That question occurred to me as well during the design phase. However, the input offset of the LME49600 is specified to ±60 mV. So worst case, half of the available output current would be consumed by the standing current between the two amp halves, assuming 1 Ω ballasting resistors. That results in poor performance. Not my cup of tea. 
It would be lovely if you could just put the LME49600s in parallel as that would lower the BOM cost considerably.
Tom
It would be lovely if you could just put the LME49600s in parallel as that would lower the BOM cost considerably.
Tom
Hey! Don't go all engineer on me!!
You are correct that the 250 mA (peak), 15-16 V (peak, assuming ±18 V rails) that the LME49600 can deliver by itself is more than enough for most headphones. Yet, some want more power. Many want 3-5 W into a 20 Ω load, which a single LME49600 cannot deliver. Do they need that much power? Most likely not. This is more about perceptions and "wants" rather than reality and needs.
With the HP-1, I aimed squarely at the high-end market. During the product concept phase I debated whether to start small or go all in. I decided the latter. I hope it pans out. I've sunk quite a bit of cash into the development...
BTW: Hot off the APx525:
* THD @ 200 mW into 600 Ω: 0.000017 % (-135 dB)
* Max. power into 600 Ω before clipping: 230 mW
Tom
You are correct that the 250 mA (peak), 15-16 V (peak, assuming ±18 V rails) that the LME49600 can deliver by itself is more than enough for most headphones. Yet, some want more power. Many want 3-5 W into a 20 Ω load, which a single LME49600 cannot deliver. Do they need that much power? Most likely not. This is more about perceptions and "wants" rather than reality and needs.
With the HP-1, I aimed squarely at the high-end market. During the product concept phase I debated whether to start small or go all in. I decided the latter. I hope it pans out. I've sunk quite a bit of cash into the development...
BTW: Hot off the APx525:
* THD @ 200 mW into 600 Ω: 0.000017 % (-135 dB)
* Max. power into 600 Ω before clipping: 230 mW
Tom
Just looking at your site and I have to admit I'm quite amazed at the absolutely ABSURD pricing of this "HP-1 Headphone Amplifier".
$225 just for a bare PCB or $899 for a fully assembled PCB???
To make matters worse, you're using one of the worst volume pots in existence for channel imbalance, especially at lower levels.
While you may have spent a lot of money designing this thing with "vanishingly low distortion", it all comes down to SQ when the dust clears.
Since you're a former National engineer, perhaps you're partial to the LME49600 and feel that it's the "ultimate" buffer chip.
On paper and in actual testing it could very well be, but specs. and SQ don't always coincide.
I've tried several iterations of the '49600 with other op-amps(including the OPA1612 and numerous other BJT and JFET op-amps) within its feedback loop and to say the resulting sound was "STERILE" would be an understatement.
Perhaps some people are tricked into believing the '49600's sterile sound is "accurate" or "transparent"...I don't know.
What I do know is that I didn't waste anymore of my time trying to optimize or improve a circuit using the sterile sounding '49600 as a buffer.
Too bad T.I. didn't discontinue the '49600 when they dicontinued some of the other popular National chips a while back.
$225 just for a bare PCB or $899 for a fully assembled PCB???
To make matters worse, you're using one of the worst volume pots in existence for channel imbalance, especially at lower levels.
While you may have spent a lot of money designing this thing with "vanishingly low distortion", it all comes down to SQ when the dust clears.
Since you're a former National engineer, perhaps you're partial to the LME49600 and feel that it's the "ultimate" buffer chip.
On paper and in actual testing it could very well be, but specs. and SQ don't always coincide.
I've tried several iterations of the '49600 with other op-amps(including the OPA1612 and numerous other BJT and JFET op-amps) within its feedback loop and to say the resulting sound was "STERILE" would be an understatement.
Perhaps some people are tricked into believing the '49600's sterile sound is "accurate" or "transparent"...I don't know.
What I do know is that I didn't waste anymore of my time trying to optimize or improve a circuit using the sterile sounding '49600 as a buffer.
Too bad T.I. didn't discontinue the '49600 when they dicontinued some of the other popular National chips a while back.
Last edited:
I don't disagree that a zero distortion amp sounds sterile. I for one, like a little second harmonic - not much, maybe 500 to 1000ppm. Doesn't take much to remove sterility. But it's not really fair to say that Tomchr's pricing of a one of a kind amp is absurd. I know how much work goes into development - it's a lot of time. Folks are welcome to buy it or not. But they should not complain it's too expensive if it's not their cup of tea. Why complain if you won't even buy it if it were priced low. Speakers have the same issue. Some people like non-flat frequency response because their ears have lost some of the HF acuity. So they like some zing - but it's not what the recording engineer meant for us to hear. If one wants truth in sound. Sterile is the norm. If one wants to sit back and enjoy, then let's add some H2 secret sauce to the sound. Maybe a tiny bit H3.
Here is my latest 2 transistor pocket class A amp. It has just enough secret sauce to satisfy:
Here is my latest 2 transistor pocket class A amp. It has just enough secret sauce to satisfy:
Last edited:
@Ammel68: Not that you're biased or anything, right?
Yeah, if your reference is a vacuum tube amp or LDR volume control that add a good percent or two of THD, thus, adding "space" or fuzz to the source material, you'll likely be disappointed by a precision amp, such as the HP-1. Each to his own.
There's an overwhelming body of research pointing to good measurements (plural) correlating with a positive listening experience. Sadly much of it is in academic journals, thus, difficult to get at without a subscription. However, Sean Olive and Floyd E. Toole do publish quite a bit in the public domain.
Also note that I don't work for free. I don't know anybody who does. Neurochrome is my passion and only source of income (by choice). I'm working for well below minimum wage on these projects, but I'm having fun.
Tom
Yeah, if your reference is a vacuum tube amp or LDR volume control that add a good percent or two of THD, thus, adding "space" or fuzz to the source material, you'll likely be disappointed by a precision amp, such as the HP-1. Each to his own.
There's an overwhelming body of research pointing to good measurements (plural) correlating with a positive listening experience. Sadly much of it is in academic journals, thus, difficult to get at without a subscription. However, Sean Olive and Floyd E. Toole do publish quite a bit in the public domain.
Also note that I don't work for free. I don't know anybody who does. Neurochrome is my passion and only source of income (by choice). I'm working for well below minimum wage on these projects, but I'm having fun.
Tom
Reminds me of Doug Self's notion of a niceness knob. I occasionally toy with the idea of a solid state tube emulator with variable THD. Want clean, bypass the circuit. Want fuzz/space/lazy, just flip a switch and dial it in to your liking.
The HP-1 is all about "straight wire with gain". No additions. Just the source material, reproduced.
Tom
The output impedance of the '600 is 10Ohms or so, therefore much less cross current. We did a HP design with three '600's in parallel with 1R ballast resistors and had zero problems with excessive cross current.
FWIW, TI app note for paralleled BUF634 don't even show ballast resistors, exactly for that reason (and that's were 10Ohms number comes from).
Last edited:
@Ammel68: Not that you're biased or anything, right?
Also note that I don't work for free. I don't know anybody who does. Neurochrome is my passion and only source of income (by choice). I'm working for well below minimum wage on these projects, but I'm having fun.
Tom
I don't expect anyone to work for free.
From reading some of your past posts, I know that you're very knowledgeable.
I also understand that PCB layout is both tedious and time consuming from laying out my own rather simple projects.
I consider it an art and have the upmost respect for people who are good at it.
Board cost is also quite expensive when it's larger than 100mm x 100mm.
As long as you enjoy what you're doing and making some money to boot, that's what really counts.
I don't see the output impedance of the LME49600 specified in the data sheet. I would be really careful about relying on the output impedance of the LME49600 by itself in a design as it is not a controlled parameter. That doesn't mean it cannot be relied on ... in one build ... but I would not rely on it in a production design.
Regardless of the output impedance, you're still applying up to 120 mV across the two output impedances if you put two LME49600s in parallel. That's not a good design in my view. Not saying it can't be done, or that it's "wrong" somehow; only that I prefer a better solution.
The BUF634 is a completely different design than the LME49600. What applies to the BUF634 may/may not apply to the LME49600.
Tom
Last edited:
Its funny how some people think anything on this forum should be almost free. Here's tom designing an amp that will probably leave multi thousand dollar boutique designs for dead in the long grass, yet people think he's over charging.
10k for a Stax valve energiser because it contains Japanese distributor marketing stories. Toms hpa offers better measured performance for a tenth of the cost.
10k for a Stax valve energiser because it contains Japanese distributor marketing stories. Toms hpa offers better measured performance for a tenth of the cost.
Thank you! Well said
Best,
Anand.
Yes, I totally agree.
Sq225917, I never said anything about "free" or "almost free" in my initial post, but feel free to make all the assumptions you want.
A simple design based on TI datasheet (49600 + 4562 + servo) I built some years
back (see here) sounds already pretty darn good for 10$ worth of components.
I am wondering if this new design sounds hundred times better
Interesting build with a servo. Did you ever post the schematic for it?
The schematic is in the LME49600 data sheet. National used the LME49720 for the DC servo which almost defeats the purpose. I've built that circuit. It works quite well, though it can be improved.
The OPA1611 provides better performance than the LME49720/LM4562 but does have a little phase wobble in the 1-10 MHz range that you have to deal with. Tradeoffs, tradeoffs.
The HP-1 contains $250 worth of parts, not including the PCB or the chassis, neither of which is cheap. I'm not sure where you get the $10 figure from. You can't even buy the chips for that. Besides, as I've noted earlier, I don't work for free. I'm in a low volume, niche market, so I have to have high margins to survive. Sorry. That's reality.
Tom
The OPA1611 provides better performance than the LME49720/LM4562 but does have a little phase wobble in the 1-10 MHz range that you have to deal with. Tradeoffs, tradeoffs.
The HP-1 contains $250 worth of parts, not including the PCB or the chassis, neither of which is cheap. I'm not sure where you get the $10 figure from. You can't even buy the chips for that. Besides, as I've noted earlier, I don't work for free. I'm in a low volume, niche market, so I have to have high margins to survive. Sorry. That's reality.
Tom
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.