[Zaubertuba]

LOL - Now that *is* funny.

 

[Zaubertuba]

Quote:

Originally Posted by amb /img/forum/go_quote.gif Whether summing with a transformer or summing via a summing amplifier (or some other means), the effect would be the same. The sum of two signals is what it is. You cannot simply connect the L- and R- outputs together, though. The two amplifier's outputs would "see" each other and fight. One or both of them would most likely be damaged as a result.

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif No, it's not the same. If you common the L- and R- outputs of a pair of bridged amplifier channels as would be the case with a four board B22 or similarly configured amplifier, the two channels see each others' outputs as an effective short circuit and output current from one channel will flow through the output of the other and vice versa. If you common the bottom secondary windings of a pair of output transformers, neither secondary sees the other and there is no current flowing from one channel's secondary into the other channel's secondary.

O.K., Trying to get back on track.

I'm an experimenter at heart, so I'd be interested in integrating an additional "mode" like this, if just for the purposes of being able to A/B it with other configurations for future designs. That is, if I could indeed pull it off without blowing anything up.

It seems to me, from my extremely "layman's" perspective, that such a mode would be a more "proactive" approach than an active ground.

But, I'm not well versed enough to understand if you common transformer secondaries, why is that not a short?

 

[Zaubertuba]

Quote:

Originally Posted by amb /img/forum/go_quote.gif You would be creating a "ground" that's swinging an out-of-phase version of the MONO signal. Think about the implications -- I don't think you'd really want to do this.

O.K. I've been thinking, maybe I didn't fully grasp this before. With an active ground, the sinking/sourcing of current on the ground board is dependent on the L+/R+ signals. The board will only sink/source what it needs based on the positive phase signals.

Summing L-/R-, though, there's really no way to split the signal back apart into it's respective L/R components once it gets to the headphone cable, is there? So you really would be sending a mono negative phase signal to both sides.

Maybe that isn't such a great idea. Am I missing something?

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif But, I'm not well versed enough to understand if you common transformer secondaries, why is that not a short?

Well, in the strictest technical sense, it is a "short" between the bottoms of the two transformer windings. But it's not a "short" in the same context as when you common the L- and R- outputs on a pair of bridged amplifier channels.

The bridged amplifier is effectively made up of four distinct amplifiers. If you common the L- and R- outputs, you're putting two amplifier outputs in parallel with each other. Which means that the L- output sees the R- output and vice versa. And since their output impedances are very low, it's tantamount to simply shorting their outputs to ground.

Now, if you common the bottoms of the secondary windings of a pair of output transformers, you're not doing the same thing as you would above. You don't have four separate amplifiers. You simply have the two secondary windings, which are in series, and not parallel.

It's like if you took two unbalanced mono amps and tied their output ground terminals together. Only with the transformers, you'd still be driving the phones from a balanced source.

se

 

[Zaubertuba]

I'm kind of starting to bark up the wrong tree, methinks. Two discrete headphone outs is still a major criteria for this design. So, barring summing outputs, are there other benefits inherent in using transformers in the output stage? Do they effect output impedance with respect to high- or lo-impedance headphones, for example?

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif I'm kind of starting to bark up the wrong tree, methinks. Two discrete headphone outs is still a major criteria for this design.

If that's what you want then that's what you should do.

Quote:

So, barring summing outputs are there other benefits inherent in using transformers in the output stage?

Their allowing you to drive TRS terminated phones from bridged amplifier outputs was the main reason I brought them up.

Quote:

Do they effect output impedance with respect to high- or lo-impedance headphones, for example?

Their winding resistance would add to the output impedance of whatever was driving them. If you used something like a Jensen or CineMag quadfilar line output transformer, the winding resistances would add to either 20 ohms if you paralleled the primaries and secondaries or 80 ohms if you kept them in series.

se

 

[Zaubertuba]

Ugh! I think I'm going to have to hold off of the transformer input-stage. It still adds $180 to the build--way over my original budget. I just won't end up getting it built at this rate.

Are there other "middle-ground" options for an unbalanced line receiver? A discrete component circuit, perchance?

Working on rev."D", BTW. Complete with 4-pin XLR.

 

[cobaltmute]

Unbalanced to balanced?

The Dynalo shows a way of going from un-balanced to balanced (see how Figure 2 ties into the main schematic to create the negative output)

 

[Zaubertuba]

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif Unbalanced to balanced? The Dynalo shows a way of going from un-balanced to balanced (see how Figure 2 ties into the main schematic to create the negative output)

Well, that does show a very integrated solution with Kevin's design, but I need a converter as a separate input stage, before the selector switch. Also, I'm trying to find a discrete solution, without any opamps in the signal path.

I don't know, looking at the OP27's block diagram, maybe that's really not possible on the cheap, either.

As an aside, it's interesting how Kevin's circuit uses the OP27 in reverse.

 

[Zaubertuba]

Or...I found this:

EDCOR - PCW600/600

Certainly not a Cinemag or Jensen, but freq. response looks allright, and it's probably more in line with the rest of the build.

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif Or...I found this: EDCOR - PCW600/600 Certainly not a Cinemag or Jensen, but freq. response looks allright, and it's probably more in line with the rest of the build.

Yeah.

Though don't know what the winding resistance is on those, which will be something worth considering if you're going to use it on the output.

I called Edcor several weeks ago and asked them what the winding resistances were on that exact model transformer.

They guy said he didn't "have that information on file" and had absolutely no interest in getting it.

I just hung up and said **** 'em.

se

 

[Zaubertuba]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif Yeah. Though don't know what the winding resistance is on those, which will be something worth considering if you're going to use it on the output.

Well, for now they'll just be for the input--so I suppose I'm O.K. It's a 1:1, but is the 600 Ohm input and output resistance appropriate for an input stage?

Also, how do I go about calculating the values for the resistors across the secondaries and the (I assume) shunt resistors? I found this, but with the shunt in series with the primary and what (I again assume) appears to to be a load resistor across the secondary this appears to be an output stage and not exactly applicable.

Since these transformers are not really high-end, I'm also wondering if I'm better off using them exclusively for just the unbalanced inputs (my unbalanced sources are relatively lo-fi, anyway). I don't want to *degrade* the signal coming from my 1212m. Although, I've read having a well-matched receiver complement can be important to the performance of a balanced cable. Yes, that's the sound of me waffling again.

 

[mugdecoffee]

One thing you might want to think about for the unbalanced input is just grounding the negative input so you'd basically have two active ground channels. You'd "lose" half the signal voltage compared to a balancing scheme (such as the input transformer) but the upside is that its practically free, just add a DPDT switch. You'd still get the channel separation and more current output.

BTW I've ordered 4 JISBOS's for a very similar amp which I'll hopefully integrate with my proto LCDuino. This thread is definitely giving me ideas.

 

[luvdunhill]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif Also, how do I go about calculating the values for the resistors across the secondaries and the (I assume) shunt resistors?

a square wave generator and oscilloscope. Also, my guess is you'll need a RC circuit across the secondary.

 

[Steve Eddy]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif Well, for now they'll just be for the input--so I suppose I'm O.K. It's a 1:1, but is the 600 Ohm input and output resistance appropriate for an input stage?

Not unless you're driving a 600 ohm input, which you're not.

Go with either the 10k/10k or the 15k/15k model.

Quote:

Also, how do I go about calculating the values for the resistors across the secondaries and the (I assume) shunt resistors?

The load is going to be the input impedance of your amp, which will ultimately be the value of whatever you're using for your volume control.

If it's 10k, the transformer may be well behaved (i.e. no significant high frequency resonance) and you don't have to do anything further. If it's much higher, or if the transformer's not well behaved with a 10k load, then you may need to add an RC snubber network to tame the resonance.

To evaluate this, you'll need what luvdunhill says above, a signal generator and a 'scope.

That's the problem with Edcor. About the only useful information they provide about their audio transformers are their physical dimensions.

Quote:

I found this, but with the shunt in series with the primary and what (I again assume) appears to to be a load resistor across the secondary this appears to be an output stage and not exactly applicable.

That series resistor shown in the illustration is only used as a means of measuring the transformer's input impedance, and can be applied to any transformer, whether an input transformer or an output transformer.

Quote:

Since these transformers are not really high-end, I'm also wondering if I'm better off using them exclusively for just the unbalanced inputs (my unbalanced sources are relatively lo-fi, anyway). I don't want to *degrade* the signal coming from my 1212m.

I certainly wouldn't use the Edcors for anything I was serious about.

Quote:

Although, I've read having a well-matched receiver complement can be important to the performance of a balanced cable. Yes, that's the sound of me waffling again.

That's cool. I have plenty of butter and maple syrup!

se