It was a pain but I rewired my OB-1'a for balanced drive using two 3.5mm stereo jacks (taken by cutting a 3.5mm male-male cable in half) and then connecting that to the 4 wires from the OB-1 (luckily it is a 4 wire headphone and not 3 wire). I learned how to solder Litz wire for the first time.
So I have the single to balanced converters running of a stack of 4 rechargeable 9v EBL batteries for a Vcc of about 34v. Max output swing is 25% of that minus 1v or 7.5v p-p. More than enough to hit max SPL after 12dB gain on the power amps.
Source is the DAC from Focsrite Solo 2G playing high res FLACs. Power amps are NHB's running about 67mA bias with the works. Red amp is Right channel and Green amp is Left channel.
First impressions... WOW!!! It's a completely different ball game. The sheer dynamics, depth, crescendos, all kill a non balanced drive amp/phone combo. It appears to be much louder - probably 12dB? Volume knob is used sparingly.
First track played was Nils Lofgren's "Keith Don't Go" a song know well. The attack on the acoustic guitar is incredible.
Next up are some well tested tracks and all are amazing sounding. It sounds like a full blown desktop amp even though they are pocket amps.
Very very happy with the result. This is going to get made into a PCB for sure and probably is the one thing one can get to really change your listening sound quality to a whole 'nother level.
Wow that is cool!
So, for an average person like me, would I benefit more in terms of sound quality from a silicon harmony or from a balanced drive?
Average person would probably like the balanced drive more. It's kind of like a roller coaster ride. You are acutely aware of your visceral sensations and mind is alert and ready to brace itself. It's a special sound - not sure it is something one can hear 24/7. But very amazing when you hear it. If you haven't heard it before, nothing I say can prepare you for what it sounds like. We just don't know what we don't know.
Some more notes...
1. After a little more listening and tests I am finding that the interconnects between the phase splitter driver board and the power amps is critical. Use of 3.5mm stereo jacks is prone to noise as jiggling it can result in changes in background noise and a lot of popping if a disconnect occurs. Balanced drive is very very sensitive - essentially increases sensitivity of the headphones and makes them very resolving of low level detail. So I think this may not be practical to do with removable interconnects between balanced driver and power amp stages.
2. Because the output of the phase splitter is already capacitively coupled, there is no need to add more caps on the input of the power stage. Or vice versa, one can removed the caps on the phase splitter output stage.
3. If carefully matched FETs such as that used in the NHB are employed, and the bias of both +ve and -ve channels on each amp is carefully matched, then one can forgo output caps altogether as the voltage on each side will balance out for net zero current flow through the voice coil. Alas, about 7v DC will be sitting on each side of the voice coil but net DC offset of order 0mV to 5mV it can work. However, an uncomfortable proposition it will be should one side be inadvertently become unbalanced due to interconnect jostle or failure. But there will probably be increased resolution and dynamics possible in a DC coupled output configuration. It may be Ok to test and experiment with full knowledge one could destroy a set of cans with one false move. Hmm... I think I will leave the output coupler caps in place
I think this will be best implemented as a new fresh *integrated* PCB with all circuits on one board with small tight traces, full use of SMT components for the input phase splitter BJT's (vs TO92 through holes that I am now using). A delay turn on relay and possibly DC detect and protect output cut-off can also be added. This is all a lot of work though... but the resulting desktop amp can be very glorious sounding.
I may even switch to a CFP BJT topology for the output stage which will give even more resolution and better dynamics.
1. After a little more listening and tests I am finding that the interconnects between the phase splitter driver board and the power amps is critical. Use of 3.5mm stereo jacks is prone to noise as jiggling it can result in changes in background noise and a lot of popping if a disconnect occurs. Balanced drive is very very sensitive - essentially increases sensitivity of the headphones and makes them very resolving of low level detail. So I think this may not be practical to do with removable interconnects between balanced driver and power amp stages.
2. Because the output of the phase splitter is already capacitively coupled, there is no need to add more caps on the input of the power stage. Or vice versa, one can removed the caps on the phase splitter output stage.
3. If carefully matched FETs such as that used in the NHB are employed, and the bias of both +ve and -ve channels on each amp is carefully matched, then one can forgo output caps altogether as the voltage on each side will balance out for net zero current flow through the voice coil. Alas, about 7v DC will be sitting on each side of the voice coil but net DC offset of order 0mV to 5mV it can work. However, an uncomfortable proposition it will be should one side be inadvertently become unbalanced due to interconnect jostle or failure. But there will probably be increased resolution and dynamics possible in a DC coupled output configuration. It may be Ok to test and experiment with full knowledge one could destroy a set of cans with one false move. Hmm... I think I will leave the output coupler caps in place
I think this will be best implemented as a new fresh *integrated* PCB with all circuits on one board with small tight traces, full use of SMT components for the input phase splitter BJT's (vs TO92 through holes that I am now using). A delay turn on relay and possibly DC detect and protect output cut-off can also be added. This is all a lot of work though... but the resulting desktop amp can be very glorious sounding.
I may even switch to a CFP BJT topology for the output stage which will give even more resolution and better dynamics.
Attachments
Last edited:
Finally got my new work space up and running in the shop. Even built myself a vent hood entirely out of junk I had laying around!
My second board is making music! I upgraded pretty much every component this time around:
-All resistors are Vishay MELF
-Higher bias via R4 value of 51R
-Bit higher dummy load via RL value of 1K
-Power rail caps are 25V 1500uF Panasonic FR
-Power bypasses are original value, but TDK C0G
-Input caps are my DIY stacked and leaded 2x 63V .68uF WIMA SMD-PPS
-Still auditioning output cap options
The sound is awesome, though so far I have only tested it with my cheap AKG K-55 cans, just to be safe about possible DC leakage with all the caps pressure fit. Even so, dynamics and detail are just superb.
I spent some time last night auditioning my DIY PPS input caps vs. the stock polyesters, and quickly heard a difference (hopefully not my subjective bias). The polyesters have a creamy sound with a lot of bloom and soft treble. The PPS opens up the sound significantly with better precision, speed, and especially more air--exactly what I was hoping for! I soldered them in without hesitation. Anyone interested can see how I assembled these caps in post #664. They require some sanding of the inner side of the input jack to slide in, at least with the caps shrink-wrapped like I did.
I need to place an order for more output caps to experiment with, but right now I am loving the sound of this amp with the trusty 25V 470uF Nichicon KA standing alone on output duty.
This amp is measuring 7.5VDC by the way.
My second board is making music!
I upgraded pretty much every component this time around:-All resistors are Vishay MELF
-Higher bias via R4 value of 51R
-Bit higher dummy load via RL value of 1K
-Power rail caps are 25V 1500uF Panasonic FR
-Power bypasses are original value, but TDK C0G
-Input caps are my DIY stacked and leaded 2x 63V .68uF WIMA SMD-PPS
-Still auditioning output cap options
The sound is awesome, though so far I have only tested it with my cheap AKG K-55 cans, just to be safe about possible DC leakage with all the caps pressure fit. Even so, dynamics and detail are just superb.
I spent some time last night auditioning my DIY PPS input caps vs. the stock polyesters, and quickly heard a difference (hopefully not my subjective bias
). The polyesters have a creamy sound with a lot of bloom and soft treble. The PPS opens up the sound significantly with better precision, speed, and especially more air--exactly what I was hoping for! I soldered them in without hesitation. Anyone interested can see how I assembled these caps in post #664. They require some sanding of the inner side of the input jack to slide in, at least with the caps shrink-wrapped like I did.I need to place an order for more output caps to experiment with, but right now I am loving the sound of this amp with the trusty 25V 470uF Nichicon KA standing alone on output duty.
This amp is measuring 7.5VDC by the way.
Attachments
Nice work StellarE ! Those little round resistors look very hot rod. Try this for output caps: 1000uF oscon + 470uF Nichicon AK + 10uF 35v Silmic II. That's what I am runnning presently. Very nice performer in all areas. You could even squeeze in 100uF Silmic with lateral pins bent and stick on remaining open space on through hole side. Then go with 2.2uF Wima on SMT side. You would have total 1572uF of 4 varieties.
Love the duct tape hood!
Try this for output caps: 1000uF oscon + 470uF Nichicon AK + 10uF 35v Silmic II. That's what I am runnning presently. Very nice performer in all areas. You could even squeeze in 100uF Silmic with lateral pins bent and stick on remaining open space on through hole side. Then go with 2.2uF Wima on SMT side. You would have total 1572uF of 4 varieties. Love the duct tape hood!
Hi X,
I see there is a serious debate about balanced vs single ended connection for headamps. Now, I know nothing about the technicalities involved - but just a stupid question: with your current balanced setup using two PCA's - is that not similar to running two monoblocks? Or will two single ended PCA's not give a similar improvement?
I see there is a serious debate about balanced vs single ended connection for headamps. Now, I know nothing about the technicalities involved - but just a stupid question: with your current balanced setup using two PCA's - is that not similar to running two monoblocks? Or will two single ended PCA's not give a similar improvement?
I do not know why there is debate. If you have an amp that has limited voltage swing (like a portable amp with two 9v or a car battery with 12v of lead acid), using balanced drive will essentially double your output drive capability. Perhaps the debate is, if you have a Class AB amp and are not limited by your rail voltage, simply increase your rails to get the same swing.
Using two separate monoblock amps but without balanced drive will only buy you extra stereo channel separation (currently -72dB to -85dB in same box depending on whether source follower or CFP topology).
Here is a measurement of the output from the pocket amp in single ended drive vs balanced drive into 270ohm load. For balanced drive each amp already has a 270ohm internal load resistor plus an external 270ohm floating load resistor connected to just the output from each side's cap coupled output. So I think the resistance (or impedance) may be 270 parallel (270x2) or 180ohm equivalent.
The voltage swing is almost doubled - while keeping the amp in a lower distortion operating range vs pushing 10v with a single amp (which is above clipping). Balanced drive is also referred to in car audio amps as bridge tied load (BTL) amps in order to realize more voltage swing.
Single ended drive into 270ohms:
Balanced drive into 180ohms but same input:
So based on the above, don't you think balanced drive for a voltage limited SE Class A amp makes a difference?
Last edited:
The output voltage swing doubled so the gain increased and the max SPL before clipping increased substantially. All this translates to better dynamics by allows higher dynamic range to be played at higher volumes. It's is quite noticeable and apparent - night and day difference.
I am working up a new design with a MOSFET CCS and BJT CFP output stage with balanced drive capability. Bias will probably be in 125mA range and still SE Class A.
I am working up a new design with a MOSFET CCS and BJT CFP output stage with balanced drive capability. Bias will probably be in 125mA range and still SE Class A.
Last edited:
Nice setup! I felt a bit of pain seeing those headphones just willy nilly in a box though =/
Finally got my new work space up and running in the shop. Even built myself a vent hood entirely out of junk I had laying around!
My second board is making music!
-All resistors are Vishay MELF
-Higher bias via R4 value of 51R
-Bit higher dummy load via RL value of 1K
-Power rail caps are 25V 1500uF Panasonic FR
-Power bypasses are original value, but TDK C0G
-Input caps are my DIY stacked and leaded 2x 63V .68uF WIMA SMD-PPS
-Still auditioning output cap options
The sound is awesome, though so far I have only tested it with my cheap AKG K-55 cans, just to be safe about possible DC leakage with all the caps pressure fit. Even so, dynamics and detail are just superb.
I spent some time last night auditioning my DIY PPS input caps vs. the stock polyesters, and quickly heard a difference (hopefully not my subjective bias
I need to place an order for more output caps to experiment with, but right now I am loving the sound of this amp with the trusty 25V 470uF Nichicon KA standing alone on output duty.
This amp is measuring 7.5VDC by the way.
I definitely would like to hear that! Good desktop choice, no doubt. I'm about to order some of the different Nichicons that will fit in the tin, from the FG, FW and KW audio grades. I am also going to try X's recommended combo, as well as the big 470uF Silmic that Funch used. Not sure If I'll be able to fit two 12.5" wide caps side by side though, and still have room for batteries.
You could just change it into a desktop amp haha use the entire tin for caps and have a power jack, and as a portable setup could have another tin of just batteries, which you could plug in
longer runtimes tooWell I still have 2200 or 3300 FGs to try, and 1000uF Silmics are backordered till Aug 3 =(
I do have the smaller values of FG/KZ/Silmics though (~47uF - 470uF) got them thinking to use as bypassess if I found another large value cap which sounded good lol