Hi All,
I've been visiting these great forums a little over a year now and I thought it was time to discuss some of my builds. The idea was to put something in between a Schiit Modi DAC and Audio Technica ATH-A700X headphones for night listening. I started with a classic: Marcello Pellerano's class A headamp scaled down form a patented power amp (©Nelson Pass). Building it was a lot of fun (designing and etching pcbs) and very instructive e.g. learning how to match MOSFETS pairs which I've never done before. The nice thing is that the amp 'worked' on the first trial. No explosions, no fire, not even smoke although it runs quite hot as expected for class A. When I say 'worked' I am a bit optimistic. There was music coming out of the headphones all right but there were also two additional nuisances, the dreaded hum and hiss. I could get rid of the hum by adding an additional tiny 330pF condensator copper side between base and collector of the ZTX450 (Pass’ variable current source device) as shown here. (post #6)
Without the hum the amp is very dynamic and punchy: it really rocks. I never experienced class A sound before. Feels like driving a Lamborghini on a winding mountain road. I wish I could get rid of the annoying hiss, not loud enough to bother rock music but very distracting for classical. Tried everything I could think of: 100ohm resistor between star ground and signal ground, no star ground, loop ground, metal case, low ripple stabilized psu. No cigar: hiss seems to be here to stay. Any suggestions welcome... The build is visible below: two channels on the left with their separate +24V LM338 low ripple power supplies on the right.
I've been visiting these great forums a little over a year now and I thought it was time to discuss some of my builds. The idea was to put something in between a Schiit Modi DAC and Audio Technica ATH-A700X headphones for night listening. I started with a classic: Marcello Pellerano's class A headamp scaled down form a patented power amp (©Nelson Pass). Building it was a lot of fun (designing and etching pcbs) and very instructive e.g. learning how to match MOSFETS pairs which I've never done before. The nice thing is that the amp 'worked' on the first trial. No explosions, no fire, not even smoke although it runs quite hot as expected for class A. When I say 'worked' I am a bit optimistic. There was music coming out of the headphones all right but there were also two additional nuisances, the dreaded hum and hiss. I could get rid of the hum by adding an additional tiny 330pF condensator copper side between base and collector of the ZTX450 (Pass’ variable current source device) as shown here. (post #6)
Without the hum the amp is very dynamic and punchy: it really rocks. I never experienced class A sound before. Feels like driving a Lamborghini on a winding mountain road. I wish I could get rid of the annoying hiss, not loud enough to bother rock music but very distracting for classical. Tried everything I could think of: 100ohm resistor between star ground and signal ground, no star ground, loop ground, metal case, low ripple stabilized psu. No cigar: hiss seems to be here to stay. Any suggestions welcome... The build is visible below: two channels on the left with their separate +24V LM338 low ripple power supplies on the right.
Attachments
Adventures in Headamp Country, part 2
Because of the half-success with discrete components, I decided to join the Dark Side of the Force.... chip amps! I know chips have a good reputation of silent operation. For instance, when the music is not playing I cannot tell if my LM3886 gainclone is on or off. What about a chip headamp? For this project I decided to trust a dry Texas Instrument app note.
On paper, the ultra-low distortion LM49600 buffer chip looked very promising especially paired with a LM4562. One half of the LM4562 is for the input stage and the other half offices as DC servo in the feedback loop. TI was kind enough to send me free LME49600/4562 samples. The build went well even if it is a bit scary to solder LME49600 skinny spider legs. The yellow conds look a bit cramped: the pcb was planned for smaller ones... Again no problems running the first tests. The result is very good even if not as punchy as the class A amp above. The amazing thing here is the complete absence of background noise: only the darkest of the most abysmal black silence. To continue the above automotive metaphor, listening to this headamp is like driving a big German sedan (a "bmer-cedes"?) on a freshly paved and deserted highway. Just you and the music. The picture below shows a soft start (©Rod Elliott) on top, the amp bottom left (with LME49600's copper side) and the shared regulated +12/-12V psu with LM338s (maybe overkill for this build but I had them laying around).
Box-in is the hard part. I like to recycle old computer junk and found an old optical backup drive enclosure which could fit all the LME49600/4562 stuff without being too cramped. The inside is a bit crappy but the outside does not look too bad even if my paintjob is perfectible. That's it.
Final comment on noise. The LME49600/4562 headamp is in fact so silent that you start noticing technical glitches in records. For instance in Lisa Stansfield's very latest album "Seven" (2014), you can actually hear the temporary noise of additional mixed-in short solo vocals in track #1 "Can't Dance". Amazing.
Thank you for reading.
Jacques
Because of the half-success with discrete components, I decided to join the Dark Side of the Force.... chip amps! I know chips have a good reputation of silent operation. For instance, when the music is not playing I cannot tell if my LM3886 gainclone is on or off. What about a chip headamp? For this project I decided to trust a dry Texas Instrument app note.
On paper, the ultra-low distortion LM49600 buffer chip looked very promising especially paired with a LM4562. One half of the LM4562 is for the input stage and the other half offices as DC servo in the feedback loop. TI was kind enough to send me free LME49600/4562 samples. The build went well even if it is a bit scary to solder LME49600 skinny spider legs. The yellow conds look a bit cramped: the pcb was planned for smaller ones... Again no problems running the first tests. The result is very good even if not as punchy as the class A amp above. The amazing thing here is the complete absence of background noise: only the darkest of the most abysmal black silence. To continue the above automotive metaphor, listening to this headamp is like driving a big German sedan (a "bmer-cedes"?) on a freshly paved and deserted highway. Just you and the music. The picture below shows a soft start (©Rod Elliott) on top, the amp bottom left (with LME49600's copper side) and the shared regulated +12/-12V psu with LM338s (maybe overkill for this build but I had them laying around).
Box-in is the hard part. I like to recycle old computer junk and found an old optical backup drive enclosure which could fit all the LME49600/4562 stuff without being too cramped. The inside is a bit crappy but the outside does not look too bad even if my paintjob is perfectible. That's it.
Final comment on noise. The LME49600/4562 headamp is in fact so silent that you start noticing technical glitches in records. For instance in Lisa Stansfield's very latest album "Seven" (2014), you can actually hear the temporary noise of additional mixed-in short solo vocals in track #1 "Can't Dance". Amazing.
Thank you for reading.
Jacques
Attachments
Hi, thanks for posting 🙂
I built the Zen head amp in 2003 I think, on bread board at first then later with Digi01 boards. I also have a bit of hiss and some hum on the left channel, ive always intended to take it apart and redo the ground wiring but ive never had the real drive to do it probably because I cant hear it over the noisy rock music that I usually listen to.
I have just built a JLH headphone amp with a kit from Silicone Ray, I put it together before Christmas, listened to it a bit but as Ive had no time to put it in a box I just put it away and got back to the Zen.
I have just bought two PCBs from OPC http://www.diyaudio.com/forums/vend...jects-available-here-bal-bal-se-se-lpuhp.html
a SE-SE and a BAL-BAL, Im looking forward to trying opamp based circuits.
Perhaps when I finish the JLH or a wire I will try and improve the ZEN 😱
Have you tried any more head amps?
Cheers
Mark.
I built the Zen head amp in 2003 I think, on bread board at first then later with Digi01 boards. I also have a bit of hiss and some hum on the left channel, ive always intended to take it apart and redo the ground wiring but ive never had the real drive to do it probably because I cant hear it over the noisy rock music that I usually listen to.
I have just built a JLH headphone amp with a kit from Silicone Ray, I put it together before Christmas, listened to it a bit but as Ive had no time to put it in a box I just put it away and got back to the Zen.
I have just bought two PCBs from OPC http://www.diyaudio.com/forums/vend...jects-available-here-bal-bal-se-se-lpuhp.html
a SE-SE and a BAL-BAL, Im looking forward to trying opamp based circuits.
Perhaps when I finish the JLH or a wire I will try and improve the ZEN 😱
Have you tried any more head amps?
Cheers
Mark.
ATH-A700x
Audiophile Closed-back Dynamic Headphones
Sensitivity 100 dB/mW at 1 kHz
Impedance 38 ohms
basically pretty sensitive, low Z headphones 2 Vrms gives ~ 120 dB SPL
most listening should be done at <80 dB
so an "amp" for these headphones should not have big V gain, unity gain would be fine with most consumer digital source
Audiophile Closed-back Dynamic Headphones
Sensitivity 100 dB/mW at 1 kHz
Impedance 38 ohms
basically pretty sensitive, low Z headphones 2 Vrms gives ~ 120 dB SPL
most listening should be done at <80 dB
so an "amp" for these headphones should not have big V gain, unity gain would be fine with most consumer digital source
Hi Mark,
The Wire is based on the same LME49600 that I have used. You will not be disappointed. For me, SMDs are a bit too small to handle and I preferred to build a conventional through hole circuit. For the class A headamp, how did the JLH compare to the Zen as far as noise is concerned? I hesitated between the two and finally picked the ZEN...
In the very old days I built a simple headamp with a single opamp (LM318 I think?) followed by 2N1711 and 2N2905 as outputs. The schematics is still floating around on internet. Noisy.
Rgds,
Jacques
very nice build. i like 🙂
my first DIY is RJM Audio kit The Sapphire. it's like a Porsche on steroids. i have nothing to compare it with apart from various commercial amps i have had threw the years. according to the blurb the Sapphire is class A push-pull. i think the biggest thing i need to address is the power supply i have for it as it's built from cheep components.
soundwise it's just amazing the amount of detail from the very very bottom to the very very top. cymbals have amazed me the most with this amp and no more do they sound like tearing paper but have texture, sound and sizzle of real cymbals. i can also pick out a bass guitar with no problem. i'm stunned with the resolution and authority of the Sapphire. ok so it gets a bit confused now and again and has a funny sort of pre echo on acoustic guitar when it gets hot and a faint grittiness to the top end on complex music that sounds like tearing paper but i'll put those faults down to the power supply.
i have changed a few components from the standard BOM that Richard sends out and this changes everything more to my liking. at the moment i'm using an OPA27 but the kit comes with OPA134 and i must say i prefer the OPA27 as i feel that the OPA134 is a bit too revealing while the OPA27 is more natural sounding, well to my ears anyway.
the source is an OPPO BDP93 and headphones are Beyerdynamic DT-880 600ohm which are not the most forgiving of headphones.
my first DIY is RJM Audio kit The Sapphire. it's like a Porsche on steroids. i have nothing to compare it with apart from various commercial amps i have had threw the years. according to the blurb the Sapphire is class A push-pull. i think the biggest thing i need to address is the power supply i have for it as it's built from cheep components.
soundwise it's just amazing the amount of detail from the very very bottom to the very very top. cymbals have amazed me the most with this amp and no more do they sound like tearing paper but have texture, sound and sizzle of real cymbals. i can also pick out a bass guitar with no problem. i'm stunned with the resolution and authority of the Sapphire. ok so it gets a bit confused now and again and has a funny sort of pre echo on acoustic guitar when it gets hot and a faint grittiness to the top end on complex music that sounds like tearing paper but i'll put those faults down to the power supply.
i have changed a few components from the standard BOM that Richard sends out and this changes everything more to my liking. at the moment i'm using an OPA27 but the kit comes with OPA134 and i must say i prefer the OPA27 as i feel that the OPA134 is a bit too revealing while the OPA27 is more natural sounding, well to my ears anyway.
the source is an OPPO BDP93 and headphones are Beyerdynamic DT-880 600ohm which are not the most forgiving of headphones.
Thanks Bibio!
I read many good things about the Sapphire. Does it get really hot when biased in class A?
Cheers,
Jacques
no, it gets warm but not hot. Richard addressed the problem with putting the OPT's back to back so when one cools the other heats up then transfers the thermal run to the cooler one and so on. i would say the temp is around 30c at full tilt (ear bleeding levels). of course i don't understand any of this.. lol
all i know is that it's one lovely headamp. yes it has it's niggles but these can be addressed. the Standard BOM is very very conservative as to performance but changing a few things here and there really starts to get things to boogie. for instance the 4.7k for volume is more than useless unless you have 16ohm ear buds, i'm up to 36k and its about right.
all i know is that it's one lovely headamp. yes it has it's niggles but these can be addressed. the Standard BOM is very very conservative as to performance but changing a few things here and there really starts to get things to boogie. for instance the 4.7k for volume is more than useless unless you have 16ohm ear buds, i'm up to 36k and its about right.
a little update on the Sapphire. after a brief discussion with Richard he gave me some clue as to the 'transient slip' problem but with a small warning of consequences. well i done what he suggested and all is well with no consequence's 🙂
i would say the little tweak has turned the Sapphire into something very special indeed. Richards new boards already address the problem so you don't have to worry there.
i'm busy opamp rolling and i must say it's very easy to tell the difference between opamps.
i would say the little tweak has turned the Sapphire into something very special indeed. Richards new boards already address the problem so you don't have to worry there.
i'm busy opamp rolling and i must say it's very easy to tell the difference between opamps.
I have been watching this thread. Thanks to all for the info.
I know that the LM49600 is manufactured for audio applications, but I don't want to mess with SMD construction. Is there any advantage to the 49600 over a BUF634 (which is still available in DIP-8 and TO-220 packages 🙂 ) ? I'm thinking of whipping up a small batch of headphone amps, since my modest prototypes (much more antiquated than the circuits discussed in this thread) have been very well received.
I do have some ideas, like using the BUF with a 12Ax7a (yes a tube 😉) front end with global feedback. Yes it's gimmicky, but that's what sells 😉 and I think I can come up with a good product that is better than some of the tube headphone amps I've seen on Flea Bay; some of which looked pretty bad to me. Of course I will experiment with discrete front end and op amp front end too.
So thank you for any advice or comments.
I know that the LM49600 is manufactured for audio applications, but I don't want to mess with SMD construction. Is there any advantage to the 49600 over a BUF634 (which is still available in DIP-8 and TO-220 packages 🙂 ) ? I'm thinking of whipping up a small batch of headphone amps, since my modest prototypes (much more antiquated than the circuits discussed in this thread) have been very well received.
I do have some ideas, like using the BUF with a 12Ax7a (yes a tube 😉) front end with global feedback. Yes it's gimmicky, but that's what sells 😉 and I think I can come up with a good product that is better than some of the tube headphone amps I've seen on Flea Bay; some of which looked pretty bad to me. Of course I will experiment with discrete front end and op amp front end too.
So thank you for any advice or comments.
Hi Fast Eddie,
The LME49600 is a sort of improved BUF with much less loise and much less distortion. Technically it is a SMD but soldering it is not a problem is you own a clothespin to hold it on the pcb. The little legs will then be flush with the printed tracks. It is mounted directly on the copper side. The pcb has to be drawn precisely to account for leg spacing but really, it is not a big deal. I was a bit reluctant myself but all went well and both LMEs were successfully mounted in a matter of minutes.
I have not seen yet hybrid implementations with the LME49600 but it looks like an original and very interesting project.
Rgds,
Jacques
The LME49600 is a sort of improved BUF with much less loise and much less distortion. Technically it is a SMD but soldering it is not a problem is you own a clothespin to hold it on the pcb. The little legs will then be flush with the printed tracks. It is mounted directly on the copper side. The pcb has to be drawn precisely to account for leg spacing but really, it is not a big deal. I was a bit reluctant myself but all went well and both LMEs were successfully mounted in a matter of minutes.
I have not seen yet hybrid implementations with the LME49600 but it looks like an original and very interesting project.
Rgds,
Jacques
The heat sink tab has to be soldered to the board too, correct?
Thank you for your advice. I worked on SMD once, but didn't like it.
Thank you for your advice. I worked on SMD once, but didn't like it.
You can if you want but you don't have to; I didn't. I my case it does not get hot (not even warm) even when played very loud for some time. The initial idea was to squeeze the LMEs between the pcb and the metal casing in order to dissipate heat. But there was not heat to get rid of, at least when operating @ nominal 2x12V. Sizewise, the LME body is close to a TO220, much bigger than those microscopic passive SMD components...
may i draw your attention to a little device called fly tying hackle pliers (the link is for an example).
they are great as heat shunts and for removing/gripping components.
Gripping length a bit short. Grandma's hi-tech wooden clothespin still better suited for this job 😛.
Well I thank everyone for their input. I like to prototype my circuits but I will work something out. I'm going to try the BUF first and do some op amp rolling maybe. I noticed that some of the new generation of op amps are SMD only. 😡 But I bet I can make an excellent headphone amp even with older devices.
Then I might try a hybrid amp. I haven't worked it out yet but I bet I can come up with something. Maybe it's off topic, but has anyone used those miniature sized lower voltage tubes that solder right to the PC board; are they any good for audio? I can do something with the 12ax7 but smaller and lower voltage sure would be a lot more practical. I don't want to make one of those creations with a tube sticking out of the chassis; the warm glow visible through the vents is gimmiky enough. 🙂 I want it to be practical and of course as compact as possible.
Then I might try a hybrid amp. I haven't worked it out yet but I bet I can come up with something. Maybe it's off topic, but has anyone used those miniature sized lower voltage tubes that solder right to the PC board; are they any good for audio? I can do something with the 12ax7 but smaller and lower voltage sure would be a lot more practical. I don't want to make one of those creations with a tube sticking out of the chassis; the warm glow visible through the vents is gimmiky enough. 🙂 I want it to be practical and of course as compact as possible.
Last edited:
- Status
- Not open for further replies.