aleph 1.2's to AX100
I just thought to let you know (with the news of new designs coming from the NP kitchen) that the 350.5 I have does outperform the Aleph in one big area. It is much quieter and I mean no noice period!!! The unit drives my big line arrays to building destroying volumes and doesn't run hot. My Aleph 1.2 is much beloved by the local electrical authorities. The aleph still does ambience or echo effects on recordings extremely well and I can tell now when studio recordings have ambience or delay added. The recording spaces are clealy identified with either unit. The 350.5 has better bass and control. It looks nicer and the prime minister will allow it in the living room. NP needs to ship the 350.5 with a politically correct power cord. He says in the manual that the power chord shouldn't cost more than the amp. Good advice but a three inch diameter chord with 0 capacitance and 0 resistance made from precious metals would be nice. dave
I just thought to let you know (with the news of new designs coming from the NP kitchen) that the 350.5 I have does outperform the Aleph in one big area. It is much quieter and I mean no noice period!!! The unit drives my big line arrays to building destroying volumes and doesn't run hot. My Aleph 1.2 is much beloved by the local electrical authorities. The aleph still does ambience or echo effects on recordings extremely well and I can tell now when studio recordings have ambience or delay added. The recording spaces are clealy identified with either unit. The 350.5 has better bass and control. It looks nicer and the prime minister will allow it in the living room. NP needs to ship the 350.5 with a politically correct power cord. He says in the manual that the power chord shouldn't cost more than the amp. Good advice but a three inch diameter chord with 0 capacitance and 0 resistance made from precious metals would be nice. dave
Hi Dave,
For my AX100 with 5.4 amps of bias (total for both sides) the calculation of output power is as follows: Peak power into 8 ohms = 5.4 squared times 8 which = roughly 233W. Divide by 2 to get average power = roughly 116W. This is covered in the WIKI. To get this output the amp needs to swing enough volts. In this case it's just over 40V.
For 200W output into 8 ohms you would need the 8 to 9 amps of total bias that the XA200 uses. For the XA160 it's 6 to 7 amps. Both of these amps use roughly 32V rails. In each case you figure the number of output transistors based on how much heat you can dissipate per.
You can't beat the physics.
I hope this helps.
Graeme
For my AX100 with 5.4 amps of bias (total for both sides) the calculation of output power is as follows: Peak power into 8 ohms = 5.4 squared times 8 which = roughly 233W. Divide by 2 to get average power = roughly 116W. This is covered in the WIKI. To get this output the amp needs to swing enough volts. In this case it's just over 40V.
For 200W output into 8 ohms you would need the 8 to 9 amps of total bias that the XA200 uses. For the XA160 it's 6 to 7 amps. Both of these amps use roughly 32V rails. In each case you figure the number of output transistors based on how much heat you can dissipate per.
You can't beat the physics.
I hope this helps.
Graeme
aleph 1.2's to AX100
GL thanks, so the class a wattage is really achieved through a healthy dose of current. Yes I will have to reconsider the reliability issue with the number of output devices.
When you say for the XA160 that it uses 7-8amps of bias is that total or per side. For instance the positive signal half would use 4 amps and the negative side would also use 4amps for a total of 8amps. Of course this is with minimal dc offset. If the current is divided between the quads than I think I will be ok. If each positive and negative side is pushing 8amps then I may have to lood at a fan. Just checking, thanks dave
GL thanks, so the class a wattage is really achieved through a healthy dose of current. Yes I will have to reconsider the reliability issue with the number of output devices.
When you say for the XA160 that it uses 7-8amps of bias is that total or per side. For instance the positive signal half would use 4 amps and the negative side would also use 4amps for a total of 8amps. Of course this is with minimal dc offset. If the current is divided between the quads than I think I will be ok. If each positive and negative side is pushing 8amps then I may have to lood at a fan. Just checking, thanks dave
Hi Dave,
The Passlabs site says roughly 7 amps for the XA160. This is the total for both halves. That means 3.5 amps for each side.
If you figure 32V rails then the total power consumption at idle is 32 + 32 times 7 = 448W in the output section plus 10% more for the power supply - 448 time 1.10 = 492W. This pretty close to the 550W figure given in the data sheet for the XA160. I think the difference might be extra losses in the the power supply filters or Passlabs may have added a fudge factor just to be on the safe side with the spec.
Graeme
The Passlabs site says roughly 7 amps for the XA160. This is the total for both halves. That means 3.5 amps for each side.
If you figure 32V rails then the total power consumption at idle is 32 + 32 times 7 = 448W in the output section plus 10% more for the power supply - 448 time 1.10 = 492W. This pretty close to the 550W figure given in the data sheet for the XA160. I think the difference might be extra losses in the the power supply filters or Passlabs may have added a fudge factor just to be on the safe side with the spec.
Graeme
aleph 1.2's to AX100
Gl-this is what I think the Aleph 1.2 is, I will measure it tonight.
I am looking around now for transformers. I would rather get anohter 350.5 for biamping.
buuuut That is a lot of money for getting lazy. dave
Gl-this is what I think the Aleph 1.2 is, I will measure it tonight.
I am looking around now for transformers. I would rather get anohter 350.5 for biamping.
buuuut That is a lot of money for getting lazy. dave
gl said:
I don't agree with you.
As i pointed out this cap has an important rule on the feedback network.
If you just consider that by varying it a little bit you obtain different frequencies and phases response, you can surely understand the importance of it.
Not to mention that the signal directly passes through it.
The only problem is that i don't know where to get a so low value of capacitance.
I'll see, as babowana suggested, to find something on backgate (ca't find anything with jensen)
Stefanoo said:
I tell 'em and I tell 'em, but they just don't listen...this is what you get for simulating. (Man, I wish I had that many spare hours in a day!)
When using a real circuit on a real board populated with real parts in the real world, you'll find that all your careful simulations fly right out the window. Stray capacitance on the board contributes some of what you'll need. After that you use the actual circuit and an oscilloscope--not a simulator--to set the final value. Nelson has suggested using twisted wire to achieve the final capacitance value. Or you can try actual caps in the circuit. Or you can use none at all. The circuit is actually quite content without capacitors in the feedback loop.
If you are looking for small value caps, I suggest polystyrene or silver/mica caps. They--particularly the silver/mica--are available down into single digit pF values.
Grey
Hi Stefano,
I respect your opinion. However, I have listened to the amps with 10pf silver mica caps, with 3-4cm of twisted 30ga. wire as caps, and with no caps at all. I could hear no difference.
The only caps I would use here are silver micas. They are small and available and cheap. I can't imagine Jensen or Blackgate making anything small enough (or cheap enough).
Cheers,
Graeme
I respect your opinion. However, I have listened to the amps with 10pf silver mica caps, with 3-4cm of twisted 30ga. wire as caps, and with no caps at all. I could hear no difference.
The only caps I would use here are silver micas. They are small and available and cheap. I can't imagine Jensen or Blackgate making anything small enough (or cheap enough).
Cheers,
Graeme
Actually, it occurs to me that there was a fellow who insisted that you shouldn't use caps in the feedback loop. Ever. This might go all the way back to tube days. I can't remember. Somebody here who has more functioning brain cells than I do at the moment will probably be able to cough up his name and his reasoning. I'm fried--the boys are teething and the ruckus they're making is enough to wake the dead. I might just skip the going home part tonight and check into a hotel.
(Just kidding.)
(I think I'm just kidding.)
(I am kidding, aren't I?)
Yes, there are circuits that go absolutely batty if you don't do something to knock down the top end, but a properly built Aleph-X is a well-behaved critter. The rise is modest, at best, and not a stability issue.
Assuming that you want to tweak the thing the best way to go about it is to use something like a 10kHz square wave. Watch the leading corner. You know how to test the shock absorbers in a car, right? Stand on the bumper. Jump off. The car should rebound up, down, and a little back up to a resting position. At most. Same thing here. There'll be a little spike on the leading corner in an uncompensated amp, followed by a few ripples that diminish. Use capacitance such that you have no more than the initial spike, followed by a dip, slight rise, and a settling back to the proper top of the square wave. That's for those of us who want as much bandwidth as we can tease out of the circuit. Them what's more conservative will want to ever so slightly round that leading corner, with no spike at all. Don't count on getting a perfectly square corner, just choose your compromise.
By the way, don't try the bumper-jump damping with a high feedback circuit. Only low NFB circuits need apply.
The Aleph-X qualifies.
Grey
(Just kidding.)
(I think I'm just kidding.)
(I am kidding, aren't I?)
Yes, there are circuits that go absolutely batty if you don't do something to knock down the top end, but a properly built Aleph-X is a well-behaved critter. The rise is modest, at best, and not a stability issue.
Assuming that you want to tweak the thing the best way to go about it is to use something like a 10kHz square wave. Watch the leading corner. You know how to test the shock absorbers in a car, right? Stand on the bumper. Jump off. The car should rebound up, down, and a little back up to a resting position. At most. Same thing here. There'll be a little spike on the leading corner in an uncompensated amp, followed by a few ripples that diminish. Use capacitance such that you have no more than the initial spike, followed by a dip, slight rise, and a settling back to the proper top of the square wave. That's for those of us who want as much bandwidth as we can tease out of the circuit. Them what's more conservative will want to ever so slightly round that leading corner, with no spike at all. Don't count on getting a perfectly square corner, just choose your compromise.
By the way, don't try the bumper-jump damping with a high feedback circuit. Only low NFB circuits need apply.
The Aleph-X qualifies.
Grey
aleph 1.2's to AX100
GL, I plugged my Aleph 1.2 into my ac condioner because it has an amp meter and I read 5amps AC at 120volts. That mean of course that the plitron 2kva buzzards are drawing 600va at idle with no load attached. I don't know exactly how va compares to watts but since my ps has two inductors my losses will be a bit greater. If I use your calculation method and I have .5 volt drop accross the source resistor x 12 which is 1.5 ohms I get 33ma x 24 devices =7.9 amps total
My actual measurements were a bit less (.44volts).
For
This gives 7amp squared x 8 or about 400watts of peak power divided by 2 to get average power for the aleph 1.2 = 200watts.
The aleph has 60 volt rails so I am confused why it needs this much voltage unless it is because it is twice as inefficient as a normal class A and it is not a bridged amp. If you know the answer let me in on it.
GL, I plugged my Aleph 1.2 into my ac condioner because it has an amp meter and I read 5amps AC at 120volts. That mean of course that the plitron 2kva buzzards are drawing 600va at idle with no load attached. I don't know exactly how va compares to watts but since my ps has two inductors my losses will be a bit greater. If I use your calculation method and I have .5 volt drop accross the source resistor x 12 which is 1.5 ohms I get 33ma x 24 devices =7.9 amps total
My actual measurements were a bit less (.44volts).
For
This gives 7amp squared x 8 or about 400watts of peak power divided by 2 to get average power for the aleph 1.2 = 200watts.
The aleph has 60 volt rails so I am confused why it needs this much voltage unless it is because it is twice as inefficient as a normal class A and it is not a bridged amp. If you know the answer let me in on it.
aleph 1.2's to AX100
sorry where you see the 12 in the formula it is a typo. I am getting tired. dave
sorry where you see the 12 in the formula it is a typo. I am getting tired. dave
aleph 1.2's to AX100
Well, I got two options for power to rebuild my 1.2's to XA's.
My current transformers are 2kva plitron torrids that are suppling 60volt rails.
I got a quote from Avel Lindberg for four at 25v+25v 1 kva torroid transformers for $500. This puts my dc rails at 35.6v at no load for the XA.
Plitron wants $1100 for their audio grade line but the secondary voltage is 30v+30v for two 2kva transformers. That would put my rails at 42 volts.
I am desperate to find a quiet unit, any other suggestions? I have used the Avel Lindberg 800VA in my A75 Pass Tharagard and it is as quite as a mouse in a cheeze patch. I am afraid to use the Chinese brands as I don't know what quality that they may be?
I can use 2-1 kva's for each monoblock and center tap one leg of each to power ground. Will four make less or more noise than two bigger ones? I am reading up on how to measure dc on my ac line using my scope. If anyone has done it let me know. I have done it a lot on the secondary side of the transformers but this isolates me from the house line. dave
Well, I got two options for power to rebuild my 1.2's to XA's.
My current transformers are 2kva plitron torrids that are suppling 60volt rails.
I got a quote from Avel Lindberg for four at 25v+25v 1 kva torroid transformers for $500. This puts my dc rails at 35.6v at no load for the XA.
Plitron wants $1100 for their audio grade line but the secondary voltage is 30v+30v for two 2kva transformers. That would put my rails at 42 volts.
I am desperate to find a quiet unit, any other suggestions? I have used the Avel Lindberg 800VA in my A75 Pass Tharagard and it is as quite as a mouse in a cheeze patch. I am afraid to use the Chinese brands as I don't know what quality that they may be?
I can use 2-1 kva's for each monoblock and center tap one leg of each to power ground. Will four make less or more noise than two bigger ones? I am reading up on how to measure dc on my ac line using my scope. If anyone has done it let me know. I have done it a lot on the secondary side of the transformers but this isolates me from the house line. dave
Hi Dave,
The Aleph 1.2 has an unbalanced output and 60V rails. That means that the output voltage is referenced to ground. The output voltage will swing +- 57 to 58V max before clipping. The Aleph-X, however, is balanced out. Let's say you have 30V rails. As one side (or output) of the amp swings to within a couple of volts of the plus rail the other side will swing symmetrically (e.g. for a sine wave) toward the minus rail. There will be 58V or so between the output posts just before clipping. At the other polarity of the sine wave the output posts will be 60V apart too but the the polarity of the output voltage will be opposite.
You have the same +-58V swing for both the 1.2 and the AX. I didn't explain this well but I hope it helps.
Graeme
The Aleph 1.2 has an unbalanced output and 60V rails. That means that the output voltage is referenced to ground. The output voltage will swing +- 57 to 58V max before clipping. The Aleph-X, however, is balanced out. Let's say you have 30V rails. As one side (or output) of the amp swings to within a couple of volts of the plus rail the other side will swing symmetrically (e.g. for a sine wave) toward the minus rail. There will be 58V or so between the output posts just before clipping. At the other polarity of the sine wave the output posts will be 60V apart too but the the polarity of the output voltage will be opposite.
You have the same +-58V swing for both the 1.2 and the AX. I didn't explain this well but I hope it helps.
Graeme
aleph 1.2's to AX100
Gl thanks, it just confirms that I did the calculations correct by using 60 volts referenced to ground. I will begin in earnest this weekend with parts collection and trying to size the perf board.
I will do pt to pt on the board. I will put together some test circuits for measuring the active devices. I think 7 amps should be enought to drive frequecnies from 2600hz to 20kz. My Aleph's ran the line arrays just fine from 35hz to 20k.
Gl thanks, it just confirms that I did the calculations correct by using 60 volts referenced to ground. I will begin in earnest this weekend with parts collection and trying to size the perf board.
I will do pt to pt on the board. I will put together some test circuits for measuring the active devices. I think 7 amps should be enought to drive frequecnies from 2600hz to 20kz. My Aleph's ran the line arrays just fine from 35hz to 20k.
I'm trying to adjust the circuit to make it fits my specifics.
I would like to get the amp correctly working on 2ohm load, therefore i simulated the circuit in order to evaluate the behaviour under the different load conditions (8, 4 and 2 ohm).
The result was: with a supply rail voltage of 27V and bias resistors set up to 100K, bias of almoust 1.1A per device, with an array of 5 output devices per quadrant, 20 in total per each channel, the amp is correctly working under 8ohm load and behave preatty good on 4ohm, but really distorces on 2ohm load.
The output power over this configuration is about 110W @ 8ohm (160W pk) , 220W (400W-pk)@ 4ohm
(by the way, why is it not dropping down by the decreasing of the load impededence? Is it normal?)
So i tried to work this around by varying the bias.
I noticed that over a certain extent of incresing bias resistor (i think over 150-160Kohm), i didn't get any improvment on that sense at all.
So i increased the number of output devices up to 24 and then 28 and the capacity to drive a 2 ohm load was drammatically improved (400W @ 2ohm- 760W-pk), but i still had a little bit of distotion on my sine wave output power form that i will figure this out later.
Here are my questions:
Is there any limit (of course the heat dissipation) on the number of power output devices that i can employee on the output stage?
Will the front end be able to drive this huge number of output devices? Is the otput impedence the transconductance or capacitance goint to be an issue?
Accordingly to the experience got from you guys, is the sound of the amp going to be impaired increasing the number of devices or there would be an improvement since the output stage has more driving capabilities?
In order to have a better drivabilty, should the output stage be my only concern?
i mean, is there any other parameter that i can work around as for instance the feedback resistor that sets the gain, or the source resistor of the output devices (which i use 0.33 ohm) or power supply, to imrpove the drivability of this amplifier under hard loads?
Best,
Stefano.
I would like to get the amp correctly working on 2ohm load, therefore i simulated the circuit in order to evaluate the behaviour under the different load conditions (8, 4 and 2 ohm).
The result was: with a supply rail voltage of 27V and bias resistors set up to 100K, bias of almoust 1.1A per device, with an array of 5 output devices per quadrant, 20 in total per each channel, the amp is correctly working under 8ohm load and behave preatty good on 4ohm, but really distorces on 2ohm load.
The output power over this configuration is about 110W @ 8ohm (160W pk) , 220W (400W-pk)@ 4ohm
(by the way, why is it not dropping down by the decreasing of the load impededence? Is it normal?)
So i tried to work this around by varying the bias.
I noticed that over a certain extent of incresing bias resistor (i think over 150-160Kohm), i didn't get any improvment on that sense at all.
So i increased the number of output devices up to 24 and then 28 and the capacity to drive a 2 ohm load was drammatically improved (400W @ 2ohm- 760W-pk), but i still had a little bit of distotion on my sine wave output power form that i will figure this out later.
Here are my questions:
Is there any limit (of course the heat dissipation) on the number of power output devices that i can employee on the output stage?
Will the front end be able to drive this huge number of output devices? Is the otput impedence the transconductance or capacitance goint to be an issue?
Accordingly to the experience got from you guys, is the sound of the amp going to be impaired increasing the number of devices or there would be an improvement since the output stage has more driving capabilities?
In order to have a better drivabilty, should the output stage be my only concern?
i mean, is there any other parameter that i can work around as for instance the feedback resistor that sets the gain, or the source resistor of the output devices (which i use 0.33 ohm) or power supply, to imrpove the drivability of this amplifier under hard loads?
Best,
Stefano.
Stefanoo said:
Stefanoo,
Your spirit and energy is great, trying tackling tech things all in one. I respect it.
Here we have many do-it-yourselves who have lack of experience. We can’t expect any big advice from them. Meanwhile, we have also many professionals whose experience is long enough. But, we can’t expect their big advices either because their experiences are all confidential and afraid of rip-offs.
Sometimes, we need self-study, and when we ask part by part during it, all non-pros and yes-pros are getting very kind. Why don’t you try building yourself, see and ask? No money for nothing.
Best Luck ^^!
Hope somebody more experienced than you and I will think differently from you.
I think this is a good community, and when i will have this insight that i'm asking for, i won't fold back on giving tips to newbies that will go throught the same steps that i'm going through 'em right now.
Let's see.
I think this is a good community, and when i will have this insight that i'm asking for, i won't fold back on giving tips to newbies that will go throught the same steps that i'm going through 'em right now.
Let's see.
- Status
- Not open for further replies.