The DC voltage across R23 and R25 is critical...but it's the gain devices in the differential, in conjunction with the level of bias and the value of the resistors, that cause that offset. It needs to be pretty close to the Vgs that the output MOSFETs expect to see.
That said, you can diddle it in any number of ways. You can increase/decrease the bias through the front end. You can change the resistor values for R23 and R25. You could replace those resistors with active loads (current sources) or a current mirror. You could change the ouput devices (as a really extreme example, consider the use of a power pentode such as the 6550, which would require something like 50V of bias). Etc.
However...
Rejigger any one of those things and you'll find that you have to change something else to compensate. It's not that it can't be done--witness the posts from folks who report having changed the front end bias and the value of the load resistors--it's that you have to stay on your toes. Be aware of the interrelation between your change and the circuit as a whole.
I could make a career building ten thousand variations on this circuit. Some of them might even be improvements. I'd rather get on to the next thing than obsess over minutiae in this one.
I'm on the verge of this problem with the X-preamp circuit I've got going in the other room. I've got four main gain devices: two different tube versions, a JFET version, and a MOSFET version. For the JFET version alone I've already got something like twenty permutations, and could easily think up twenty more. It's enough to drive a man to drink. (I'll have the Cabernet, please...If you listen carefully, you can hear Nelson laughing at me from three thousand miles away--he faces this problem every day.) You have to draw the line somewhere. There has to be a 'final' version.
Consider it from Nelson's point of view. Suppose he were to build the XA with five different front ends, but the same power output. How many would he sell? Damn close to zero. People don't want choices, they want answers. They'd get bogged down in trying to decide which front end they wanted, so they'd turn to the magazines to decide which one was 'best.' One reviewer would like version one, and another would like version two. Stalemate. The consumer would throw up his/her hands and buy something from Radio Shack.
I'm not trying to discourage people from trying variations on the Aleph-X. It's just that it's time-consuming to redesign the circuit twenty times for higher power, JFET front end, bipolar outputs, etc. I don't have enough hours in my days to begin with and get flustered when I have to do something like break off and go to work. (Not a problem today, fortunately.) If I could figure out a way to make audio pay (instead of costing me money) I'd be able to devote more time to it.
<i>Sigh.</i>
Don't get me wrong. I love this hobby. But it can take over your life if you're not careful, and it doesn't give much in return.
Kinda like my ex-wife, come to think of it...
Speaking of variations on preamps--do you guys have a consensus as to how much gain you'd like? CD folks will probably go for something on the order of, say, 10 dB. Those who listen to vinyl (like me) might want as much as 20-25dB. I'm tempted to peg it on the upper end and let folks knock it back down if they want, but then you get into more variations...
Where'd I leave my glass of wine?
Grey
That said, you can diddle it in any number of ways. You can increase/decrease the bias through the front end. You can change the resistor values for R23 and R25. You could replace those resistors with active loads (current sources) or a current mirror. You could change the ouput devices (as a really extreme example, consider the use of a power pentode such as the 6550, which would require something like 50V of bias). Etc.
However...
Rejigger any one of those things and you'll find that you have to change something else to compensate. It's not that it can't be done--witness the posts from folks who report having changed the front end bias and the value of the load resistors--it's that you have to stay on your toes. Be aware of the interrelation between your change and the circuit as a whole.
I could make a career building ten thousand variations on this circuit. Some of them might even be improvements. I'd rather get on to the next thing than obsess over minutiae in this one.
I'm on the verge of this problem with the X-preamp circuit I've got going in the other room. I've got four main gain devices: two different tube versions, a JFET version, and a MOSFET version. For the JFET version alone I've already got something like twenty permutations, and could easily think up twenty more. It's enough to drive a man to drink. (I'll have the Cabernet, please...If you listen carefully, you can hear Nelson laughing at me from three thousand miles away--he faces this problem every day.) You have to draw the line somewhere. There has to be a 'final' version.
Consider it from Nelson's point of view. Suppose he were to build the XA with five different front ends, but the same power output. How many would he sell? Damn close to zero. People don't want choices, they want answers. They'd get bogged down in trying to decide which front end they wanted, so they'd turn to the magazines to decide which one was 'best.' One reviewer would like version one, and another would like version two. Stalemate. The consumer would throw up his/her hands and buy something from Radio Shack.
I'm not trying to discourage people from trying variations on the Aleph-X. It's just that it's time-consuming to redesign the circuit twenty times for higher power, JFET front end, bipolar outputs, etc. I don't have enough hours in my days to begin with and get flustered when I have to do something like break off and go to work. (Not a problem today, fortunately.) If I could figure out a way to make audio pay (instead of costing me money) I'd be able to devote more time to it.
<i>Sigh.</i>
Don't get me wrong. I love this hobby. But it can take over your life if you're not careful, and it doesn't give much in return.
Kinda like my ex-wife, come to think of it...
Speaking of variations on preamps--do you guys have a consensus as to how much gain you'd like? CD folks will probably go for something on the order of, say, 10 dB. Those who listen to vinyl (like me) might want as much as 20-25dB. I'm tempted to peg it on the upper end and let folks knock it back down if they want, but then you get into more variations...
Where'd I leave my glass of wine?
Grey
I would like to go for a version capable to handle small signal from a phono-preamp. If you have a tube-version, even better ! I would be great to have a switch to reduce gain for CD though for instance in increasing the cathode-resistor (without bypass cap obviously) and get eben better local feedback...
Best Regards
Best Regards
So I am assuming I am the only one seeing different drops on the source resistors and a few tens of mV of offset?
Matching the 044 to even 10mV requires a truckload of mosfets! I wonder if using 240 would improve on things.
A little exercise in Excel has just proven to my satisfaction that putting 6 mosfets in parallel would bring one's realistic ability to match within 30mV to within a 6 parallel mosfet average of 5mV with 3mV standard deviation on the randomization and averaging repeat. I repeated the randomization 100 times.
Randomizing the source resistors most likely leads to an even narrower range.
So there, IMO paralleling may very well be the answer.
Matching the 044 to even 10mV requires a truckload of mosfets! I wonder if using 240 would improve on things.
A little exercise in Excel has just proven to my satisfaction that putting 6 mosfets in parallel would bring one's realistic ability to match within 30mV to within a 6 parallel mosfet average of 5mV with 3mV standard deviation on the randomization and averaging repeat. I repeated the randomization 100 times.
Randomizing the source resistors most likely leads to an even narrower range.
So there, IMO paralleling may very well be the answer.
GRollins said:
Holy.. what kind of speakers you guys run?
I have a Sony SACD and modded it to balanced out, I need no gain, I attenuate my passive with at least 10dB. My speakers are 85dB or so. Am I missing something here?
Cheers!
(Shiraz Cabernet BTW 🙂
/Peter
Blitz,
Cathode bypass?
What cathode bypass?
Ain't no cathode bypasses 'round here.
Don't believe in them.
grataku,
Sounds like you're having the same difficulty with the 044s that I had with the 644s. I bought--I think--twenty 044s and got four nearly perfectly matched devices out of the set. The 644s drove me to distraction. I bought a hundred of the things and they were all over the map. Kept changing, too. Ugh. Last time I mess with 644s...
Monte Carlo analysis, I did not do. I just set out to build the thing.
Pan,
Not everyone has as hot an output as you do. Also, it will depend greatly on the sensitivity/gain of the amplifier that the preamp is used with. Best to leave some elbow room. I've seen amps wouldn't do full output until they were fed 3V. Assuming that someone starts with, say, a Pearl phono stage (somewhat less than 40dB gain, if I recall), then puts a moving coil in front of it, then drives a first generation Zen...you're going to need a lot of go-juice to make it come together.
Actually, the worst case scenario would probably be one of those hot/cold followers. Those things need an amplifier just to drive the inputs.
(Just kidding.)
My speakers are upper eighties for efficiency, but I routinely use 15-20dB from the line stage because I'm running a moving coil into a 40dB phono stage. I will--eventually--get a more up-to-date phono stage going (got an idea for a variation on the Ono), but Rome wasn't built in a day. Others are probably in worse shape than I am. Gotta think about what other people might need.
Grey
Cathode bypass?
What cathode bypass?
Ain't no cathode bypasses 'round here.
Don't believe in them.
grataku,
Sounds like you're having the same difficulty with the 044s that I had with the 644s. I bought--I think--twenty 044s and got four nearly perfectly matched devices out of the set. The 644s drove me to distraction. I bought a hundred of the things and they were all over the map. Kept changing, too. Ugh. Last time I mess with 644s...
Monte Carlo analysis, I did not do. I just set out to build the thing.
Pan,
Not everyone has as hot an output as you do. Also, it will depend greatly on the sensitivity/gain of the amplifier that the preamp is used with. Best to leave some elbow room. I've seen amps wouldn't do full output until they were fed 3V. Assuming that someone starts with, say, a Pearl phono stage (somewhat less than 40dB gain, if I recall), then puts a moving coil in front of it, then drives a first generation Zen...you're going to need a lot of go-juice to make it come together.
Actually, the worst case scenario would probably be one of those hot/cold followers. Those things need an amplifier just to drive the inputs.
(Just kidding.)
My speakers are upper eighties for efficiency, but I routinely use 15-20dB from the line stage because I'm running a moving coil into a 40dB phono stage. I will--eventually--get a more up-to-date phono stage going (got an idea for a variation on the Ono), but Rome wasn't built in a day. Others are probably in worse shape than I am. Gotta think about what other people might need.
Grey
resistor questions...
Hi,
I'm running 12 volt rails and in order to use V1,2 and 3, I found that I needed to lower the values of R11 and R33 to 22.1k. I also had to lower the value of R26 to 200 ohms. Does this seem like the correct approach?
Many thanks,
John Inlow
Hi,
I'm running 12 volt rails and in order to use V1,2 and 3, I found that I needed to lower the values of R11 and R33 to 22.1k. I also had to lower the value of R26 to 200 ohms. Does this seem like the correct approach?
Many thanks,
John Inlow
Grey,
Thanks for the input, I will stop worrying about a few mV.
I look forward to the preamp, I am a CD person so 0-10dB is
perfect.
Chad,
I changed my PS from all C to a CRC and it cut the noise
down too something I can live with. I will try a CRCRC
later too see if that helps further.
grataku
I am seeing very close drops on the source resistors.
The DC offset I can take to 0 via the CCS and the
abs DC offsets are well below the acceptable range Nelson stated near the begining of all these posts.
As for matching the 044's I think I am living right,
I bought 25 and have 8 all within 1mV.🙂
wessol
Thanks for the input, I will stop worrying about a few mV.
I look forward to the preamp, I am a CD person so 0-10dB is
perfect.
Chad,
I changed my PS from all C to a CRC and it cut the noise
down too something I can live with. I will try a CRCRC
later too see if that helps further.
grataku
I am seeing very close drops on the source resistors.
The DC offset I can take to 0 via the CCS and the
abs DC offsets are well below the acceptable range Nelson stated near the begining of all these posts.
As for matching the 044's I think I am living right,
I bought 25 and have 8 all within 1mV.🙂
wessol
I have a balanced output as Pan have.
At the output of a Pearl the things are different. With a Grado I was feeling allright but with a low - output MC, 10db of gain for the preamplifier wasn't enough .Really.
At the output of a Pearl the things are different. With a Grado I was feeling allright but with a low - output MC, 10db of gain for the preamplifier wasn't enough .Really.
From what Grey is saying, I can assume that we should match both sets of output devices relative to each other and then a use a well matched pair of input devices. That should minimise the offset problems we see.
I must have gotten the 044 lot from hell!
No way I see transistors with a Vgs within 1mV!
Carpenter,
it looks like you are doing what you have to do.
HAPPY NEW YEAR TO ALL!!!!!!!!
No way I see transistors with a Vgs within 1mV!Carpenter,
it looks like you are doing what you have to do.
HAPPY NEW YEAR TO ALL!!!!!!!!
The New Year couldn't have started better, indeed!
I am Happy to report that my AlephX with dead stable 65mV DC offset works like a charm on Northcreek Rhythm Signature speakers. As far as I can tell it does everything better than my suped-up 30W Hiraga Le Classe A with Sanken transistors!
Fair or not I am driving the AX from the balanced BOSOZ source and the Hiraga from the SE output. The difference I have heard was pretty much the same I have heard when I went from class B solid state to class A push-pull. Where am I now? Class A++?? 😀
No doubt I must be delusional. I am going to throw everything out the window and build a gainclone! 😉
ASAIG a new set of 044 the 65mV DC offset will be history too!
I am Happy to report that my AlephX with dead stable 65mV DC offset works like a charm on Northcreek Rhythm Signature speakers. As far as I can tell it does everything better than my suped-up 30W Hiraga Le Classe A with Sanken transistors!
Fair or not I am driving the AX from the balanced BOSOZ source and the Hiraga from the SE output. The difference I have heard was pretty much the same I have heard when I went from class B solid state to class A push-pull. Where am I now? Class A++?? 😀
No doubt I must be delusional. I am going to throw everything out the window and build a gainclone! 😉
ASAIG a new set of 044 the 65mV DC offset will be history too!
I don't look for better than 10 mV. If you manage togrataku said:I must have gotten the 044 lot from hell!
find better, it doesn't tend to hold that over any real
temperature range.
C3P0: <i>My parts are showing?...I'm naked!</i>
The prototype here has no case. Vagrant breezes tend to cause the output offset to drift.
If the DC offset bothers you sufficiently, there's always the servo solution.
In the meantime, please feel free to be delusional. You've earned the right.
Grey
The prototype here has no case. Vagrant breezes tend to cause the output offset to drift.
If the DC offset bothers you sufficiently, there's always the servo solution.
In the meantime, please feel free to be delusional. You've earned the right.
Grey
those darned, vagrant breezes............
G-d I love my AX! Thanks Grey. And thank you Nelson.
John
G-d I love my AX! Thanks Grey. And thank you Nelson.
John
GRollins said:
Really? After you told me about the water cooled aleph episode, I made it a point to go at the amp with a fan and that produced only a 100 mV change in the absolute offset but nothing at all happened to the real DC at the speaker.
There is something to be said about thermally coupling the input differential, eh, eh, eh 😉
The Hiraga is a different beast, thr DC starts @ 120 mV and settles to around 3mV over the course of 20 minutes.
Ah, but is the fan blowing on the <i>front</i> end?
Bear in mind that the prototype front end MOSFETs aren't next to each other so they don't track thermal changes as well as other layouts.
Cooling the back end doesn't seem to be quite such an issue...at least regarding DC offset.
Grey
Bear in mind that the prototype front end MOSFETs aren't next to each other so they don't track thermal changes as well as other layouts.
Cooling the back end doesn't seem to be quite such an issue...at least regarding DC offset.
Grey
a bit of information to share regarding chokes.....
Hi all,
I've finished my first channel of the amazing AX. I'm using an L/C power supply filtering system. From the start it has amazed me with beautiful sound.
But..........
Up to now, I've experienced a challenging problem with an awful sounding low frequency hum in my high frequency horns (cut off point: 400hz). I've played with the star ground to the point of seeing stars! I've place a steel plate beneath the pcb to block electromagnetism. I tried snubber circuits across the chokes, I've tried Nelson's thermistor ideas. I've tried dual bridges, single bridges, hell I was ready to buy a bridge, any bridge. I've been so bummed I've even been snotty to my wife......... true diy addiction and the pain of failure.
You'd think I would dwell on my success of having successfully constructed an AX monoblock. Nope; two days of nothing but that annoying hum......
So, today at work, a little voice in my head, you probably know the kind I'm referring to, suggested the strangest thing: switch the power leads on one of the chokes. Only one! I gave it some thought; the chokes are in close proximity to one another due to space requirements. Perhaps the chokes' electromagnetic fields were inducing voltage into one another. Perhaps reversing the power leads on one choke would create a cancellation effect. Hummmmm. The short answer is that the idea worked. The ac ripple out of the speaker outputs was 0.008 volts (very loud on my horn) and now it's down to 0.002 volts! All this from simply reversing the power supply leads on one choke.
I wrote this because I'm truly fond of the sound the AX creates with an L/C filtered power supply; rich, sweet and gutsy! I didn't want to give up the choke. If I were to compare my sound with that of Shawn Wessol's AX I'd say they are like day and night. Shawn's amp sounds great, but on my horns the amp has an edge that grabs the upper high frequencies and casts them in steel. I'm a little bit hard of hearing and didn't mind at all, but my wife with the "golden ears" found it hard to take. The choke has changed her mind and she now favors the AX. I think that the two items go hand in hand. The AX is bright and the choke is a bit dull. Put them together and it's almost like a peanut butter cup.
Sorry for the long winded story, but it had to be told.
John Inlow
Hi all,
I've finished my first channel of the amazing AX. I'm using an L/C power supply filtering system. From the start it has amazed me with beautiful sound.
But..........
Up to now, I've experienced a challenging problem with an awful sounding low frequency hum in my high frequency horns (cut off point: 400hz). I've played with the star ground to the point of seeing stars! I've place a steel plate beneath the pcb to block electromagnetism. I tried snubber circuits across the chokes, I've tried Nelson's thermistor ideas. I've tried dual bridges, single bridges, hell I was ready to buy a bridge, any bridge. I've been so bummed I've even been snotty to my wife......... true diy addiction and the pain of failure.
You'd think I would dwell on my success of having successfully constructed an AX monoblock. Nope; two days of nothing but that annoying hum......
So, today at work, a little voice in my head, you probably know the kind I'm referring to, suggested the strangest thing: switch the power leads on one of the chokes. Only one! I gave it some thought; the chokes are in close proximity to one another due to space requirements. Perhaps the chokes' electromagnetic fields were inducing voltage into one another. Perhaps reversing the power leads on one choke would create a cancellation effect. Hummmmm. The short answer is that the idea worked. The ac ripple out of the speaker outputs was 0.008 volts (very loud on my horn) and now it's down to 0.002 volts! All this from simply reversing the power supply leads on one choke.
I wrote this because I'm truly fond of the sound the AX creates with an L/C filtered power supply; rich, sweet and gutsy! I didn't want to give up the choke. If I were to compare my sound with that of Shawn Wessol's AX I'd say they are like day and night. Shawn's amp sounds great, but on my horns the amp has an edge that grabs the upper high frequencies and casts them in steel. I'm a little bit hard of hearing and didn't mind at all, but my wife with the "golden ears" found it hard to take. The choke has changed her mind and she now favors the AX. I think that the two items go hand in hand. The AX is bright and the choke is a bit dull. Put them together and it's almost like a peanut butter cup.
Sorry for the long winded story, but it had to be told.
John Inlow