Hi Davide,
I'm glad to hear you narrowed it down to an HF oscillation and not DC! Now we can try and figure out how to help you!
I should have realized that it could have been that, and your multimeter was misrepresenting the presence of HF as DC.
A few tips that might help:
1 - Be very mindful of wiring. Use proper shielded balanced cables between the legato and The Wire.
2 - Don't let the output go anywhere near the input.
3 - Try lifting the BW pin as qusp suggested
4 - Toss a 100pF cap in parallel (on top of) the input R's, or add one in parallel with the feedback resistor on the second stage, and one in parallel with the resistor to ground on the + input of the second stage. All of these will reduce the likelihood of HF oscillation.
I'm quite surprised you're seeing this, and I would really focus on 1, 2 and 3 before resorting to the last option. As qusp mentioned, this is a very wide BW amp, and requires close attention to be paid to the way you wire it up. Sloppy input wiring, or looping the output back near the input can cause HF oscillations.
Keep us updated on your progress with this! I'd like to know what ultimately solves the issue.
Regards,
Owen
I'm glad to hear you narrowed it down to an HF oscillation and not DC! Now we can try and figure out how to help you!
I should have realized that it could have been that, and your multimeter was misrepresenting the presence of HF as DC.
A few tips that might help:
1 - Be very mindful of wiring. Use proper shielded balanced cables between the legato and The Wire.
2 - Don't let the output go anywhere near the input.
3 - Try lifting the BW pin as qusp suggested
4 - Toss a 100pF cap in parallel (on top of) the input R's, or add one in parallel with the feedback resistor on the second stage, and one in parallel with the resistor to ground on the + input of the second stage. All of these will reduce the likelihood of HF oscillation.
I'm quite surprised you're seeing this, and I would really focus on 1, 2 and 3 before resorting to the last option. As qusp mentioned, this is a very wide BW amp, and requires close attention to be paid to the way you wire it up. Sloppy input wiring, or looping the output back near the input can cause HF oscillations.
Keep us updated on your progress with this! I'd like to know what ultimately solves the issue.
Regards,
Owen
i'm quite surprised as well, for my first test and the way its still set up since yesterday, i just lashed it together, no signal ground connection between ivy 3 and the wire, signal wires pretty short and direct, but not twisted. i have dead silent background and about 1mv dc.
i did find last night that it draws a bit more current than i thought it would, i'm running both the ivy and the wire with a single amb sigma22 bipolar supply set for +/-12vdc and while in use the psu hardly gets warm, but i walked away and left it without any music playing and came back to find the s22 radiating quite a bit of heat, like too much heat, i guess because its not a class a amp and without the wire chewing the energy and putting it into the headphones the mosfets (which ARE running class A) and heatsinks on the psu had to dissipate too much. at least thats what i figured. with the ivy 3 and the wire i guess its requirements are about 350-400ma (+/-175-200ma) for ivy and about +/-400+ma for the wire? probably it could use some larger sinks on the s22 if i'm to keep using it like that, but i'll probably just knock up a perfboard lt1764a/lt1175 reg for it to keep it compact.
even with chewing 600ma @ +/-12v its within spec for the s22, but probably not with only 4 x 1.5" board mount heatsinks
i did find last night that it draws a bit more current than i thought it would, i'm running both the ivy and the wire with a single amb sigma22 bipolar supply set for +/-12vdc and while in use the psu hardly gets warm, but i walked away and left it without any music playing and came back to find the s22 radiating quite a bit of heat, like too much heat, i guess because its not a class a amp and without the wire chewing the energy and putting it into the headphones the mosfets (which ARE running class A) and heatsinks on the psu had to dissipate too much. at least thats what i figured. with the ivy 3 and the wire i guess its requirements are about 350-400ma (+/-175-200ma) for ivy and about +/-400+ma for the wire? probably it could use some larger sinks on the s22 if i'm to keep using it like that, but i'll probably just knock up a perfboard lt1764a/lt1175 reg for it to keep it compact.
even with chewing 600ma @ +/-12v its within spec for the s22, but probably not with only 4 x 1.5" board mount heatsinks
Last edited:
So, as it's better not to go out today, and family already evacuated, I had all the time to chase this problem.
Actually I solved it putting a 47 ohm carbon resistor for each input, as a stopper. Now I have no bias/oscillation. I can still see something with the scope, it's not perfectly clean compared to the on-board converter of the legato, but I think is just because I have flying wires and no case.
I am listening now with my F5. It sounds very good. At first glance the bass are better, fuller.
Another thing, you should try the dual mono. It's a big improvement. I am using two buffalo II and one legato. Cannot wait to try the D1.
One question: how do you advice to to the grounding ? Should I connect the ground to the safety earth with a termistor "a la Pass" ?
So I can officially say that I successfully built my first SMD project.
Thanks,
Davide
Actually I solved it putting a 47 ohm carbon resistor for each input, as a stopper. Now I have no bias/oscillation. I can still see something with the scope, it's not perfectly clean compared to the on-board converter of the legato, but I think is just because I have flying wires and no case.
I am listening now with my F5. It sounds very good. At first glance the bass are better, fuller.
Another thing, you should try the dual mono. It's a big improvement. I am using two buffalo II and one legato. Cannot wait to try the D1.
One question: how do you advice to to the grounding ? Should I connect the ground to the safety earth with a termistor "a la Pass" ?
So I can officially say that I successfully built my first SMD project.
Thanks,
Davide
glad you sorted it all out, guess you'll have to find something else to distract you now 
mine is only grounded through power supply for the moment, signal ground is floating, but then i'm not using it for a se convertor. just follow the grounding scheme you have used with the rest of your build, power supply grounded to your power supply star ground, signal ground connected to the signal star ground and then the whole lot connected together with digital ground as well and in turn to safety ground through a thermistor. thats how i do it anyway.
stay safe mate
mine is only grounded through power supply for the moment, signal ground is floating, but then i'm not using it for a se convertor. just follow the grounding scheme you have used with the rest of your build, power supply grounded to your power supply star ground, signal ground connected to the signal star ground and then the whole lot connected together with digital ground as well and in turn to safety ground through a thermistor. thats how i do it anyway.
stay safe mate
Last edited:
yeah without going to more upmarket types, the neutriks are the best of the bunch. for cable connectors, the neutrik XX series male and the female are the best imo. i've linked you the gold plated version, but they are also available in silver. the difference between the x and xx series is purely mechanical and cosmetic, they are the latest one, but they have to continue to stock the old ones because these are out there in the millions in recording studios etc.
for panel mount here is the male and female
yes there are hundreds, particularly the panel mounts, because of the reason above, but also because there are several different models, each of those comes with either silver or gold plate (though annoyingly you cannot get male silver plated panel mount DL type with black chrome housing, only in the nickel housing, although you can get the female. at least at mouser anyway. each of those models comes in pcb mount in straight or right angle, solder cup, screw terminal. the ones above are black chrome with gold plate and soldercup, also allows you to mount the from the rear of the panel and be flush mount
are you going dual mono with the amps too? because personally if not i would go with a single 4 pin xlr, rather than 2 x 3 pin. or better still, if you have total control of your setup and have built all elements, use a 25 pin or 9 pin dsub, or a small multipin lemo connector, or mini 4 pin xlr.
one thing with balanced also, forget ground unless you need the ground as a reference for the signal between the 2 devices, its more trouble than its worth imo, just leave it out, it does nothing except create another possibility for a ground loop and increase cable capacitance. i would say perhaps you should use the shield, given you have proved to have enough interference to cause oscillation, but leave out ground, so pin 1 is nothing, the shield connects to the housing/shield tab (not all connectors have this, but standard ones do) and chassis ground, thats all.
i mostly dont even use a shield, as balanced signals coupled with well designed balanced gear has the advantage of common mode noise rejection and will delete noise that is common to both phases. for me they cause more trouble than they are worth, but i dont have so much problems with interference. some signals like aes3 balanced digital need the ground reference or they dont function and some gear might need a bias, but most differential audio signals do not and without those you can create a very low capacitance, cheap, thin and quick to make cable; this carries to using other types of connector too; 2 x thin twisted pairs and thats it.
for panel mount here is the male and female
yes there are hundreds, particularly the panel mounts, because of the reason above, but also because there are several different models, each of those comes with either silver or gold plate (though annoyingly you cannot get male silver plated panel mount DL type with black chrome housing, only in the nickel housing, although you can get the female. at least at mouser anyway. each of those models comes in pcb mount in straight or right angle, solder cup, screw terminal. the ones above are black chrome with gold plate and soldercup, also allows you to mount the from the rear of the panel and be flush mount
are you going dual mono with the amps too? because personally if not i would go with a single 4 pin xlr, rather than 2 x 3 pin. or better still, if you have total control of your setup and have built all elements, use a 25 pin or 9 pin dsub, or a small multipin lemo connector, or mini 4 pin xlr.
one thing with balanced also, forget ground unless you need the ground as a reference for the signal between the 2 devices, its more trouble than its worth imo, just leave it out, it does nothing except create another possibility for a ground loop and increase cable capacitance. i would say perhaps you should use the shield, given you have proved to have enough interference to cause oscillation, but leave out ground, so pin 1 is nothing, the shield connects to the housing/shield tab (not all connectors have this, but standard ones do) and chassis ground, thats all.
i mostly dont even use a shield, as balanced signals coupled with well designed balanced gear has the advantage of common mode noise rejection and will delete noise that is common to both phases. for me they cause more trouble than they are worth, but i dont have so much problems with interference. some signals like aes3 balanced digital need the ground reference or they dont function and some gear might need a bias, but most differential audio signals do not and without those you can create a very low capacitance, cheap, thin and quick to make cable; this carries to using other types of connector too; 2 x thin twisted pairs and thats it.
do not connect to one side, sounds like you need to read the sticky on system grounding, only EVER do that if you have grounding problems with systems out of your control like commercial gear and you need to fix temporarily. being a diyer you should sort those issues out first before trying a stopgap thing like connecting shield one side imo. the shield should only be connected to the chassis ground anyway (on both sides).
there is no danger of ground loops unless you have done something wrong and if you have done something wrong and connected the chassis to power supply or signal ground directly, rather than through a loop breaker. connecting it on one side will do nothing anyway if you have ground connected; it will just flow back up the shield and back up the ground wire again, acting as an aerial and shunting all of the noise to signal ground. often people even recommend doing that with an rca cable, when ground and shield are one and the same thing and connected directly to each other. its a common old wives tale/myth, that i wish could be stamped out once and for all.
/rant
anyway this is OT
there is no danger of ground loops unless you have done something wrong and if you have done something wrong and connected the chassis to power supply or signal ground directly, rather than through a loop breaker. connecting it on one side will do nothing anyway if you have ground connected; it will just flow back up the shield and back up the ground wire again, acting as an aerial and shunting all of the noise to signal ground. often people even recommend doing that with an rca cable, when ground and shield are one and the same thing and connected directly to each other
. its a common old wives tale/myth, that i wish could be stamped out once and for all./rant
anyway this is OT
Last edited:
This might be of some help: Grounding and Shielding Audio Devices
and: Sound System Interconnection
regards,
sam
and: Sound System Interconnection
regards,
sam
yeah read all that long ago, but they will never recommend to omit the ground on this cable as i have recommended, because they have a one size fits all look. in this case there is absolutely no need for the ground connection, there is no return signal current on the ground pin needed going from 'the wire' to his amp, so there is no reason to use it. the shield may be handy, but should only be connected to chassis ground and the barrel of the xlr connector, not pin 1
Last edited:
indeed there is good info about that and many other subjects on the rane site, very good reading and indeed it was a good choice. if i were to link anywhere, which i probably should have; it would have been there. there is also a very good thread on grounding in the articles section on this site
Yep read that a while ago.
I'm not that old but in my younger years I repaired microphone cables routinely. They were all connected shield to shell on one side only. The chassis side if I remember correctly. That was thirty years ago. I'm in my forties. So that shows how far behind!
I appreciate the link and the pointers. I'm at the bleeding edge in loudspeaker design and whacky cabinet construction. But electronics I have to blow off the dust once again. Can't be all things at once.
I enjoy this thread and all the comments. Thanks to Owen and all the crew for a great project. I'm about to solder mine up and enshrine the little monster in a nice case. I'll post some pics. I think I will go with rechargeable battery power. Should give a quiet supply. Do the gurus think I still need a regulator? Although if I remember the regulator type 300 series should not dissipate to much power at the current draw. Only a few milliamps.
I'm not that old but in my younger years I repaired microphone cables routinely. They were all connected shield to shell on one side only. The chassis side if I remember correctly. That was thirty years ago. I'm in my forties. So that shows how far behind!
I appreciate the link and the pointers. I'm at the bleeding edge in loudspeaker design and whacky cabinet construction. But electronics I have to blow off the dust once again. Can't be all things at once.
I enjoy this thread and all the comments. Thanks to Owen and all the crew for a great project. I'm about to solder mine up and enshrine the little monster in a nice case. I'll post some pics. I think I will go with rechargeable battery power. Should give a quiet supply. Do the gurus think I still need a regulator? Although if I remember the regulator type 300 series should not dissipate to much power at the current draw. Only a few milliamps.
yeah the recommendation used to be to connect at the source side only and we are still fighting to get people to forget this . sounds like i could use a mind meld with you, being that i'm about to embark on my first set of diy loudspeakers. being a headphone guy primarily till now, i'm pretty sorted for great quality source and transport, but i've left my studio monitors in the dust, so have some scanspeak drivers on the way and i'm way in the deep end with the cabinets and dont have and dont plan to get the tools to route the baffles, so will be trying to find a fellow member locally to make them from my design. wouldnt have a clue how to make a cabinet thats up to the quality of the drivers or the rest of my system. and kits arent really my thing.
going digital crossover, ackodac with Owens iv, aleph jx for bass and f5x for highs. so just need to come up with a nice solid sealed enclosure, if i want more bass i'll add subs.
now your request, depends on the battery type? how are you getting rid of the dc at the output from the split battery divider? a cap? seems a shame to do that after all the good work owen has done using no caps in the signal path. and certainly if you feel you need a regulator, its a long way around if you ask me. this amp has fantastic psrr and there are lots of high quality low noise regulator circuits, batteries in this case will create more problems than they solve imo unless you need it to be portable. i love them for the right situation, particularly lifepo4, but preferably you should use a very low Z out battery with no regulator and at the very most a cap unless you need to drop the voltage. i would say just use a high quality ic reg, like lt1764a with lt1175. or a sigma22 if you want discrete. salas regs would work nicely to, but will be large and create a lot of heat.
guru?? i think not, but hopefully thats of some use
. sounds like i could use a mind meld with you, being that i'm about to embark on my first set of diy loudspeakers. being a headphone guy primarily till now, i'm pretty sorted for great quality source and transport, but i've left my studio monitors in the dust, so have some scanspeak drivers on the way and i'm way in the deep end with the cabinets and dont have and dont plan to get the tools to route the baffles, so will be trying to find a fellow member locally to make them from my design. wouldnt have a clue how to make a cabinet thats up to the quality of the drivers or the rest of my system. and kits arent really my thing. going digital crossover, ackodac with Owens iv, aleph jx for bass and f5x for highs. so just need to come up with a nice solid sealed enclosure, if i want more bass i'll add subs.
now your request, depends on the battery type? how are you getting rid of the dc at the output from the split battery divider? a cap? seems a shame to do that after all the good work owen has done using no caps in the signal path. and certainly if you feel you need a regulator, its a long way around if you ask me. this amp has fantastic psrr and there are lots of high quality low noise regulator circuits, batteries in this case will create more problems than they solve imo unless you need it to be portable. i love them for the right situation, particularly lifepo4, but preferably you should use a very low Z out battery with no regulator and at the very most a cap unless you need to drop the voltage. i would say just use a high quality ic reg, like lt1764a with lt1175. or a sigma22 if you want discrete. salas regs would work nicely to, but will be large and create a lot of heat.
guru?? i think not, but hopefully thats of some use
If you want loudspeaker help you can contact me through private e-mail no problem.
No shunt regulators may apply in a portable situation.
No output cap if I can help it. I don't want a cap to be the dominant qualitative factor.
Battery is for portability.
Most likely three possibilities. Dual rail 9 volts, as in two per rail. Easy but not low impedance source.
Or SLA 12 volt. The SLA option would not be to portable but play a long, long time.
One other option would be a couple of laptop packs and a laptop charger. But that comes in about 19 volts. To high for the OpAmp if I remember correctly.
No shunt regulators may apply in a portable situation.
No output cap if I can help it. I don't want a cap to be the dominant qualitative factor.
Battery is for portability.
Most likely three possibilities. Dual rail 9 volts, as in two per rail. Easy but not low impedance source.
Or SLA 12 volt. The SLA option would not be to portable but play a long, long time.
One other option would be a couple of laptop packs and a laptop charger. But that comes in about 19 volts. To high for the OpAmp if I remember correctly.
hmm, ok so you will need to make sure the input ground matches the midrail voltage 'ground' or there will be offset at the output.
i would not use either of those types of battery for this type of application. despite this just being a few opamps (well more than a few and the buffers are pretty hungry), the current draw is actually reasonably large for a headphone amp. sla is too slow for creating a reasonable virtual ground with just a battery divider, also noisy and heavy and your laptop batteries will require protection circuits so they dont explode in your pocket.
lifepo4 is best option without question, it has good energy density, massive output current, low output impedance, better than most high grade caps and is much more tolerant of over/undercharge. i recommend looking at the linear tech range of ldos, some are nicely high current, low noise and are well suited for battery power, very well suited, in fact most of the range was designed at the very least with it in mind. they also have a convenient shutdown pin that can be used for basically putting the unit into a very low power sleep state with a single relay or quick and dirty switch. your psu will need to source/sink 800ma or there abouts, each lifepo4 fro a123 can output 40A in bursts.
anyway we are probably way OT, i've done a lot of research into this area with the design of my portable buffalo 2 with balanced headphone amp (basically a modified IV stage) so have looked into shutdown circuits, dc protect etc etc. maybe we can help each other.
i would not use either of those types of battery for this type of application. despite this just being a few opamps (well more than a few and the buffers are pretty hungry), the current draw is actually reasonably large for a headphone amp. sla is too slow for creating a reasonable virtual ground with just a battery divider, also noisy and heavy and your laptop batteries will require protection circuits so they dont explode in your pocket.
lifepo4 is best option without question, it has good energy density, massive output current, low output impedance, better than most high grade caps and is much more tolerant of over/undercharge. i recommend looking at the linear tech range of ldos, some are nicely high current, low noise and are well suited for battery power, very well suited, in fact most of the range was designed at the very least with it in mind. they also have a convenient shutdown pin that can be used for basically putting the unit into a very low power sleep state with a single relay or quick and dirty switch. your psu will need to source/sink 800ma or there abouts, each lifepo4 fro a123 can output 40A in bursts.
anyway we are probably way OT, i've done a lot of research into this area with the design of my portable buffalo 2 with balanced headphone amp (basically a modified IV stage) so have looked into shutdown circuits, dc protect etc etc. maybe we can help each other.
Last edited:
You are referring to lithium polymer batteries?
Because that is what most of the newer laptop packs are comprised of. They even come with wonderful chargers that are purpos built with all the safety bells and whistles.
Hence my attraction to a cheap prebuilt version of a battery pack.
Second point.
Two battery packs. Positive and negative for a split rail power supply.
I agree on the potential regulators. The LM300 series is not a low drop out device. But many of the Linear Tech ones have more ripple. But as pointed out already the ripple rejection is very good on the Op-amps.
Has anybody done a current consumption calc or am I just to lazy to find it in the thread?
Cuz that would be a make or break for the battery version of this amp.
Because that is what most of the newer laptop packs are comprised of. They even come with wonderful chargers that are purpos built with all the safety bells and whistles.
Hence my attraction to a cheap prebuilt version of a battery pack.
Second point.
Two battery packs. Positive and negative for a split rail power supply.
I agree on the potential regulators. The LM300 series is not a low drop out device. But many of the Linear Tech ones have more ripple. But as pointed out already the ripple rejection is very good on the Op-amps.
Has anybody done a current consumption calc or am I just to lazy to find it in the thread?
Cuz that would be a make or break for the battery version of this amp.
LME49600 Headphone Amp Design
I've had interest in the LME49600 and plan to build a similar, but lower cost and more DIY-friendly, design with unbalanced RCA inputs. I would also include some bandwidth limiting as others have suggested. The OPC design here is essentially a balanced version of the one on the National datasheet and in their application note AN-1768:
National LME49600 AN-1768 Application Note
The design I have in mind would be virtually identical to the one in the app note above with perhaps a few useful additions. It would be significantly less expensive than the OPC version as instead of 6 expensive op amps you only need 2 dual op amps (there's no issue sharing the servo with the gain stage for each channel in a single package as National did in their reference design). And the dual parts are barely more expensive than the single versions OPC used. So that eliminates almost $20 of op amps.
I would also use through hole parts, instead of SMT as in OPC's design, for everything but the LME49600 (and it's fairly big and relatively easy to solder). The dual op amps used in the National design are available in an 8 pin DIP and could even be socketed for those who like to experiment with different op amps.
I have a Prism dScope III which is a close competitor to the APx5xx analyzer OPC used. I also have an ABX set up that allows full blind listening tests. And I use the digital AudioDiffmaker with my Benchmark ADC1 as well as analog (i.e. "Hafler Distortion Test") audio differencing to evaluate amplifiers under real world conditions. So expect lots of tests.
Part of my interest in the LME49600 is to have an "ultra low distortion" entry in an extensive comparison between genres of headphone amps. I plan to conduct full dScope measurements, blind ABX listening, and audio differencing on the National LME49600 reference design, a popular lower feedback discrete Class-A design (i.e. Gilmore/Kumisa/Balkishan), and a lowly single IC "cmoy" design using a decent op amp like the OPA551.
If the National reference LME49600 amp is of interest for others wanting PC boards, please let me know? If there's enough other interest it will change how much effort I put it into the PCB and documentation as well as my purchase quantity for the PCBs.
If there's a lot of interest I could also consider a group buy of parts and/or even having the boards fully made at the assembly contractor I use. If the volumes are really high, then a full SMT version that can be done on their mask/pick/place/reflow lines and be even cheaper.
I suspect most interested in the LME49600 already participated in OPC's buy? But I wanted to at least see if there was some added interest for either those who missed out or those who might be interested in a lower cost unbalanced input through-hole version that was more DIY-friendly?
If this gets much interest, I'll start a new thread as it's really a different project. But this seemed like a good place to start.
I've had interest in the LME49600 and plan to build a similar, but lower cost and more DIY-friendly, design with unbalanced RCA inputs. I would also include some bandwidth limiting as others have suggested. The OPC design here is essentially a balanced version of the one on the National datasheet and in their application note AN-1768:
National LME49600 AN-1768 Application Note
The design I have in mind would be virtually identical to the one in the app note above with perhaps a few useful additions. It would be significantly less expensive than the OPC version as instead of 6 expensive op amps you only need 2 dual op amps (there's no issue sharing the servo with the gain stage for each channel in a single package as National did in their reference design). And the dual parts are barely more expensive than the single versions OPC used. So that eliminates almost $20 of op amps.
I would also use through hole parts, instead of SMT as in OPC's design, for everything but the LME49600 (and it's fairly big and relatively easy to solder). The dual op amps used in the National design are available in an 8 pin DIP and could even be socketed for those who like to experiment with different op amps.
I have a Prism dScope III which is a close competitor to the APx5xx analyzer OPC used. I also have an ABX set up that allows full blind listening tests. And I use the digital AudioDiffmaker with my Benchmark ADC1 as well as analog (i.e. "Hafler Distortion Test") audio differencing to evaluate amplifiers under real world conditions. So expect lots of tests.
Part of my interest in the LME49600 is to have an "ultra low distortion" entry in an extensive comparison between genres of headphone amps. I plan to conduct full dScope measurements, blind ABX listening, and audio differencing on the National LME49600 reference design, a popular lower feedback discrete Class-A design (i.e. Gilmore/Kumisa/Balkishan), and a lowly single IC "cmoy" design using a decent op amp like the OPA551.
If the National reference LME49600 amp is of interest for others wanting PC boards, please let me know? If there's enough other interest it will change how much effort I put it into the PCB and documentation as well as my purchase quantity for the PCBs.
If there's a lot of interest I could also consider a group buy of parts and/or even having the boards fully made at the assembly contractor I use. If the volumes are really high, then a full SMT version that can be done on their mask/pick/place/reflow lines and be even cheaper.
I suspect most interested in the LME49600 already participated in OPC's buy? But I wanted to at least see if there was some added interest for either those who missed out or those who might be interested in a lower cost unbalanced input through-hole version that was more DIY-friendly?
If this gets much interest, I'll start a new thread as it's really a different project. But this seemed like a good place to start.