I'm looking for a similar design like the Simplistic Phono RIAA but for Headphone Amp, for mine needs not be necessary be portable, I own a Senhheisser HD600 headphone with high impedance so the Sennhs will be comfortably with high gain to drive well & get it's maximum perfomance.
All suggestions will welcome, thanks for your attention.
All suggestions will welcome, thanks for your attention.
There are several here
HeadWize Library - Projects (check the forum here as well).
DIY (Do-It-Yourself) Discussions - Head-Fi: Covering Headphones, Earphones and Portable Audio
HeadWize Library - Projects (check the forum here as well).
DIY (Do-It-Yourself) Discussions - Head-Fi: Covering Headphones, Earphones and Portable Audio
This should be a challenge to Salas. The world needs a simplistic HP amp
Salas, save us!
How Loud Do you Want to Go?
HD600 spec sheet shows 97 dB SPL at 1 mW, corresponding to 0.55 V rms across the 300 Ohm nominal impedance. How loud do you want to go? 120 dB SPL is both the threshold of pain, and like standing near a jet plane on take-off. To do this, you need to deliver 7.7 Volts RMS. That starts to limit some options...so your choice of maximum output (e.g. SPL) could be important...
HD600 spec sheet shows 97 dB SPL at 1 mW, corresponding to 0.55 V rms across the 300 Ohm nominal impedance. How loud do you want to go? 120 dB SPL is both the threshold of pain, and like standing near a jet plane on take-off. To do this, you need to deliver 7.7 Volts RMS. That starts to limit some options...so your choice of maximum output (e.g. SPL) could be important...
What about 20dB, rise time 200 nanoseconds, class A, closed loop frecuency 1 MHz, output impedance less than 1 ohm, etc.
Link Headphone Erno Borbely: http://www.borbelyaudio.com/pics/Web309line.pdf
Well in DIY real world don't need all mine wishes, but I know that here in the forum are people can do very similar for a fraction of price & of course the enjoyment to DIY
Link Headphone Erno Borbely: http://www.borbelyaudio.com/pics/Web309line.pdf
Well in DIY real world don't need all mine wishes, but I know that here in the forum are people can do very similar for a fraction of price & of course the enjoyment to DIY
This should be a challenge to Salas. The world needs a simplistic HP amp
Salas, save us!
I want saved by Salas too
I would hope more DIY audio people are aware of dynamic music peak power requirements
its not relevant that "120 dB is too loud" - for continuous listening
but its not too much for the mS peaks of drum hits, cymbal crashes - in fact clipping is often "painful" to hear if you have real live music dynamics
yes, if you listen to "loudness war" dynamic compressed music you may never need even 100 dB for all day long safe average SPL levels
but if you want to recreate "live music" you need lots of headroom for dynamic peaks
I believe that HD600 can be driven adequately from > +/-12 Vsuppy op amp circuits if you don't give up too much output swing with the bias headroom requirements of the output buffer - as some MOSFET buffer circuit do unfortunately
the current requirement of the 300 Ohm load can be handled directly by some higher output current op amps - skipping the buffer complexity and V loss
its not relevant that "120 dB is too loud" - for continuous listening
but its not too much for the mS peaks of drum hits, cymbal crashes - in fact clipping is often "painful" to hear if you have real live music dynamics
yes, if you listen to "loudness war" dynamic compressed music you may never need even 100 dB for all day long safe average SPL levels
but if you want to recreate "live music" you need lots of headroom for dynamic peaks
I believe that HD600 can be driven adequately from > +/-12 Vsuppy op amp circuits if you don't give up too much output swing with the bias headroom requirements of the output buffer - as some MOSFET buffer circuit do unfortunately
the current requirement of the 300 Ohm load can be handled directly by some higher output current op amps - skipping the buffer complexity and V loss
Last edited:
Thanks to post the schematic, Salas told me that is a very good design so you have all mine trust, initially I see a lot of components so I can't not do it, I have several newbie questions before do mine attempt:
-VF2 is INPUT?
-VF1 is OUTUT?
-Transistors must be matcheds?
-C9 & C11 = 1m, is 1 picofarad?
-Leds voltage value, CQX35A is a reference from a component supplier?
-Measurement for setup? IS1 4m is 4mA across T7 BF246A?
-Measurement for setup? IS2 4m is 4mA across R19 1K?
-Measurement for setup? IS3 50m is 50mA across R3 4.7R?
-Opamp LME49710 is active or is a servo DC offset?
-The design is DC coupled?
-I don't see any cap on the signal way
-Wich is the gain?
-Needs any mod for 300R load like mine Sennhs?
Also let mek now if this BOM for one channel is correct:
Trimmers
P1 - 100R linear
P3 - 10K are logarithmic & serves as volume or is also linear?
Resistors
R1+R2+R16 - 47R
R3+R4 - 4.7R
R5 - 10R
R6+R7 - 220R
R8+R9+R18 - 470R
R10 - 150R
R11 - 100R
R12 - 3.4K
R13 - 680R
R14 - 340R
R15 - 220K
R17 - 340R
R19 - 1K
Leds
1+2 V???
Capacitors
C1+C2+C4+C6+C10+C14 - 330uF
C3+C5 - 1 uF
C7+C12 - 4.7uF
C8+C13 - 10nF
C9+C11 - 1m, is 1pF???
Transistors
T1+T5+T11 - BC560C or better 517
T2+T6+T12 - BC550C or better 516
T3 - BD140
T4 - BD139
T7+T9 - BF246A or 2N3369
T8+T13 - 2SK170BL or 2N3369
T10 - OPAMP LME49710 or 2N2608
T14 - 2SJ074BL
Casually have you a blank PCB?
Thanks again to share the design.
so you have all mine trust, initially I see a lot of components so I can't not do it, I have several newbie questions before do mine attempt:-VF2 is INPUT?
-VF1 is OUTUT?
-Transistors must be matcheds?
-C9 & C11 = 1m, is 1 picofarad?
-Leds voltage value, CQX35A is a reference from a component supplier?
-Measurement for setup? IS1 4m is 4mA across T7 BF246A?
-Measurement for setup? IS2 4m is 4mA across R19 1K?
-Measurement for setup? IS3 50m is 50mA across R3 4.7R?
-Opamp LME49710 is active or is a servo DC offset?
-The design is DC coupled?
-I don't see any cap on the signal way
-Wich is the gain?
-Needs any mod for 300R load like mine Sennhs?
Also let mek now if this BOM for one channel is correct:
Trimmers
P1 - 100R linear
P3 - 10K are logarithmic & serves as volume or is also linear?
Resistors
R1+R2+R16 - 47R
R3+R4 - 4.7R
R5 - 10R
R6+R7 - 220R
R8+R9+R18 - 470R
R10 - 150R
R11 - 100R
R12 - 3.4K
R13 - 680R
R14 - 340R
R15 - 220K
R17 - 340R
R19 - 1K
Leds
1+2 V???
Capacitors
C1+C2+C4+C6+C10+C14 - 330uF
C3+C5 - 1 uF
C7+C12 - 4.7uF
C8+C13 - 10nF
C9+C11 - 1m, is 1pF???
Transistors
T1+T5+T11 - BC560C or better 517
T2+T6+T12 - BC550C or better 516
T3 - BD140
T4 - BD139
T7+T9 - BF246A or 2N3369
T8+T13 - 2SK170BL or 2N3369
T10 - OPAMP LME49710 or 2N2608
T14 - 2SJ074BL
Casually have you a blank PCB
?Thanks again to share the design.
VF2 is the input and VF1 is the output.
i did not match the transitors.
C9 / C11 is 1000uF that is 1mF.
The green led is not critical. I took a standart 10mA, 3mm diameter. You can use 5mm too. Forward voltage of green Leds is around 1.9 to 2.1 V. You can try other colours too.
usually red ones have slighly less forward voltage.
IS1 is the current through the Led. Again the absolute value is not critical. I use a Fet cascode here because a single Fet got too hot.
IS2 is the current that i force out of the OPamp to switch off one set of output transistors.
Theoretical the Opamp now works in class a up to 4mA. The sound with that tweak got a very little bit warmer but the circuit works as well without that tweak.
IS3 is the idle current in the output stage so it works into class a up to 50mA.
The output trasistors need some cooling. I used some stick on profiles. See the pickture. If you use bigger cooling profiles you can raise the class a operation up to 100mA by raising R1, R2 but i found that not necessary. The circuit as is can swing 1.5V into 30 Ohm class a and 5V into 100 Ohm class a. So in 100 Ohm it has 250mW class a. You can measure the idle current over R3, R4. So you will measure 0.235 V over R3 when you have 50mA idle current. Again exact value is not critical.
The LME49710 is the amplifying Opamp and gain is R12/R3 so 5 times. You can raise the 3.4kOhm to 6.8kOhm if you need more gain. The 680 Ohm is small because i wanted lowest noise as posible. That is aso the reason i use a buffer at the input. By the way you can adjust the DC at the output with P1. A servo is not neccesarry because the circuit has very stable working points. For your Sennheiser no modification is neccessary. The circuit simply will never leave class a with that high impedance.
Your Bom is correct except the 1mF that is 1000uF. Use the Fets i recommended because i did not try others. Maybe Salas can give you advice what kinds can be substituted when you can not buy the ones i use.
Better use the BC516 / 517. The circuit is then a bit faster and lower in distortion.
Even with the BC550C / 560C performance is very good and when you do not find the darlingtons any small signal bipolar pair will work,
P1 is a linear 20 gang trimmer that sets DC offset. P3 is a 10kOhm log potentiometer.
I originally used a Bourns 91 but later substited a TKD potmeter that sounded much better. You can use the blue Alps RK27 here. It has a good performance-price ratio and has surprising good channel tracking. The TKD is the best i tried but it is very expensive. Even when i buy it direct in Tokyo it costs at least 40,-€ each. You can use a law fake 100 Kohm ( 15kOhm in parallel) like on Elloits sound page. A pot i found that is not expensive and sounds good is from OMEG, an english company. You can get them from Farnell england and from Schuro in Germany.
In my prototype i used an LME49710 opamp because it has one of the lowest distortions
of modern OPamps. The NE5534A is nearly as good in this arangement because i use very linear shunt feedback. The only "disadvantage" of my circuit is , that it is phase inverting. This usually has no audible consequences because absolute phase is not preserved on most recordings. i can design a phase switch if you think it is important.
You can use other opamps then the LME49710 but i do not know the 2N2608. That sounds more like a transitor to me. You can use OPA134 or the better OPA827. There is a plethora of other options but as i said my circuit minimises audible differences between different makes.
I do not have a PCB yet. I simply did it dead bug style. See the photo.
i did not match the transitors.
C9 / C11 is 1000uF that is 1mF.
The green led is not critical. I took a standart 10mA, 3mm diameter. You can use 5mm too. Forward voltage of green Leds is around 1.9 to 2.1 V. You can try other colours too.
usually red ones have slighly less forward voltage.
IS1 is the current through the Led. Again the absolute value is not critical. I use a Fet cascode here because a single Fet got too hot.
IS2 is the current that i force out of the OPamp to switch off one set of output transistors.
Theoretical the Opamp now works in class a up to 4mA. The sound with that tweak got a very little bit warmer but the circuit works as well without that tweak.
IS3 is the idle current in the output stage so it works into class a up to 50mA.
The output trasistors need some cooling. I used some stick on profiles. See the pickture. If you use bigger cooling profiles you can raise the class a operation up to 100mA by raising R1, R2 but i found that not necessary. The circuit as is can swing 1.5V into 30 Ohm class a and 5V into 100 Ohm class a. So in 100 Ohm it has 250mW class a. You can measure the idle current over R3, R4. So you will measure 0.235 V over R3 when you have 50mA idle current. Again exact value is not critical.
The LME49710 is the amplifying Opamp and gain is R12/R3 so 5 times. You can raise the 3.4kOhm to 6.8kOhm if you need more gain. The 680 Ohm is small because i wanted lowest noise as posible. That is aso the reason i use a buffer at the input. By the way you can adjust the DC at the output with P1. A servo is not neccesarry because the circuit has very stable working points. For your Sennheiser no modification is neccessary. The circuit simply will never leave class a with that high impedance.
Your Bom is correct except the 1mF that is 1000uF. Use the Fets i recommended because i did not try others. Maybe Salas can give you advice what kinds can be substituted when you can not buy the ones i use.
Better use the BC516 / 517. The circuit is then a bit faster and lower in distortion.
Even with the BC550C / 560C performance is very good and when you do not find the darlingtons any small signal bipolar pair will work,
P1 is a linear 20 gang trimmer that sets DC offset. P3 is a 10kOhm log potentiometer.
I originally used a Bourns 91 but later substited a TKD potmeter that sounded much better. You can use the blue Alps RK27 here. It has a good performance-price ratio and has surprising good channel tracking. The TKD is the best i tried but it is very expensive. Even when i buy it direct in Tokyo it costs at least 40,-€ each. You can use a law fake 100 Kohm ( 15kOhm in parallel) like on Elloits sound page. A pot i found that is not expensive and sounds good is from OMEG, an english company. You can get them from Farnell england and from Schuro in Germany.
In my prototype i used an LME49710 opamp because it has one of the lowest distortions
of modern OPamps. The NE5534A is nearly as good in this arangement because i use very linear shunt feedback. The only "disadvantage" of my circuit is , that it is phase inverting. This usually has no audible consequences because absolute phase is not preserved on most recordings. i can design a phase switch if you think it is important.
You can use other opamps then the LME49710 but i do not know the 2N2608. That sounds more like a transitor to me. You can use OPA134 or the better OPA827. There is a plethora of other options but as i said my circuit minimises audible differences between different makes.
I do not have a PCB yet. I simply did it dead bug style. See the photo.
Thanks a lot for the information, now I have a clear idea about the project & hows operate.
-I have a couple NE5534AP, could I use instead the LME49710?
-Wich is the power recommended for the Tx of 12V, 30VA-50VA?
-What's between the reg. & the headamp, an inductor? wich value Ohms & Henry?
-Do you have a design for cross-feed, in mine opinion is very good to have it because you minimise the stereo effect in headphones & if it's well implemented for me the sound is better focused & more realistic.
-Will be super to have the option for work in phase or inverted, you are right I used the Borbely I/V inverted & seems that's only can hear the difference in several records in others not like you said. But if you don't have time now I can wait no problem.
Thanks again for your time.
-I have a couple NE5534AP, could I use instead the LME49710?
-Wich is the power recommended for the Tx of 12V, 30VA-50VA?
-What's between the reg. & the headamp, an inductor? wich value Ohms & Henry?
-Do you have a design for cross-feed, in mine opinion is very good to have it because you minimise the stereo effect in headphones & if it's well implemented for me the sound is better focused & more realistic.
-Will be super to have the option for work in phase or inverted, you are right I used the Borbely I/V inverted & seems that's only can hear the difference in several records in others not like you said. But if you don't have time now I can wait no problem.
Thanks again for your time.
You can measure IS2 over R19, the NE5534 is fine but will need a small cap ( ca 15pF) over Pin 1 / 8 when gain is under x5. A 30VA transformer is fine. There is some disagreement if a big or smaller Transformer sounds better. You an use Block Transformers from Reichelt. They are split bobbin, sound good and are inexpnsive.
The inductor is not nesserary. I use 27mH ferites that have a resistance of 4 Ohm. You can get inductors like that from Mundorf and Intertechnik. They are used for crossover networks in speakers. Hammond makes open frame coils with higher values. Ask Salas, he is a much better expert then me in PSU work.
I will design a pahse inverter for you. Comes later the day. I need some rest after work.
I am no expert on crossfeed eather. Lnkwitz did something and i have seen a lot of designs on HeadWise. I have not tried other then Linkwitz and have always switched it off after a while. The same problem like PSU issues. There are a lot of opinions which circuit is the best. I can do some research for you an try something out. It is interesting. Maybe i find something that is convincing.
The inductor is not nesserary. I use 27mH ferites that have a resistance of 4 Ohm. You can get inductors like that from Mundorf and Intertechnik. They are used for crossover networks in speakers. Hammond makes open frame coils with higher values. Ask Salas, he is a much better expert then me in PSU work.
I will design a pahse inverter for you. Comes later the day. I need some rest after work.
I am no expert on crossfeed eather. Lnkwitz did something and i have seen a lot of designs on HeadWise. I have not tried other then Linkwitz and have always switched it off after a while. The same problem like PSU issues. There are a lot of opinions which circuit is the best. I can do some research for you an try something out. It is interesting. Maybe i find something that is convincing.
I make both HeadWize designs "low Z" & "high Z", I prefer low Z sounds good in mine system but the only problem it's spend near 3dB & I am using now the Borbely balanced lineamp as headamp driving the HD600, the Erno lineamp it's designed to see an input impedance of 10K for power-amps not 300 ohms like Sennhs HD600 headphones also I was to change the gain in the lineamp due to mine setup Borbely power amps 75WRMS class A + Wilson Audio Specialties 94dBs 4 ohms, it's killer get easy highs SPL of 114dBs or more, so I reduced the gain but doing that in the lineamp now don't have punch to drive well mines Senhs HD600, that's the reason I'm looking for a headamp.
About cross-feed if it's well do it, is easy to hear: the stereo effect isn't too separated is more centered, not exactly in the middle but is more natural with headphones than hear near two voices/instruments when you have only one voice/instrument, of course all depends how is recorded but IMHO is a big step to improve the listening with headphones.
So if you don't have time, don't worry about cross-feed, I can add to your headamp.
Link HeadWize: http://gilmore2.chem.northwestern.edu/projects/showfile.php?file=cmoy1_prj.htm
Schematic:
About cross-feed if it's well do it, is easy to hear: the stereo effect isn't too separated is more centered, not exactly in the middle but is more natural with headphones than hear near two voices/instruments when you have only one voice/instrument, of course all depends how is recorded but IMHO is a big step to improve the listening with headphones.
So if you don't have time, don't worry about cross-feed, I can add to your headamp.
Link HeadWize: http://gilmore2.chem.northwestern.edu/projects/showfile.php?file=cmoy1_prj.htm
Schematic:
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I personally wouldn't say no to a DAO with Salas shunts..
[I know for a fact I'm not the only one..
If I do the JG headamp sure the regs will be Salas 1.0 with remote sensing
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.