There's no denying the LME49600 has much higher bandwidth than my discrete diamond buffer, and it probably has lower distortion as well. I do realize it's just a buffer and needs a gain stage in front of it. I thought about trying a combination of my A2 front end with an LME49600 output. But really, I just like playing with discrete circuits, and that's the main reason I've been doing things the hard way. There's plenty of LME49600 composite HPA projects floating around the forum and the internet, so I didn't feel I had much to add there.
The only thing i find trully objectionable in your circuit are the supply voltages...way too high and dangerous for headphones.You should probably consider a higher power buffer and make it a power amplifier.
Have a Neurochrome HP-1 here, which IIRC uses LME49600 output buffers. Compared it to an original Pass HPA-1, a modified HPA-1, and an unnamed HPA proto. The HP-1 is kind of interesting in that it has a 3-position gain switch which appears to change negative feedback (judging by how the sound changes with switch position). Guess which HPAs are most and least preferred for listening?
Im going to go with hp-1. Well implemented 49600 composite is pretty much sota from what i know
Interesting comment. i wasn't expecting this. Plus and minus twelve to eighteen volts seems to be pretty common for headphone amp projects. Do you mean "dangerous" as in frying the cans, or are you talking about hearing damage?
I wanted to have enough current and voltage swing to drive high-impedance and/or low-efficiency headphones, and overall to have plenty of headroom, in theory to improve linearity. You can make a good argument that this is overkill. I'd be interested to know your opinion on how much voltage is enough.
If you think the +-17V supply in the A1/A2 is too high, you must be very alarmed about the +-24V in the HPA-1. Interestingly, the A3 prototype clips at about the same voltage as my other amps even with its higher supply. You lose about 4-5 Volts due to the MOSFET gate bias.
Thanks for the suggestion, but I didn't want to build a speaker amplifier. Probably an even easier solution would have been to scrap the DIY entirely and just buy a Topping A30 Pro. Oh, wait. I did buy a Topping A30 Pro. Listening to it now. Sounds nice.
Lets look at a LME49600 ....
1) basically a diamond buffer with protection circuitrary (can be bad?)
2) great bandwudth (but do we need it?...this is not a video device we are building)
3) way underbiased for our application
4) for best results needs to be in a global feedback loop (complexity)
5) it's predecessor the BUF634 and 2002 sound better (why?)
6)....and so on.
I humbly suggest that you do not let distortion and other non-related specs. guide your design. This is all I have to say on the matter. 🙁
Jam
1) basically a diamond buffer with protection circuitrary (can be bad?)
2) great bandwudth (but do we need it?...this is not a video device we are building)
3) way underbiased for our application
4) for best results needs to be in a global feedback loop (complexity)
5) it's predecessor the BUF634 and 2002 sound better (why?)
6)....and so on.
I humbly suggest that you do not let distortion and other non-related specs. guide your design. This is all I have to say on the matter. 🙁
Jam
So 250mA aren't enough for headphones...underbiased...that's crap.I can bet it doesn't need any feedback loop to sound perfectly fine on any headphones EVER made...I mean no audible distortion for a human being...My germanium hphamp runs on +-3 v at 13 mAmps idle current on 250 ohms cans and it's damn loud and doesn't need any more base.
I think I proved I can design a discrete amp that is pretty much indistinguishable from SOTA commercial amps (e.g., Topping). Tom Christiansen sells his HPA-1 (great name...) for $899. Why spend that much money when you can buy a JDS Labs Atom+ for a hundred bucks?
I'm not necessarily a die-hard audiophile and could be happy with my A30 Pro, and then move on to other pursuits. The HPA-1 clone sounds different. I'm confident it's not imagination. I'm doing this project now, initially at Mark's bidding, because I wanted to better understand why some people like things that aren't all the same.
I am happy for you. Fortunately, nothing I'm doing here is keeping you from enjoying your project, and vice-versa.
Most of my career was spent testing and fixing things and if i'd have a 1700 bucks pair of headphones the last thing i want to see it's getting them fried by an accidental short circuit.There's quite a while already since i only consider capacitor coupling for headphones no matter the quality and price and i never screw it.My cans are only 200 bucks but in my country these are really expensive cans...
What has bias got to do to do with current capacity? If class B amps are your thing more power to you. 🙄
I see that this will be a pointless debate.....some people say that ignorance is bliss.
Regards,
Jam
I see that this will be a pointless debate.....some people say that ignorance is bliss.
Regards,
Jam
I agree there's a risk any time you plug expensive headphones into a DIY amplifier. There's risk with commercial amps, and speaker amps, too. Worse would be burning down the house or electrocuting someone! You have to decide what level of risk you can live with. I try to do quality construction work, and I test with cheap headphones at first. Even my "good" headphones weren't that expensive.
I don't want to knock someone else's project (though I reserve the right to do so in the future, haha). I'm glad you're getting good results with the germanium design and that the low voltages and currents involved give you peace of mind.
18ma of bias is a lot for 250 ohms cans as mine.Even for 32 ohms depending on sensitivity might be a lot but a conservative 50mA of bias will drive any headphones maybe exept some Hifiman which usually need around 16 x more power than usual 32 ohm cans. But 18ma for 64...600 ohms studio grade headphones is enough to keep them in class A most of time in their whole power range.Maybe i was too radical about saying " ever made" , but most well known good quality headphones won't need more than that to spend most of their time in class A.
Instead of arguing the only way to know what matters for audio is to do listening tests with everyone at the same time and place using the same equipment. Since we can't practically do that, what we are more or less left with is to build some identical amplifiers and compare notes. Pass HPA-1 is a successful commercial design. Its still the standard HPA used by Stereophile for evaluating headphones. They explain why they choose to use it. Beyond that it won multiple awards including a couple of 'product of the year' awards. Theory and AP measurements aside, there is something or other going on with it that has made it exceptional. That is to say, its not completely logical to discount all of that in favor of AP 'figure of merit' numbers. My suggestion to Henry was to try it and see for himself. Since a schematic was reverse engineered and published in the forum, anyone is free to give it a try and form their own opinions. Otherwise its just arguments about theory and models. IMHO most people who buy HPAs care more about the actual listening experience.
The HPA-1 was designed to drive almost any type of headphone, being a $3500.00 device, has to peform with impedances as low as 12ohms and some pretty ineffecient headphones as well. If you understand the design brief you will understand that the LME49600 is not suitable for the task.....I have tried it with dissapointing results. It might be suitable for some inexpensive headphones but not in the arena the HPA-1 is competring in.
Last edited:
It sounds like you haven't tried enough mate! I designed hybrid commercial amplifiers for 3...5000 bucks headphones I'm not even trying to disband a sound circuit, only the necessity of its high supply voltage for dynamic headphones.I have a friend driving 1 cubic meter horns equiped with 8 ohm Fostex Fe 206 speakers with both paralleled channels of a Tpa 6120 for almost 10 years now...And he never felt the need for a higher power.Well lme49600 has 18ma idle current while tpa 6120 has 15ma idle current just that it has an internal op amp which you can attach externally to lme49600.As far as i know the cleanest well known amp till today is still tpa6120...I have no idea what high quality means for you...I know what high end can mean, but high quality means only one thing: no distortions that you can hear.
I'm not knocking any project...with the right transistors your design can be supplied with +-5 v with no problem and no significant distortions.There's no wise decision in keeping transistors hot or even warm.If it would be so no ic on the market would hit 0.000x % distortions before its 50 trz would melt.Class A is invoked to lower crossover distortions in push pull circuits not to enhance any harmonic response..No circuit can have lower thd and pretend being high quality amplifiers without solving crossover distortions first but higher voltage supply is offering you exactly this : lower distortions and high slewrate.If you can get it cheaper and simpler why bother building fantastically complicated designs that waste electricity for something you can get out of a few batteries?