my frends and i on my headphone Site are working on devoloping a High quality Headphone Amplifier for the DIY comunity using a Differential Output drive that uses another Amplifier to sink and source the load current for the normaly grounded common headphone conection. This and other design techniques I would like to get input from the people hear on some remaining design issues. http://elvencraft.com/ppa/
Interesting that you've chosen the HA5002 as output buffers. I have some experience using it, in both preamp and headamp apps... excellent chip.
- why have you chosen to use two per channel? I would have thought that one would give you plenty of current capacity.
- the 1k slew limiting resistors before the 5002 is presumably for stability margin? In my experience, if grounding is done with just a little care, it will not be needed.
- there will be just a little offset voltage from the 5002 (typ. 8-10 mV in my application), so I use it dc coupled. How much offset do you expect the 8610 to generate?
- imho, the 5002 sounds very, very good
Given the extreme attention you've given to supply and choice of chips, I feel that you may have overkill on aspects such as supply regulation and output buffering. $ for $, I would cut down on these, and look for a better solution on the volume pot.
Looks like a very interesting project. Good luck!
- why have you chosen to use two per channel? I would have thought that one would give you plenty of current capacity.
- the 1k slew limiting resistors before the 5002 is presumably for stability margin? In my experience, if grounding is done with just a little care, it will not be needed.
- there will be just a little offset voltage from the 5002 (typ. 8-10 mV in my application), so I use it dc coupled. How much offset do you expect the 8610 to generate?
- imho, the 5002 sounds very, very good
Given the extreme attention you've given to supply and choice of chips, I feel that you may have overkill on aspects such as supply regulation and output buffering. $ for $, I would cut down on these, and look for a better solution on the volume pot.
Looks like a very interesting project. Good luck!
Thank you for your insight. The prototype of this amp has a DC offset of less than 1 mV with the bass boost on and about 0.001 with it off using the AD-8610BR or the OPA-627B. yes most of the HA-5002's i have tested typicaly run about 6 mV. The reason i use two Buffers is for low impedance headphones like the 12 ohm Sony MDR-F1 and AKG K-1000. Other wise as you sed just one is plenty.
Is anything wrong with the Alps Blue pot?
Is anything wrong with the Alps Blue pot?
You're welcome, and I'm interested to track the progress. However I won't be joining any of the headphone groups, I think -- too many groups to read as it is!
No, I've no negative comments about the Alps, but I know that my HA5002 buffer quite easily resolves differences between supply caps and attenuators, and given the quality of the other components, there is room for investigation on pots.
If I recall, Thorsen and others have made comparisons and comments on log pots, linear + law fake, etc. One option to check is a linear pot with quality law fake.
FWIW, I like Panasonic FC, Rubycon ZA and ZL and Cerafines on the 5002. I find the Silmics to smooth , and Oscons way top lean with them. My Cerafines seemed to take a long, long time to burn in with the 5002.
No, I've no negative comments about the Alps, but I know that my HA5002 buffer quite easily resolves differences between supply caps and attenuators, and given the quality of the other components, there is room for investigation on pots.
If I recall, Thorsen and others have made comparisons and comments on log pots, linear + law fake, etc. One option to check is a linear pot with quality law fake.
FWIW, I like Panasonic FC, Rubycon ZA and ZL and Cerafines on the 5002. I find the Silmics to smooth , and Oscons way top lean with them. My Cerafines seemed to take a long, long time to burn in with the 5002.
I can relate to having lot's of groups to read, i also have that problem. Regarding Electrolytic capacitors. This Amp uses rail to rail bulk capacitence rather than the more conventional rail to ground. do you still think that this still matters?
Yes the HA-5002 has great resolving powers as the AD-8610 also has. these two Chips work well together. Other op amps do not sound so good in this Amp OPA-132, NE-5532 and similar sound real Harsh in this amp, The few op amps that do sound good in addition to the AD-8610 are the OPA-627 in the Ground Channel and OPA-637 in the Gain channels. the AD-825 also is good. although at this point i have not tried a vast range of Op amps
Yes the HA-5002 has great resolving powers as the AD-8610 also has. these two Chips work well together. Other op amps do not sound so good in this Amp OPA-132, NE-5532 and similar sound real Harsh in this amp, The few op amps that do sound good in addition to the AD-8610 are the OPA-627 in the Ground Channel and OPA-637 in the Gain channels. the AD-825 also is good. although at this point i have not tried a vast range of Op amps
I can agree with the mentioned overkill parts but I think the current source is an obvious overkill.mb said:
How much distortion do you get without this current source?
How much do you get with a plain resistor?
How much do you get with a non-cascode current source?
I, "the overkill guy" would have settled for a non-cascode current source.
Otherwise a nice jobb.
BTW: Is Morsel a "she"? I have never come acrossed a female electronic professional! Never!
BTW 2 :I think it's amazing that Halo hasn't been around.....16 posts a day, every day and then boom, silence.
peranders>
Glad to hear your input, I also was hoping for Halo also as he offers some different insights on these things. The cascode Current source is IMHO not overkill from the Perspective of sound quality. I did Listening tests some time ago on the differences Between these in this application http://www4.head-fi.org/forums/showthread.php?s=&threadid=3254&highlight=Cascode+current+source. the Performance of this Amp as it now is is so good that the Numbers would indicate perfection, However i fall into the Objectionest camp and place listening test foremost and if it also measures good great.
yes Morsel is quite a talented Electronic professional Female. While it is true more men than women get into Electronics those that do are real good at it.
Hi kevin>
Glad to hear your input i thought this would be of interest since i know you like balanced bridged output's.
Glad to hear your input, I also was hoping for Halo also as he offers some different insights on these things. The cascode Current source is IMHO not overkill from the Perspective of sound quality. I did Listening tests some time ago on the differences Between these in this application http://www4.head-fi.org/forums/showthread.php?s=&threadid=3254&highlight=Cascode+current+source. the Performance of this Amp as it now is is so good that the Numbers would indicate perfection, However i fall into the Objectionest camp and place listening test foremost and if it also measures good great.
yes Morsel is quite a talented Electronic professional Female. While it is true more men than women get into Electronics those that do are real good at it.
Hi kevin>
Glad to hear your input i thought this would be of interest since i know you like balanced bridged output's.
Op-Amp Change HA-5160
I used the Comp pin Pin 8 on this device and left the output Disconnected. So pin 8 of the op amp was inserted into the socket pin 6. I used IC socket pins for the cascaded Bias to the Op amp output so removing that resistor removed the connection. The intitial DC offset with the BB off was 1.2 mV with the BB on this was 10 times that at 13mV still not Bad at all. for some reason this went down to 6.5mV from 13mV by operating With a 13.2 Volt supply while a 24 volt supply produced 13mV. Dose this make sense?
Initial listening test started with my Panasonic SL-SW860 player and I got a clean and detailed sound and on par in the Mids and Highs as the OPA-637/627 set up. The Midrange seamed slightly recessed compared to the BB chips. One thing that Clearly stands out with the configuration is clarity and blackness of background. My initial impression of gladiator DVD was quite startling and Noted the intense detail that was being reviled. it was if I almost could hear DA converter Problems which should not have been the case with a mash 1 Bit used in the technics DVD-A40. On mp3's the better ones were listen able but the decoding artifacts were quit abit more apparent with this configuration than I remember them with all the Other op-amps operating through there output stage and current sourced. the Exception was the AD-744 that also has a comp pin. it sounds allot better in the PPA through its comp pin than its own output pin. However this device still has grain in the uppermost octaves and thus even in the Comp pin mode is not suitable for the PPA with the Highly revealing 5002's. This HA-5160 is quite another matter no grain what so ever anywhere. This is also like the OPA-637 a fast sounding Op amp. real nice quick transients most noticeable in the Mids and highs. The bass has as much extension as other op-amps but seamed to be in the BB camp rather than the AD on bass quality. The Bass was non the less not deemed bad or inaccurate just as one would not think the BB sound is any less accurate than the AD sound. We Introduce the Intersil sound with the HA-5160. ill report more when I further evaluate this chip but initial impressions are good. up coming is the analog test and re comparing to the OPA-637 set up as sonically that is closer to this chip than the AD-8610 or the AD-843. this chip as my last checking is available from newark
I used the Comp pin Pin 8 on this device and left the output Disconnected. So pin 8 of the op amp was inserted into the socket pin 6. I used IC socket pins for the cascaded Bias to the Op amp output so removing that resistor removed the connection. The intitial DC offset with the BB off was 1.2 mV with the BB on this was 10 times that at 13mV still not Bad at all. for some reason this went down to 6.5mV from 13mV by operating With a 13.2 Volt supply while a 24 volt supply produced 13mV. Dose this make sense?
Initial listening test started with my Panasonic SL-SW860 player and I got a clean and detailed sound and on par in the Mids and Highs as the OPA-637/627 set up. The Midrange seamed slightly recessed compared to the BB chips. One thing that Clearly stands out with the configuration is clarity and blackness of background. My initial impression of gladiator DVD was quite startling and Noted the intense detail that was being reviled. it was if I almost could hear DA converter Problems which should not have been the case with a mash 1 Bit used in the technics DVD-A40. On mp3's the better ones were listen able but the decoding artifacts were quit abit more apparent with this configuration than I remember them with all the Other op-amps operating through there output stage and current sourced. the Exception was the AD-744 that also has a comp pin. it sounds allot better in the PPA through its comp pin than its own output pin. However this device still has grain in the uppermost octaves and thus even in the Comp pin mode is not suitable for the PPA with the Highly revealing 5002's. This HA-5160 is quite another matter no grain what so ever anywhere. This is also like the OPA-637 a fast sounding Op amp. real nice quick transients most noticeable in the Mids and highs. The bass has as much extension as other op-amps but seamed to be in the BB camp rather than the AD on bass quality. The Bass was non the less not deemed bad or inaccurate just as one would not think the BB sound is any less accurate than the AD sound. We Introduce the Intersil sound with the HA-5160. ill report more when I further evaluate this chip but initial impressions are good. up coming is the analog test and re comparing to the OPA-637 set up as sonically that is closer to this chip than the AD-8610 or the AD-843. this chip as my last checking is available from newark
This is the Compleated Amp so far almost done. I have decided upon a set of (4) Buffers as the THD went down with each Buffer addition with 2 volts RMS into 32 ohm Resistive loads.
The Results of using 1/2/3/4 Buffers per Ch
@ 1KHz> (1) Buffer (2) Buffers (3) Buffers (4) Buffers
THD, %: 0.051 / 0.025 / 0.014 / 0.0084
IMD, %: 0.045 / 0.024 / 0.020 / 0.019
The Op-Amps used in the Above are the AD-8610AR
The Sound is fantastic with the halmark of this Amp Being soundstage. since i got such great imaging with Headphones I thought some low volume listening with Gasp! a pair of small loudspeakers (Radio Shack Pro77 Metal box type) were in order and what a great treat that was with some parts coming from below the Floor. Most Loudspeaker Amps will not do that with alot more costly loudspeakers. Overall I am happy with the Measured and sonic performance of this Amp. BTW this version uses OPA-637 in gain channels and OPA-627 in the Ground Driver Channel.
Inside Photo
The Results of using 1/2/3/4 Buffers per Ch
@ 1KHz> (1) Buffer (2) Buffers (3) Buffers (4) Buffers
THD, %: 0.051 / 0.025 / 0.014 / 0.0084
IMD, %: 0.045 / 0.024 / 0.020 / 0.019
The Op-Amps used in the Above are the AD-8610AR
The Sound is fantastic with the halmark of this Amp Being soundstage. since i got such great imaging with Headphones I thought some low volume listening with Gasp! a pair of small loudspeakers (Radio Shack Pro77 Metal box type) were in order and what a great treat that was with some parts coming from below the Floor. Most Loudspeaker Amps will not do that with alot more costly loudspeakers. Overall I am happy with the Measured and sonic performance of this Amp. BTW this version uses OPA-637 in gain channels and OPA-627 in the Ground Driver Channel.
Inside Photo
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.