Hi,
I think I got the gainclone bug
Digging on the web I saw that the jeff rowland concentra uses some concepts that are mentioned in this forum regarding the gainklones, like the use of LM3886 and battery supply. Taking a look at
http://www.jeffrowland.com/review16b.htm
we can see that:
1 - input is through a tranformer and an AD815
2 - output are 6 x LM3886
TI has an application ("Complete Audio Amplifier") where they use a OPA134 input opamp, that serves to make the treble and bass controls, and then a simple OPA548 (replaceable by the usual suspects).
My question/thoughts are if these more elaborated designs can be better than the gainclone. And if there is a point in converting an unbalanced signal to balanced signal inside the amp to send it to the further stages. And if the transformer is necessary or the input opamp can do it.
Miguel
I think I got the gainclone bug
Digging on the web I saw that the jeff rowland concentra uses some concepts that are mentioned in this forum regarding the gainklones, like the use of LM3886 and battery supply. Taking a look at
http://www.jeffrowland.com/review16b.htm
we can see that:
1 - input is through a tranformer and an AD815
2 - output are 6 x LM3886
TI has an application ("Complete Audio Amplifier") where they use a OPA134 input opamp, that serves to make the treble and bass controls, and then a simple OPA548 (replaceable by the usual suspects).
My question/thoughts are if these more elaborated designs can be better than the gainclone. And if there is a point in converting an unbalanced signal to balanced signal inside the amp to send it to the further stages. And if the transformer is necessary or the input opamp can do it.
Miguel
Hehehe. I like how their marketing literature tries to mislead people, referring to a power opamp as an "Intelligent Power Transistor with Gain" and implies that these devices are simply replacing the typical output transistors in an amplifier. 47 Labs has some laughable marketing verbiage too, but at least they call a spade a spade when it comes to the Gaincard being based on a power opamp.
Well, the Concentra will certainly give you power power than a Gainclone using a single power opamp. Other than that, I don't see any particular advantage.
As for converting an unbalanced signal to a balanced signal inside the amp, well, seeing as the Concentra uses a bridged output scheme, you need to produce a symmetrical signal to drive the two halves.
As for the transformer, I'm not sure what you mean by the input opamp doing it. Not sure what the "it" is that you're referring to. An opamp certianly can't provide electrical isolation. It can't offer as good common-mode rejection performance from balanced sources without trimming of the output impedances of the source to make sure they're as matched as possible and it can't give you anywhere near the common-mode rejection that a transformer can when fed from unbalanced sources.
I've been using power opamps in my own amps for over a dozen years. While the power opamps I've used have changed over the years, the basic configuration has always been input transformer/power opamp. It's worked well for me.
My amps have all been just power amps though. However you could simply put a pot or swtiched attenuator between the transformer and the power opamp. However this would only work using the power opamp in non-inverting mode. Otherwise, the attenuator would need to be upstream from the input transformer.
se
miguel2 said:
Well, the Concentra will certainly give you power power than a Gainclone using a single power opamp. Other than that, I don't see any particular advantage.
As for converting an unbalanced signal to a balanced signal inside the amp, well, seeing as the Concentra uses a bridged output scheme, you need to produce a symmetrical signal to drive the two halves.
As for the transformer, I'm not sure what you mean by the input opamp doing it. Not sure what the "it" is that you're referring to. An opamp certianly can't provide electrical isolation. It can't offer as good common-mode rejection performance from balanced sources without trimming of the output impedances of the source to make sure they're as matched as possible and it can't give you anywhere near the common-mode rejection that a transformer can when fed from unbalanced sources.
I've been using power opamps in my own amps for over a dozen years. While the power opamps I've used have changed over the years, the basic configuration has always been input transformer/power opamp. It's worked well for me.
My amps have all been just power amps though. However you could simply put a pot or swtiched attenuator between the transformer and the power opamp. However this would only work using the power opamp in non-inverting mode. Otherwise, the attenuator would need to be upstream from the input transformer.
se
This has already been discussed previously. I think R Jones owns a JR 3886 based amp and he used it as reference while working with the STK modules. Apparently this particular JR is not that great sounding. But i would imagine the sound is different and in some ways better than a gaincard. The multiple parallel devices will certainly improve drive of more difficult loads and generally bass quality. The bridging will further improve dynamics and bass as it solves problems of ground currents/psu well. Arguably the best way to implement bridging is using a quality transformer for the phase inversion. I think JR uses Jensen. So it may be a very valid exercise to improve on some sonic aspects of gainclone while hopefully not losing out on others (immediacy, microdynamics).
cheers
peter
cheers
peter
I guess the transformer we are talking about here would have to be something very well made, which means expensive. Wouldn't a nice pair of opamps costing 1euro or such do the same trick, feeding one in the + input and the other on the - input? I guess the feedback is different and would have to be adjusted.
Transformers:
1. provide galvanic isolation,
2. are immune to some nasties that active devices are not (thanks to the laws of physics),
3. do not generate some nasties that active devices do (again, thanks to the laws of physics).
Jeff Rowland does indeed use Jensen transformers. And yes, they are expensive, but this IS high end, right?
However, active devices can sound good in this role.
mlloyd1
1. provide galvanic isolation,
2. are immune to some nasties that active devices are not (thanks to the laws of physics),
3. do not generate some nasties that active devices do (again, thanks to the laws of physics).
Jeff Rowland does indeed use Jensen transformers. And yes, they are expensive, but this IS high end, right?
However, active devices can sound good in this role.
mlloyd1
miguel2 said:
tbla,
Go to http://www.jeffrowland.com/techtalk.htm and scroll down the list. There are several articles in the TechTalk section that address the company's use of input transformers. The models you listed are called out specifically because they don't use input transformers. And it all depends on how you pay for your women. Those of us on the long term plan really have to pay out the nose.
Go to http://www.jeffrowland.com/techtalk.htm and scroll down the list. There are several articles in the TechTalk section that address the company's use of input transformers. The models you listed are called out specifically because they don't use input transformers. And it all depends on how you pay for your women. Those of us on the long term plan really have to pay out the nose.
Re: single ended to balanced...
That'd be one way to convert an unbalanced input to a balanced output. Or you could use the more generic and ubiquitous approach using a dual opamp chip.
But the issue here is input transformers, not simply single-ended to balanced conversion.
I believe in their new gear instead of input transformers, Rowland is using InGenius line receivers from <a href="http://www.thatcorp.com">THAT Corporation</a>.
Bill Whitlock (President of Jensen Transformers) patented an amplifier topology which gives high common-mode rejection even when there is an imbalance in the source impedances (the worst case being unbalanced sources). Transformers are very good at this and can give you triple digit common-mode rejection even from unbalanced sources. But typical active circuits can only achieve this level of rejection when the source impedances are precisely balanced.
se
tbla said:
That'd be one way to convert an unbalanced input to a balanced output. Or you could use the more generic and ubiquitous approach using a dual opamp chip.
But the issue here is input transformers, not simply single-ended to balanced conversion.
I believe in their new gear instead of input transformers, Rowland is using InGenius line receivers from <a href="http://www.thatcorp.com">THAT Corporation</a>.
Bill Whitlock (President of Jensen Transformers) patented an amplifier topology which gives high common-mode rejection even when there is an imbalance in the source impedances (the worst case being unbalanced sources). Transformers are very good at this and can give you triple digit common-mode rejection even from unbalanced sources. But typical active circuits can only achieve this level of rejection when the source impedances are precisely balanced.
se
Philo said:tbla, ... And it all depends on how you pay for your women. Those of us on the long term plan really have to pay out the nose.
I don't know about that, I just bought one of these for my wife so she can stay in touch and and entertain herself wherever the US Navy decides to send her. Could of bought a lot of nice drivers and amp parts for that much money
http://www.fujitsu-siemens.com/rl/products/notebooks/professional/lifebook/lifebooks/lifebooks.html
Attachments
Re: The other side of the moon...
So is any conductive loop. But this is really only an issue with high impedance input transformers, which are typically supplied with magnetic shielding.
Which is why the best transformers use the best core materials with the smallest hysteresis loops. This doesn't become significant except at the lowest frequencies and the highest input levels. Linearity increases with frequency and by the time you get to a couple hundred hertz, is virtually non-existent.
If your source component is outputting enough DC to actually saturate the core, it's probably time to have your source component repaired.
Certainly DC currents in the primary or secondary can degrade performance. But it's rather trivially easy to keep these down to levels that they're not terribly significant.
Guess it depends where you draw the line for "limited." The input transformers I use have a bandwidth of 100kHz. I don't see how that's limiting in any meaningful way. And output transformers can have bandwidths into the tens of megahertz.
So what exactly do you consider "bandwith limited"?
se
Tube_Dude said:
So is any conductive loop. But this is really only an issue with high impedance input transformers, which are typically supplied with magnetic shielding.
Which is why the best transformers use the best core materials with the smallest hysteresis loops. This doesn't become significant except at the lowest frequencies and the highest input levels. Linearity increases with frequency and by the time you get to a couple hundred hertz, is virtually non-existent.
If your source component is outputting enough DC to actually saturate the core, it's probably time to have your source component repaired.
Certainly DC currents in the primary or secondary can degrade performance. But it's rather trivially easy to keep these down to levels that they're not terribly significant.
Guess it depends where you draw the line for "limited." The input transformers I use have a bandwidth of 100kHz. I don't see how that's limiting in any meaningful way. And output transformers can have bandwidths into the tens of megahertz.
So what exactly do you consider "bandwith limited"?
se
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.