a higher rated transformer and 12 output devices would give you more room to experiment with higher bias. Is that "better" for you?
still4given said:
I continue to look at this question, and I think it ends up being
a subjective call. More devices means lower distortion at low
frequencies, but depending on the amplifier circuit, you can start
seeing trade offs at the top end.
For followers you can get away with a lot, but when you're
using outputs in Common Source (voltage and current gain)
mode, I lean toward the minimum number that will work reliably.
That usually means about 25 watts dissipation per device.
😎
Nelson,
25W/mosfet will give you a junction temperature 26-27C above the heatsink. Given your spec of 50-55C heatsink temperature, the junction temperature will be 75-80C. Is that pretty much the max you would use in a commercial amp? Thanks.
25W/mosfet will give you a junction temperature 26-27C above the heatsink. Given your spec of 50-55C heatsink temperature, the junction temperature will be 75-80C. Is that pretty much the max you would use in a commercial amp? Thanks.
A comment and a question:
1) My Aleph-X monoblocks are absolutely amazing, and both sound and measure exceptionally well. Thanks Nelson and Grey! However, my amps have one tiny design flaw that I thought I'd pass along for the benefit of future builders. I designed for +/-15V rails and a 5-5.5A bias per channel, for a continuous power dissipation of ~150W per channel and a max signal output of ~50W. Each channel has a 225VA Plitron transformer. While the trafo's seem completely able to supply sufficient power for the demands of the amplifier, they are emitting an audible hum (not from the speakers, directly from the trafo's). The hum is low enough that I can't hear it even at very low music playback levels, but it is audible during the silence between tracks. I figured a 50% overhead margin on the trafo's should be enough, but apparently if you want to avoid mechanical hum from the trafo's you need more.
2) I configured my Aleph-X's to run each IFRP044 device at 37.5 watts. I can keep my hand pressed against the heatsinks, they are hot enough to be uncomfortable, but not so hot that I can't leave them there. How much am I shortening the life of the MOSFET's? I'm starting to wonder how soon I'm going to need the extra set of matched output transistors I have packed away.
Cheers, Terry
1) My Aleph-X monoblocks are absolutely amazing, and both sound and measure exceptionally well. Thanks Nelson and Grey! However, my amps have one tiny design flaw that I thought I'd pass along for the benefit of future builders. I designed for +/-15V rails and a 5-5.5A bias per channel, for a continuous power dissipation of ~150W per channel and a max signal output of ~50W. Each channel has a 225VA Plitron transformer. While the trafo's seem completely able to supply sufficient power for the demands of the amplifier, they are emitting an audible hum (not from the speakers, directly from the trafo's). The hum is low enough that I can't hear it even at very low music playback levels, but it is audible during the silence between tracks. I figured a 50% overhead margin on the trafo's should be enough, but apparently if you want to avoid mechanical hum from the trafo's you need more.
2) I configured my Aleph-X's to run each IFRP044 device at 37.5 watts. I can keep my hand pressed against the heatsinks, they are hot enough to be uncomfortable, but not so hot that I can't leave them there. How much am I shortening the life of the MOSFET's? I'm starting to wonder how soon I'm going to need the extra set of matched output transistors I have packed away.
Cheers, Terry
More Grist for the Mill
A bit of follow up info on my question:
After calculating my way through the heat transfer specs for the heatsink, insulator pad, and transistors, etc, my transistors should not exceed 120°C at the junction on a hot summer day, compared to the 175°C max rating. Further, after derating the max power spec of the transistor for the 120°C operating temperature, the max power rating of the transistor is 66W, compared to the 37.5W I am operating them at.
Any ideas as to the real world effects I might experience?
A bit of follow up info on my question:
After calculating my way through the heat transfer specs for the heatsink, insulator pad, and transistors, etc, my transistors should not exceed 120°C at the junction on a hot summer day, compared to the 175°C max rating. Further, after derating the max power spec of the transistor for the 120°C operating temperature, the max power rating of the transistor is 66W, compared to the 37.5W I am operating them at.
Any ideas as to the real world effects I might experience?
Hi Terry,
I think you under-calculated the ratings a bit.
With William's spreadsheet at hand I can see that the transformer
should be a 300 to 330 VA for a factor two and preferably more to be really safe. So 225VA is IMO a bit low since these beasts have to deliver full time. Do they hum at low bias?
Also, the 37.5W each mosfet has to get rid off is a bit on the high side. Personally I'd try to mount an extra pair or get some more heat sinks or both if possible.
Glad you like the sound. I'm deeply impressed with mine as well.
/Hugo 🙂
I think you under-calculated the ratings a bit.
With William's spreadsheet at hand I can see that the transformer
should be a 300 to 330 VA for a factor two and preferably more to be really safe. So 225VA is IMO a bit low since these beasts have to deliver full time. Do they hum at low bias?
Also, the 37.5W each mosfet has to get rid off is a bit on the high side. Personally I'd try to mount an extra pair or get some more heat sinks or both if possible.
Glad you like the sound. I'm deeply impressed with mine as well.
/Hugo 🙂
Hey Hugo,
Yes, the loudness of the hum is directly proportional to the bias level. I'm going to learn to live with the hum though, as I find I really prefer the way the amp sounds at the higher bias level. It sounds more liquid but at the same time the highs seem better defined and the amp is more dynamic overall. I know it sounds contradictory, but that's exactly what I perceive.
On the other front, I reviewed the Zen V5-7 articles, and in those Nelson is running the output transistors at up to 75W each, and I have one big heatsink for every two fets. I'm just trying to get a feel for what I am up against. If I have to change the fets out every year, I'll have to rethink things, but if I can get 5 years that doesn't concern me much.
Cheers, Terry
Yes, the loudness of the hum is directly proportional to the bias level. I'm going to learn to live with the hum though, as I find I really prefer the way the amp sounds at the higher bias level. It sounds more liquid but at the same time the highs seem better defined and the amp is more dynamic overall. I know it sounds contradictory, but that's exactly what I perceive.
On the other front, I reviewed the Zen V5-7 articles, and in those Nelson is running the output transistors at up to 75W each, and I have one big heatsink for every two fets. I'm just trying to get a feel for what I am up against. If I have to change the fets out every year, I'll have to rethink things, but if I can get 5 years that doesn't concern me much.
Cheers, Terry
I got my Aleph-X up and running on the test bench. The problem
is that this quote from Mr. Pass has always been in the back of
my mind.
You can run them at any rail voltage you like, but below 25
volts or so you start into the part of the curve where every
volt counts, so I recommend that at least as a + to - figure.
I have 15 volt transformers that i've already purchased.
Plus i'm not entirely sure how to add fets for more power
if i were to up the rails to 25V.
How will the XA60 and XA30 get around this rule of thumb?
Eviedntly Grey's circuit must be a little bit different then the
production XAs?
is that this quote from Mr. Pass has always been in the back of
my mind.
You can run them at any rail voltage you like, but below 25
volts or so you start into the part of the curve where every
volt counts, so I recommend that at least as a + to - figure.
I have 15 volt transformers that i've already purchased.
Plus i'm not entirely sure how to add fets for more power
if i were to up the rails to 25V.
How will the XA60 and XA30 get around this rule of thumb?
Eviedntly Grey's circuit must be a little bit different then the
production XAs?
metalman said:
Terry, Congratulations! I know you once expressed interest in my boards, but it looks like the final product is not moving along as you might want. Anyway, what about adding to/replacing your heatsinks once🙂?
moe29 said:
moe29, that's why NP loves the 300VA 18V-0-18V Plitron transformer...he must have a warehouse full of it.
You would add more fets if you are increasing total bias current. This is to distribute the total current thereby dissipating about 25W for every fet. So if you have +/-25V supplies you can run about 1A on each branch of output fet.
XA60 = +/-34V(?); XA30=+/-25V(?)
Blues said:
OK, this is what is confusing me. If Nelson says that when going below 25V rails the voltage starts to become important the why is it so bad to go above 25V. I'm having a tough time finding 18-0-18 300vA to 400vA tranformers. What would I have to do to use something like these 22-0-22 400vA transformers?
Would it be better to use a 16-0-16 300vA?
Thanks, Terry
metalman,
You're welcome.
Yes, I'd say that 225VA is pushing it. I'm not surprised that the transformers hum a bit. As long as they don't get too hot, you'll be okay. Just make sure that you provide plenty of ventillation for the transformers.
37.5W is pretty warm--use as much heatsinking as you can lay your hands on.
There's absolutely no question that the circuit responds well to higher bias. I played extensively with the bias on one of my Aleph 2s and the circuit can become rather thin sounding at low bias. Raising the bias improves everything except your electricity bill.
moe29,
The 25V rail figure is not an absolute number in any way. What happens is that the MOSFET capacitance becomes more of a problem below this number. This makes the the MOSFETs a little more difficult to drive and distortion rises as a result. I put some distortion readings in the original Aleph-X thread on this.
One thing to do when faced with higher capacitance is to provide more drive current, perhaps with a follower, although there are other options.
Blues,
An XA with 34V rails will develop in excess of 200W. I expect that the rails would be lower, but with changes in the front end in order to increase the drive.
still4given,
I'm not sure why you feel that it's bad to go over 25V. The only penalty is increased heat dissipation in the devices. Of course, the power supply and heatsinking will need to be increased, but that's not bad, per se, just expensive.
Grey
You're welcome.
Yes, I'd say that 225VA is pushing it. I'm not surprised that the transformers hum a bit. As long as they don't get too hot, you'll be okay. Just make sure that you provide plenty of ventillation for the transformers.
37.5W is pretty warm--use as much heatsinking as you can lay your hands on.
There's absolutely no question that the circuit responds well to higher bias. I played extensively with the bias on one of my Aleph 2s and the circuit can become rather thin sounding at low bias. Raising the bias improves everything except your electricity bill.
moe29,
The 25V rail figure is not an absolute number in any way. What happens is that the MOSFET capacitance becomes more of a problem below this number. This makes the the MOSFETs a little more difficult to drive and distortion rises as a result. I put some distortion readings in the original Aleph-X thread on this.
One thing to do when faced with higher capacitance is to provide more drive current, perhaps with a follower, although there are other options.
Blues,
An XA with 34V rails will develop in excess of 200W. I expect that the rails would be lower, but with changes in the front end in order to increase the drive.
still4given,
I'm not sure why you feel that it's bad to go over 25V. The only penalty is increased heat dissipation in the devices. Of course, the power supply and heatsinking will need to be increased, but that's not bad, per se, just expensive.
Grey
GRollins said:
Hi Grey,
It's not that I think it's bad to go over 25V. For some reason, that is the impression I got from what folks have been telling me. I'm just trying to settle on a tranformer so I can work out the parts list and get started building. I'm having trouble finding a low voltage high amperage transformer. I just don't want to get everything done and then find I didn't buy a big enough transformer. they are usually one of the biggest expenses when building an amp. I'd rather get it right the first try.
Blessings, Terry