With my HD800 with Bal-Bal and the Auzentech's 3v, or so, single ended output, I have to turn down the PC volume controls to only a few % to get it quiet enough.
HD800 are very revealing. I could easily hear the differences made to the sound by everything between the DAC output and the Bal-Bal, and I'm talking about extra solder joints, not just the type of cable, plugs, etc. I could verify for myself other peoples findings about the difference to the sound between gold plated and silver plated XLR's ,etc.
I also found the Bal-Bal to be sensitive to power supply as when I added some Black Gate caps across the polymer caps, the BG's did their normal thing.
Other than that I've have not tried different regulators and PSU parts so far with this apart from initially feeding the Bal-Bal from the soundcard 7808 and 7908 regs. I then instead soldered on the PSU from my old CD player's DAC / analogue stage, the spec of which I'd arrived at over years of trying stuff. It's not perfect I think and I suspect the regs could be better, but currently it has discrete regs, BG's x 10,000uF, Nichicon KZ x 20,000uF, a custom toroid with 100VA around a 300VA M0 core. There are no leads, switches fuses, the Tx primaries are plugged into the 'super spur' and it's permanently on. I strongly suggest you do not do the same unless you know what you're doing.
I found the soundcard supply was causing noise, 'beeping' and 'whistling' that changed with the HDD seek! The external PSU stopped all that. There is an only just perceptible amount of that now, coming in on the single ended signal lines I guess. Maybe one day I'll get the soundcard analogue stage off there and over by the Bal-Bal, maybe the DAC chip too.
HD800 are very revealing. I could easily hear the differences made to the sound by everything between the DAC output and the Bal-Bal, and I'm talking about extra solder joints, not just the type of cable, plugs, etc. I could verify for myself other peoples findings about the difference to the sound between gold plated and silver plated XLR's ,etc.
I also found the Bal-Bal to be sensitive to power supply as when I added some Black Gate caps across the polymer caps, the BG's did their normal thing.
Other than that I've have not tried different regulators and PSU parts so far with this apart from initially feeding the Bal-Bal from the soundcard 7808 and 7908 regs. I then instead soldered on the PSU from my old CD player's DAC / analogue stage, the spec of which I'd arrived at over years of trying stuff. It's not perfect I think and I suspect the regs could be better, but currently it has discrete regs, BG's x 10,000uF, Nichicon KZ x 20,000uF, a custom toroid with 100VA around a 300VA M0 core. There are no leads, switches fuses, the Tx primaries are plugged into the 'super spur' and it's permanently on. I strongly suggest you do not do the same unless you know what you're doing.
I found the soundcard supply was causing noise, 'beeping' and 'whistling' that changed with the HDD seek! The external PSU stopped all that. There is an only just perceptible amount of that now, coming in on the single ended signal lines I guess. Maybe one day I'll get the soundcard analogue stage off there and over by the Bal-Bal, maybe the DAC chip too.
Last edited:
Hi everyone! I will be building the BAL-BAL version soon and have a couple questions for you fine folks
1) What is the recommended toroidal for the BAL-BAL? 10VA, 20VA, 12v, 15v?
2) Alongside #1 what's a good IEC/fuse and switch?
3) My Yulong DA8 DAC bal out is 4.2v and I'll be using the internal preamp with HD-800 headphones. Will unity gain suffice, or do I need to raise it? If I do, what are the recommended resistors to switch out?
Thanks so much in advance!
Hi Brunk,
To answer your questions:
1) Any dual secondary 10VA or higher transformer will do the trick, and the secondary voltage depends on how high you want to run your rails. The BAL-BAL can accept a maximum of +/- 16V and for that you probably want 15VAC secondaries on the transformer. If you're alright with 12V rails then you can reduce that to dual 12VAC secondaries.
2) No ideas here... I use those ungodly expensive integrated units from Schurter with a switch, fuse and IEC inlet all in one, but discreet parts are probably cheaper and more DIY friendly.
3) As Ian said, 4.2V is likely going to be more than enough drive voltage. A quick calculation tells me you'll be at roughly 115dBSPL at that voltage output with unity gain on The Wire. If you really want to hit 120dBSPL then you would need a gain of 2 on the BAL-BAL, but I don't recommend that. I get away with 2VRMS on the HE-6 which gives a maximum of about 102dBSPL and I find that to be about perfect.
Cheers,
Owen
Many headphones have an absolute max voltage of ~5Vac.
Your 4.2Vac output is getting close to absolute max.
A unity gain is suitable.
Check the output current capability and see if it suits driving your headphones.
With a unity gain buffer fed from a digital 4.2Vac Source you will need a lot of attenuation to help stop you going deaf.
Headphones are typically 90dB to 100dB / milliwatt.
4.2Vac into 32ohms is ~ +27dBmW (550mW)
Your 4.2Vac output is getting close to absolute max.
A unity gain is suitable.
Check the output current capability and see if it suits driving your headphones.
With a unity gain buffer fed from a digital 4.2Vac Source you will need a lot of attenuation to help stop you going deaf.
Headphones are typically 90dB to 100dB / milliwatt.
4.2Vac into 32ohms is ~ +27dBmW (550mW)
A friend came over a few weeks ago with his HD650 and MDAC which we fed via USB from my 'pute.
Then swapped over to the HD800, Bal-Bal, soundcard.
He thought the latter was ridiculously good compared to the former.
And I know it can be better still.
Owen, it would be rather good if your new regulators meant all that excess PSU I'd had to use was rendered ineffective, which is what they say the super teddy regs do. I've not tried those so far due to some apathy compounded by difficulty obtaining the low noise Toshiba transistors.
Then swapped over to the HD800, Bal-Bal, soundcard.
He thought the latter was ridiculously good compared to the former.
And I know it can be better still.
Owen, it would be rather good if your new regulators meant all that excess PSU I'd had to use was rendered ineffective, which is what they say the super teddy regs do. I've not tried those so far due to some apathy compounded by difficulty obtaining the low noise Toshiba transistors.
Brunk,
Like I posted on Head-Fi... the Amveco (TE62043-ND from Digikey) transformer works quite well. It may not be the cheapest option, but it's a really good quality trafo.
It's 10VA, has dual primaries, dual 15VAC secondaries. You can wire it up for center tap (if you use a power supply that requires this) or in parallel, which I think is what OPCs new supply requires.
I didn't want to answer "over there" because I thought it would be more fruitful here, but yeah like OPC says... Unity Gain is fine for your source voltage. In my experience, I have two DACs -- one that does 1.4Vrms (AMB Gamma 2) and the other at 2Vrms (Curryman ES9023). The Wire really needed a gain of 2 for the Gamma 2's 1.4Vrms output to drive most of my headphones nicely -- LCDs, Thunderpants, DT880-250. With 2VRMS Unity Gain is fine for most headphones. However, I will say that with my recently acquired AKG K240 Sextetts (voltage hungry) you really need some extra gain even with the 2VRMS of the Curryman. So I'm glad I chose a gain of 2. I have resistors for a gain of 4, but at this point I'm not going to bother.
Like I posted on Head-Fi... the Amveco (TE62043-ND from Digikey) transformer works quite well. It may not be the cheapest option, but it's a really good quality trafo.
It's 10VA, has dual primaries, dual 15VAC secondaries. You can wire it up for center tap (if you use a power supply that requires this) or in parallel, which I think is what OPCs new supply requires.
I didn't want to answer "over there" because I thought it would be more fruitful here, but yeah like OPC says... Unity Gain is fine for your source voltage. In my experience, I have two DACs -- one that does 1.4Vrms (AMB Gamma 2) and the other at 2Vrms (Curryman ES9023). The Wire really needed a gain of 2 for the Gamma 2's 1.4Vrms output to drive most of my headphones nicely -- LCDs, Thunderpants, DT880-250. With 2VRMS Unity Gain is fine for most headphones. However, I will say that with my recently acquired AKG K240 Sextetts (voltage hungry) you really need some extra gain even with the 2VRMS of the Curryman. So I'm glad I chose a gain of 2. I have resistors for a gain of 4, but at this point I'm not going to bother.
Last edited:
Here you go
BAL-BAL = 2.325" x 1.625"
SE-SE = 1.825" x 1.150"
BAL-SE = 2.325" x 1.625"
PSU = 2.525" x 1.150"
http://www.diyaudio.com/forums/head...-headphone-amplifier-pcbs-76.html#post2748697
Hey all, have a question (might be something obvious I'm overlooking) in regards to my design that I'll be using with The Wire.
I realize that it is unity gain, so whatever coming in is what will be going out. However, what exactly should be the voltage that is coming out of the The Wire to my headphone jack?
Right now I have HD598s, but I am trying to design a DAC+Amp strong enough for anything I might throw at at.
I realize that it is unity gain, so whatever coming in is what will be going out. However, what exactly should be the voltage that is coming out of the The Wire to my headphone jack?
Right now I have HD598s, but I am trying to design a DAC+Amp strong enough for anything I might throw at at.
If my memory serves me, I believe he said a gain of 2 gives him just barely the volume he needs for the HE-6 at full blast.
I do in fact run unity gain with a source that provides 2VRMS @ 0dBFS when I use the HE-6.
I frequently listen at full volume though (ideally where you should be listening) and I find it just right on most recordings.
Understand though that we're into an area that is highly subjective, and some people might listen at levels that are easily 12dB higher than I like. Some will listen 12dB lower.
I've been handed headphones at audio shows from someone saying "give these a listen" and the levels have been both ear blisteringly loud, and so quiet I could hardly tell anything was playing. It really is about personal preference, music selection, mood, time of day, environment, and a million other factors.
The general point I'd like to make though is not to be afraid of running the volume at full tilt, and that being a little short on gain is probably better than being way too high. I know some people who run a gain structure that basically means they never get above 1/4 volume, and that's just poor planning. Maximum volume should be just that, the highest volume you're likely to listen at. If you've never got beyond half, you're throwing a lot of dynamic range out the window for what amounts to a mental placebo.
Cheers,
Owen
I frequently listen at full volume though (ideally where you should be listening) and I find it just right on most recordings.
Understand though that we're into an area that is highly subjective, and some people might listen at levels that are easily 12dB higher than I like. Some will listen 12dB lower.
I've been handed headphones at audio shows from someone saying "give these a listen" and the levels have been both ear blisteringly loud, and so quiet I could hardly tell anything was playing. It really is about personal preference, music selection, mood, time of day, environment, and a million other factors.
The general point I'd like to make though is not to be afraid of running the volume at full tilt, and that being a little short on gain is probably better than being way too high. I know some people who run a gain structure that basically means they never get above 1/4 volume, and that's just poor planning. Maximum volume should be just that, the highest volume you're likely to listen at. If you've never got beyond half, you're throwing a lot of dynamic range out the window for what amounts to a mental placebo.
Cheers,
Owen
Probably just whatever orientation gets it the farthest away. The leads running into and out of the amp are probably more susceptible to picking up a stray magnetic field than anything on the board itself.
Some experimentation might be in order though, as it's hard to visualize without seeing everything.
Cheers,
Owen
Perhaps I should stop dreaming
Fortunately I am able to follow your advice Owen. I was anxious to exploit the charactistic of properly wound R-cores that results in cancellation of leakage flux at the short-leg ends. In combination with the noise immunity inherent in balanced circuits I was dreaming of snugging a Bal-Bal right up next to the R-core.
:: Shilchar Technologies Limited ::
Fortunately I am able to follow your advice Owen. I was anxious to exploit the charactistic of properly wound R-cores that results in cancellation of leakage flux at the short-leg ends. In combination with the noise immunity inherent in balanced circuits I was dreaming of snugging a Bal-Bal right up next to the R-core.
:: Shilchar Technologies Limited ::
Hi Wolfsin,
It might work just fine, so I wouldn't rule it out. I have no experience at all with R-cores so the side null might be low enough not to have an impact.
My experience with toroids though is that distance is always your friend. Orientation is also critical, but the surefire way to eliminate magnetic coupling is to just keep the transformer far away from everything that carries a signal.
Give it a try, but make sure you have a plan B
Cheers,
Owen
It might work just fine, so I wouldn't rule it out. I have no experience at all with R-cores so the side null might be low enough not to have an impact.
My experience with toroids though is that distance is always your friend. Orientation is also critical, but the surefire way to eliminate magnetic coupling is to just keep the transformer far away from everything that carries a signal.
Give it a try, but make sure you have a plan B
Cheers,
Owen
It's not a potentiometer ground.
On the input side it's the return wire for the input circuit.
On the output side it is the return wire for the output circuit.
Both those circuits MUST be complete for the current to flow around each circuit.
The input is fed with a two wire input signal.
The output sends a two wire signal to the next stage.
On the input side it's the return wire for the input circuit.
On the output side it is the return wire for the output circuit.
Both those circuits MUST be complete for the current to flow around each circuit.
The input is fed with a two wire input signal.
The output sends a two wire signal to the next stage.
OPC and all:
I am uncertain about the final BOM for the new version of the PSU. There were a number of posts sometime in October and a note by OPC, stating that there would be an update or final BMO posted somewhere, but I am unable to find it. Is there a wiki for the new boards?
Thank you all for your help and contributions to this fun project.
I am uncertain about the final BOM for the new version of the PSU. There were a number of posts sometime in October and a note by OPC, stating that there would be an update or final BMO posted somewhere, but I am unable to find it. Is there a wiki for the new boards?
Thank you all for your help and contributions to this fun project.