Roll Call: Who's building, built, or thinking of building a beta22?
Dec 31, 2013 at 2:38 PM Post #3,138 of 3,218
Fishski13:
 
I do have 2.5" heatsinks with a 100VA trafo. I am currently running gain of x5, but I do have spare components to increase the gain to x8 if I needed to. So if I went on the path of high quiescent current with gain of 8, are you saying I should be able to drive the HE-6 properly? Or do you think it is still safer me to go with LCD3? The beta 22 is my only good amp and I am also concerned about it being too powerful (too little pot play) with gain of 8 for my other future headphones =P
 
Jan 1, 2014 at 2:23 AM Post #3,140 of 3,218
  Fishski13:
 
I do have 2.5" heatsinks with a 100VA trafo. I am currently running gain of x5, but I do have spare components to increase the gain to x8 if I needed to. So if I went on the path of high quiescent current with gain of 8, are you saying I should be able to drive the HE-6 properly? Or do you think it is still safer me to go with LCD3? The beta 22 is my only good amp and I am also concerned about it being too powerful (too little pot play) with gain of 8 for my other future headphones =P

 
kind of opening a can of worms...and i've never heard the HE6.
 
the power delivery of the amplifier should be capable of driving the HPs specs with "adequate" headroom and allowing for quieter mastered recordings.  assuming no significant hearing loss, 110dB is plenty.
 
assuming that the HE6 is 78dB/mW and 43R impedance:
 
105dB = 501mW, 4.6Vrms, 108mA
110dB = 1.6W, 8.3Vrms, 193mA
115dB = 5W, 14.7rms, 340mA
120dB =  15.9W, 26.2Vrms, 608mA
 
the O22 has no current limiting.  effective continuous current delivery is going to be determined by the heatsinking, TX VA, and fuse rating.  the stock 1.5" heatsinks are good for 1A continuous and peak much larger, but we're no where near those limits driving the HE6.
 
assuming a 2Vrms unbalanced source, to achieve at least the 14.7Vrms needed to deliver 115dB, you need a gain of 8x.
 
in a push-pull Class A design, total Class A current delivery will be 2x the quiescent current.  so a quiescent current bias of 170mA = 340mA, or the current delivery required to stay in Class A at 115dB.  if the load draws more than 340mA, then you're in Class A/B.
 
if my Benchmark DAC1 as a HP amp or my QRV-08 sounds thin driving my K701 or Fostex T50RP, it's not because its underpowered - it's just not the best pairing.
 
going back to my B1/F5 anecdote, if any HP is tuned or dampened with an under-driven amplifier, this may affect the sonics if the HP is being driven by a "normal" amplifier.  i really have no idea and am just throwing out ideas as to why people like under-driven speaker amplifiers powering the HE6.
 
i would think 100VA would be enough for a 3 channel B22 and 200mA quiescent current bias.  you could even take the bias up higher with a 2 channel.  the only reason that i see to take the bias up would be for sonics.  changing the quiescent current bias will affect the sonics somewhat.  going with 2.5" heatsinks is never a bad idea with a B22 - more so on the B22 boards themselves since they're dropping more voltage and dissipating more power across a single MOSFET/heatsink than across the pass MOSFETs in the O22.
 
a gain of 8x would be fine even with the LCD3.  they're not terribly efficient and you should have no noise issues.  i didn't have any noise issues when driving my K701 and 8x gain.  i dropped the gain to 5x though.  you should be able to get away with a 24 stepped attenuator with decent volume control as well.  once you start driving more efficient HPs, you may want to consider a pot instead or dropping the gain.  changing the gain isn't a huge deal.
 
Jan 2, 2014 at 1:22 AM Post #3,141 of 3,218
I thought the HE-6 has sensitivity of 83.5dB/mW with impedance of 50Ohm? I do not know how to do the calculations myself, but will the numbers stay roughly the same? say for 115dB level, will the current draw be still similar to 340mA? Also, on the AMB website, it says the power output for B22 is 18Wrms into 8Ω unbalanced. Is that the maximum power output the amp is capable of in single ended mode regardless of the transformer rating? If so, what limits the output power?
 
Jan 2, 2014 at 12:56 PM Post #3,142 of 3,218
I thought the HE-6 has sensitivity of 83.5dB/mW with impedance of 50Ohm? I do not know how to do the calculations myself, but will the numbers stay roughly the same? say for 115dB level, will the current draw be still similar to 340mA? Also, on the AMB website, it says the power output for B22 is 18Wrms into 8Ω unbalanced. Is that the maximum power output the amp is capable of in single ended mode regardless of the transformer rating? If so, what limits the output power?

 


oops. my numbers were from memory and incorrect. here's what Tyll measured: http://www.innerfidelity.com/headphone-data-sheet-downloads . 77dB/mW and 53R impedance. which means that on paper, the B22 should be able to drive the HE6 to 113dB with 14.5Vrms and 275mA. if going by the HiFiMan specs, then would would be looking at 119dB with about the same power demands.

P=V*I. the max rail voltages the B22 is spec'd for is +/-30V. you cannot get anymore than the 15.2Vrms uloaded spec. the rated VA for the O22 PS is appropriate for the spec'd B22 power. you can increase the VA for more Class A bias to get more power in Class A, but i honestly have no idea how much more current you could realistically get in Class A/B.

i can PM you later on how to perform the calcs.
 
Jan 4, 2014 at 6:09 AM Post #3,143 of 3,218
Hello!
 
Have anyone use custom PSU for beta22. Something like this http://www.ebay.com/itm/ULN-PS7-Ultra-Low-Noise-6-0A-High-Current-Power-Supply-for-High-Quality-Audio/231109887461?rt=nc&_trksid=p2047675.m1851&_trkparms=aid%3D222002%26algo%3DSIC.FIT%26ao%3D1%26asc%3D261%26meid%3D3875593153771070086%26pid%3D100005%26prg%3D1088%26rk%3D2%26rkt%3D5%26sd%3D330869867631%26. This PSU with ultralow noise. Original PSU sigma 22 has 102.5 db noise against 132 on ebay's one.
 
Jan 4, 2014 at 1:30 PM Post #3,144 of 3,218
  Hello!
 
Have anyone use custom PSU for beta22. Something like this http://www.ebay.com/itm/ULN-PS7-Ultra-Low-Noise-6-0A-High-Current-Power-Supply-for-High-Quality-Audio/231109887461?rt=nc&_trksid=p2047675.m1851&_trkparms=aid%3D222002%26algo%3DSIC.FIT%26ao%3D1%26asc%3D261%26meid%3D3875593153771070086%26pid%3D100005%26prg%3D1088%26rk%3D2%26rkt%3D5%26sd%3D330869867631%26. This PSU with ultralow noise. Original PSU sigma 22 has 102.5 db noise against 132 on ebay's one.

 
i'm not sure if the open-loop gain here defines the PSRR.  i would pose the question at amb.org.  
 
Jan 5, 2014 at 12:51 PM Post #3,145 of 3,218
the B22 has several design feature that give higher PSRR - separately filtered C-multiplier/Gyrator front end supply, cascodes between all rail referenced and ground referenced stages, grounded shunt R, complementary "balanced" compensation
 
these should add up to awesome PSRR for the B22 amplifier
 
a "super regulator" supply for the B22 is more "gilding the lily" than puting a Pd reflective plating layer inside a heavy Au casting (bright Ni or other shiny layer is put under cheap, thin
Au "plating" because the Au so thin it is actually partially transparent )
 
 
there is a relation between gains, PSRR in monolithic op amps which don't have a ground pin - all gains are referenced to one or the other of the supply pins - but the B22 doesn't have that limitation
 
Apr 11, 2014 at 4:57 PM Post #3,146 of 3,218
Well looks like something is in the works....
 

 
Apr 12, 2014 at 1:15 PM Post #3,148 of 3,218
Some progress with better lighting:
 

 
 
Didn't realize it takes so long to solder one amp board! I hope to get another one set up tonight to test out dual mono unbalanced.
 
Apr 12, 2014 at 3:59 PM Post #3,149 of 3,218
Is the Par-Metal case a 16" x 16"? Looks like you have plenty of room in there, unlike the 12 x 12 cases that are popular here.
 
Take it from an old DIY'er and don't get in a hurry. Disaster will ensue! Take it slow and enjoy the ride, and don't compromise anything 
to get it done quickly. 
 
Apr 12, 2014 at 6:09 PM Post #3,150 of 3,218
It's a 16x12x3. I didn't like the look of the terraced board arrangement in a 12x12, so I opted for a bigger case which also comes with extra ventilation holes on the top panel. I finished the second board but I stopped because I'm a little daunted by the input wiring. I really want to apply power to the boards to see if everything is okay but I think getting the input and outputs properly wired first will be better...
 

Users who are viewing this thread

Back
Top