[Zaubertuba]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif ****. Sorry. I got the EMail notification for this but had something else going on at the time and forgot to get back over here to reply.

Sheesh Steve, you've been such an awesome mentor through this whole thing--I probably can cut you a little slack.

Quote:

Looks fine, but you still need to change the ground symbols on the XLR pin 1's and the ground loop breaker to chassis grounds.

Double *Doh!* ...duly noted.


Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif My original plans were to add 2x gain with the feedback resistor hack but I started out with just the default buffer configuration. It drives my k240s (600ohms) to a louder listening level for most things and a normal listening level for quieter thigns. My k701s got very loud, beyond listening levels for sure but not absolutely painful. I think when I get this cased up I'll end up adding some gain for headroom and also in case I run single ended headphones out of it.

Good point about also running single-ended. That's a key criteria in my design so I'll probably start out with 2x (total 4x double-ended) gain.

Question for both of you (and anyone else for that matter). Am I correct in assuming that adding the gain will throw all the documented test voltages out of whack? If so, how will I test it when I build? The same question could be asked about biasing the boards, I think.

Quote:

For the PSU regulator, I was going to use a couple of 7815 regulators on a piece of perfboard which shouldn't cost more than about $5 all together. You seem like a pretty thrifty diyer which I like to think of myself as. The Opus I have only cost around $100 with only USB input. I saved money by buying the bare boards and parts separately and by making a simple 317 regulator on perfboard. My bijou cost well under $250 before I started upgrading it by just repurposing an old tape player enclosure. Basically, there are plenty of ways to cut corners.

I think I'm pretty thrifty--though hacking together your own regulator like that has me beat.

I'm already planning on using some existing broken stereo components to help case my builds. $100 for an Opus sounds pretty awesome!

And I was thinking about a balanced Bijou, so if I stick to that it'll be a little pricier, but again, that's down the road a bit.

 

[mugdecoffee]

Adjusting the gain shouldn't throw off any voltages. The change in voltages would be the same (almost all of them at least) for adding 2x gain and for just doubling the input signal voltage.

Biasing was really easy actually. It worked exactly as it was supposed to. The only thing that threw me off a little bit is that one of the test points is negative and one positive but they should be the same

Even if you're a bit uneasy about P2P wiring and perfboards, using 7815/7915 (or any other 78xx/79xx) regulators only requires the regulator and a two caps. Its super simple and a good first P2P project. And its super cheap.

 

[Zaubertuba]

Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif Even if you're a bit uneasy about P2P wiring and perfboards, using 7815/7915 (or any other 78xx/79xx) regulators only requires the regulator and a two caps. Its super simple and a good first P2P project. And its super cheap.

P2P/Perfboard doesn't scare me (my a47 is all on perfboard). I just haven't found a "simple" audio-quality PS circuit design, which is why I was planning on the LCBPS (two treads barely save you any money).

I had a comment at one point that the LCBPS seemed to be overkill, but I'm kind of OCD about power quality--especially for an amplifier stage. How well-regulated is your design?

EDIT: Well, OCD for my newer builds, anyway. Don't look at what I did for the a47.

 

[Zaubertuba]

So, the moment I *say* I haven't found a decent circuit diagram...I find something--and fairly easy, I might add.

How about this (another Rod Elliot design)?

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif How about this (another Rod Elliot design)?

Nothing I can recommend against it.

se

 

[cobaltmute]

How about this (from the LT1033 datasheet):

Replace the LT parts with LM317/337. Sim'med as working for me and it worked for the person who posted the idea on diyAudio.

I'm slowly working up a PCB design for this.

 

[Zaubertuba]

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif How about this (from the LT1033 datasheet): Replace the LT parts with LM317/337. Sim'med as working for me and it worked for the person who posted the idea on diyAudio. I'm slowly working up a PCB design for this.

You're talking about this thread? Wow! I just read through it. He's definitely putting the design through it's paces.

Any thoughts on particular merits of these two designs (besides that they're simple and should also be relatively cheap to implement--an attractive trait for a novice diy'r like me

)? Somebody pointed out to me in an earlier thread that big caps don't really matter for a balanced amp since it's not sinking/sourcing return currents.

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif Somebody pointed out to me in an earlier thread that big caps don't really matter for a balanced amp since it's not sinking/sourcing return currents.

That's just plain incorrect.

There's a popular mythology among the headphone crowd that "balanced" amplifiers and three channel "active ground" amplifiers somehow bypass ground. That the load currents are instead routed to the output stages of the bridged channels and then magically evaporate into the ether or dribble out onto the floor.

The reservoir capacitance of no less importance in a "balanced" amp than an unbalanced amp.

se

 

[mugdecoffee]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif That's just plain incorrect. There's a popular mythology among the headphone crowd that "balanced" amplifiers and three channel "active ground" amplifiers somehow bypass ground. That the load currents are instead routed to the output stages of the bridged channels and then magically evaporate into the ether or dribble out onto the floor. The reservoir capacitance of no less importance in a "balanced" amp than an unbalanced amp. se

I think the idea is that if say the positive 10V rail drops to 8V, this would pull the output of each channel down by about a volt but in a balanced amp, both positive and negative channels would drop resulting in no change in the potential between the positive and negative outputs. This is how balanced amps supposedly cancel power supply noise that appears at the output so smoothing caps would be less critical. The idea is sound though I'm sure stiff rails would still benefit a balanced amp in practice.

 

[cobaltmute]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif You're talking about this thread? Wow! I just read through it. He's definitely putting the design through it's paces. Any thoughts on particular merits of these two designs (besides that they're simple and should also be relatively cheap to implement--an attractive trait for a novice diy'r like me )? Somebody pointed out to me in an earlier thread that big caps don't really matter for a balanced amp since it's not sinking/sourcing return currents.

Building the dual tracking one allows you to have an adjustable regulator for later. The fixed one is easier to build.

I would still put some capacitance after the reg. To my understanding this helps transient response.

 

[Steve Eddy]

Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif I think the idea is that if say the positive 10V rail drops to 8V, this would pull the output of each channel down by about a volt but in a balanced amp, both positive and negative channels would drop resulting in no change in the potential between the positive and negative outputs.

The rail voltage dropping doesn't pull the output down. It lowers the point at which you would start to enter clipping, but shy of clipping, the output will be the same.

Quote:

This is how balanced amps supposedly cancel power supply noise that appears at the output so smoothing caps would be less critical.

Bridged amps don't cancel power supply noise. At least not those that are made up using a pair of complimentary push-pull amps.

What bridging does is make any power supply noise common-mode. What cancels the noise is the headphones, which are differential in nature.

se

 

[Zaubertuba]

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif Building the dual tracking one allows you to have an adjustable regulator for later. The fixed one is easier to build. I would still put some capacitance after the reg. To my understanding this helps transient response.

I like the idea of an adjustable regulator. So I'd just add (I suppose it could be anything between 200mf to 1,000mf) two electrolytics paralleled between the 2.2mf caps and the diodes at the output?

Since the datasheet diagram points out R1 and R5 can be trimmed, would there be any benefit to substituting trimpots to fine-fine tune the tracking?

Also, Rod bypasses the electrolytics with ceramics in his design. Any reason to do this with the tracking supply as well?

 

[Zaubertuba]

I keep coming back to how I like the idea of using transformers in the input stage. I found these just now. From what I could find, it looks like they're 1:1, 20k.

Looking at transformers with Steve got me thinking. My 1212m can run at +4dBu, but pretty much any unbalanced consumer source is going to be running at the -10dBv standard. Considering this, perhaps I should use a unity-gain pair like the above ampex units for the balanced input and use a step-up transformer pair to convert the unbalanced input. Would that not make more sense from a gain standpoint?

 

[Steve Eddy]

Personally I'd use the 1:1's for both balanced and unbalanced inputs.

As for unbalanced consumer sources, unless you're talking about old tape decks or tuners or something, the output of a CD player and the like is going to be greater than -10dBV.

Red Book CD output is 2 volts RMS (+8.23dBu) from an unbalanced output.

se

 

[Zaubertuba]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif Personally I'd use the 1:1's for both balanced and unbalanced inputs. As for unbalanced consumer sources, unless you're talking about old tape decks or tuners or something, the output of a CD player and the like is going to be greater than -10dBV. Red Book CD output is 2 volts RMS (+8.23dBu) from an unbalanced output. se

Noted. Thanks!

Another question: In the original transformer input stage, I ended up using four transformers. If I'm following correctly this was to enable a separate ground path for each signal when switched to dual-headphone mode. Am I right on that?

If so...is that the only reason to be using 4 transformers?