@ethanolson you are correct on the ODA DAC supporting 24/96. Thanks. And you bring up a good point about Windows automatically using native bit rates rather than forcing re-sampling.
It's also worth mentioning until a few releases ago, it was very difficult to get bit accurate audio out of OS X and iTunes. Benchmark Media wrote an article on it as they discovered the DAC1 performed poorly on Macs.
Perhaps it's just me, but my experience with support of anything more exotic than a USB keyboard or mouse (HID) in Linux has been mixed. Even some storage class devices fail to mount without opening a console session. Linux and USB has very much been "Plug & Pray" if you're not a hardcore Linux geek. It's gradually getting better but it seems eternally about a decade behind Windows for ease-of-use when it comes to adding hardware.
It's also worth mentioning until a few releases ago, it was very difficult to get bit accurate audio out of OS X and iTunes. Benchmark Media wrote an article on it as they discovered the DAC1 performed poorly on Macs.
Perhaps it's just me, but my experience with support of anything more exotic than a USB keyboard or mouse (HID) in Linux has been mixed. Even some storage class devices fail to mount without opening a console session. Linux and USB has very much been "Plug & Pray" if you're not a hardcore Linux geek. It's gradually getting better but it seems eternally about a decade behind Windows for ease-of-use when it comes to adding hardware.
I agree about the plug&pray... It isn't always as easy as promoted. But in the end how should the standards be interpreted? As they are written? Then we should have a lot to do on the support side of our organizations. Best practices - a broader keyhole - I believe that this is the dominant way today. MS-centric - well some distributors have choosen this path and does what MS tells them to do. These are the ones first out with drivers in new releases of windows.
My support lies with the standards even though the majority might settle for best practises. With the majority I mean a majority who cares about the functionality between architectures. The MS clan or the others doesn't bother as long it is working for them and why should they. Unification is the way but that doesn't sell well in the salesdepartments.
Yes, thank you ethanolson 24/96 is enough for me as a starter - I have other options for higher bitrates.
Brgds
My support lies with the standards even though the majority might settle for best practises. With the majority I mean a majority who cares about the functionality between architectures. The MS clan or the others doesn't bother as long it is working for them and why should they. Unification is the way but that doesn't sell well in the salesdepartments.
Yes, thank you ethanolson 24/96 is enough for me as a starter - I have other options for higher bitrates.
Brgds
Last edited:
Now, the question was not aimed for me but the values, I believe, is a few pages back. I went back to the original 2.7M since I see the behaviour more as a signal to recharge the batteries and I want as long as possible between recharges. As it is constructed today I don't see a foolproof way of getting this part to work in every case. People will put any possible load on the amp and the amp will run on batteries of any brand and capacity. As long as these things ain't under the designers control there is not much he can do. It could have been a lot worse - it could have been something affecting sound quality.
Here are the boundaries from RS earlier post:
"R25 = 1.2M - 1.5M (or parallel a second 2.7M on top of the exiting 2.7M)
R9 = 33K - 36K (or parallel 270K on top of the existing 40.2K)
Lowering R9 to around 34K significantly raises the "turn back on voltage" to where it should stay off for a longer period of time if not until you power cycle the amp and/or charge the batteries."
Now, when given time by the wife, kids, dogs, work and any other material or unmaterial request I will solder in some potentiometers instead of these resistors and run an trial/error sequence to find what fits my setup of batteries/load. It is itchy this...
The first board I found it on used the Tenergy cells, on the second I have Fisher Amps NiMh Accu 270mAh and they have very different behavioural signatures so I believe the solution will differ in needed resistor values. Is there is a "window" that fits all? I don't know but hey! This is the fun part - this is DIY! Pick your values that works for you and be happy with it until you change the batteries. You could even report what works with different choises regarding batteries. I will report, when I get the Fisher and Tenergy working, my chosen values.
Brgds and I hope it was OK to toss in my coin...
Last edited:
@ethanolson & Turbon, I put a few notes in the O2 Details article and updated the XLS BOM spreadsheet regarding the low voltage shutdown issue and the resistor values. I still need to go through what's in stock at Mouser, incorporate the other BOM suggestions, update the schematic, regenerate the PDF documentation package, update the article, update Google Docs, etc.
Also, Turbon, the mod doesn't really change the run time. Ni-Mh batteries have a very flat discharge curve. By the time they're below 7 volts you have only a few minutes out of 400+ minutes left. And the mod mainly changes the turn-on voltage, not the turn-OFF, voltage.
Also, Turbon, the mod doesn't really change the run time. Ni-Mh batteries have a very flat discharge curve. By the time they're below 7 volts you have only a few minutes out of 400+ minutes left. And the mod mainly changes the turn-on voltage, not the turn-OFF, voltage.
RS, I believe that the batteries is causing this but they arn't the problem. If you discarge them slowly the is no rebound effect but if you load them heavy they will have an bounce back effect. I am pretty shure that this will need different approaches depending on the choice of batteries and load. I will when given time try to find the right equation for HD-650 combined with the Tenergs and Fisher batteries and I'm shure of that the resistor values will differ. You should have implemented logic instead from the beginning - when it decides to shut off -it will stay shut off until you decide to turn it on again... I mean a real shutoff - nothing happening behind the on/off circuit can do anything to alter the state.
Brgds
Brgds
Last edited:
@Turbon, the HD650 is a really easy load for the O2 at 300 ohms. Even driving the HD650 to peaks of 110 dB SPL only requires about 2.5 Vrms and the O2 is capable of 5+ Vrms on battery power.
The main issue here is the internal resistance of the batteries. All batteries have an internal impedance, and it's significant for 9 V Ni-Mh batteries due to the small size of the each of the 7 cells inside. The resistance rises dramatically as the batteries discharge. So, in effect, you end up with a "perfect" 8.4 volt DC voltage source with a significant value of resistance in series. It's the drop across that resistance that causes the battery voltage to fall with the 25 - 35 mA load of the O2. Using 2 volts drop that's about 60 ohms of series resistance.
When the amp shuts itself off, the drop across the internal resistance falls dramatically as now there is only 0.5 mA load on the batteries. So you get 0.0005 * 60 = 0.03 volts drop and the voltage rises dramatically.
Different batteries are optimized for different things. To add more mAH capacity to a battery without making it bigger generally means a higher internal resistance--especially towards the end of their discharge curve.
The main issue here is the internal resistance of the batteries. All batteries have an internal impedance, and it's significant for 9 V Ni-Mh batteries due to the small size of the each of the 7 cells inside. The resistance rises dramatically as the batteries discharge. So, in effect, you end up with a "perfect" 8.4 volt DC voltage source with a significant value of resistance in series. It's the drop across that resistance that causes the battery voltage to fall with the 25 - 35 mA load of the O2. Using 2 volts drop that's about 60 ohms of series resistance.
When the amp shuts itself off, the drop across the internal resistance falls dramatically as now there is only 0.5 mA load on the batteries. So you get 0.0005 * 60 = 0.03 volts drop and the voltage rises dramatically.
Different batteries are optimized for different things. To add more mAH capacity to a battery without making it bigger generally means a higher internal resistance--especially towards the end of their discharge curve.
Yes, I read what you are writing RS. So do you expect the new values to work in every case? (Sorry for beeing such pain in the butt).
The interesting part happens when the design decides to shut parts of it down and gives the batteries time to rebound and the design decides to turn it on again... I would rather se that after there is a shutdown there is no coming back until I decide so with a press on a button.
Brgds
The interesting part happens when the design decides to shut parts of it down and gives the batteries time to rebound and the design decides to turn it on again... I would rather se that after there is a shutdown there is no coming back until I decide so with a press on a button.
Brgds
Last edited:
I expect the new values to significantly help. So far, out of perhaps 1500 O2 PC boards in "circulation", only 3 or 4 people have commented on this problem with the old values. The original design works perfectly here on my prototypes. I didn't see the need for a more costly or complex design that may not have fit on the already very crowded O2 PC board.
A comparator was already needed to measure the battery voltage and this circuit elegantly makes do with just a dual comparator. The O2 is about maximum audio performance for the lowest cost. As long as the headphones are not slammed with several volts of DC when one battery dies first or becomes disconnected, the power control circuit is performing its primary function. And it does so with even the original resistor values. Adding more circuitry wouldn't improve the audio performance at all and would only make the amp more expensive and no longer fit in the B2-080 enclosure.
A comparator was already needed to measure the battery voltage and this circuit elegantly makes do with just a dual comparator. The O2 is about maximum audio performance for the lowest cost. As long as the headphones are not slammed with several volts of DC when one battery dies first or becomes disconnected, the power control circuit is performing its primary function. And it does so with even the original resistor values. Adding more circuitry wouldn't improve the audio performance at all and would only make the amp more expensive and no longer fit in the B2-080 enclosure.
OK RS, I will implement your proposals ASAP (in a few weeks...).
I will of course keep you updated on whatever.
Now I have 3 boards where 2 is positive on this. The 3'rd aint tested to this level yet. 1 of the boards being positive was not built by me but to your specs even if I ave found some other anomalies in gain behaviour.
I have ordered parts for a 4'th kit where I will test your proposals. To bad that the BOM I found on your site still seems to be the old on with a 2.7M at R25...
Maybe you can give a link to the latest.
I might add that this issue has a very small sonic impact, it's totally in the time domain. It's really amazing soundwise if I haven't said so earlier.
Brgds
I will of course keep you updated on whatever.
Now I have 3 boards where 2 is positive on this. The 3'rd aint tested to this level yet. 1 of the boards being positive was not built by me but to your specs even if I ave found some other anomalies in gain behaviour.
I have ordered parts for a 4'th kit where I will test your proposals. To bad that the BOM I found on your site still seems to be the old on with a 2.7M at R25...
Maybe you can give a link to the latest.
I might add that this issue has a very small sonic impact, it's totally in the time domain. It's really amazing soundwise if I haven't said so earlier.
Brgds
Last edited:
@Turbon, thanks for catching the BOM. I did update it, but Google Docs saved it has another version (i.e. a different link). It should be fixed.
I'm glad you like the amp! I'm not sure what you mean about sonic impact? The power management circuit has no effect on the amp during normal operation. The MOSFETs are either on or off. Once the batteries are so low as to trigger shutdown, you shouldn't be trying to listen to the amp anymore. And when the MOSFETs are on, the comparator circuit has no effect at all on sound quality.
I'm glad you like the amp! I'm not sure what you mean about sonic impact? The power management circuit has no effect on the amp during normal operation. The MOSFETs are either on or off. Once the batteries are so low as to trigger shutdown, you shouldn't be trying to listen to the amp anymore. And when the MOSFETs are on, the comparator circuit has no effect at all on sound quality.
Hi there, I'm planing on building one of these beauties, I have only two questions, witch box looks better, the black or the blue one ?
the second one is more interesting, I'm planning on driving my AKG k1000 and my Grado r225 from it, so I'm interested if anyone had experience driving the k1000s from this amp.
According to this I should get 99.8dB out of my headphones, witch is plenty (I'm currently running the from my Lehmann BCL clone, and it gets loud enough
(but it could be better...) - so how can I measure the output voltage of my matrix mini-i so i can set the gain precisely, so it wont clip at full volume.
Thanks
the second one is more interesting, I'm planning on driving my AKG k1000 and my Grado r225 from it, so I'm interested if anyone had experience driving the k1000s from this amp.
According to this I should get 99.8dB out of my headphones, witch is plenty (I'm currently running the from my Lehmann BCL clone, and it gets loud enough
(but it could be better...) - so how can I measure the output voltage of my matrix mini-i so i can set the gain precisely, so it wont clip at full volume.Thanks
The connection between the LED and series ballast resistor R6 is also connected to the + inputs of comparators U2a & U2b. So the forward voltage of this particular LED is critical to the proper comparator operation. Substitutes can be used for the LED, but they require measurements (and probably changing other resistor values also) to get the protection circuit "balanced" again and operating correctly.
Hi there, I'm planing on building one of these beauties, I have only two questions, witch box looks better, the black or the blue one ?
the second one is more interesting, I'm planning on driving my AKG k1000 and my Grado r225 from it, so I'm interested if anyone had experience driving the k1000s from this amp.
I should get 99.8dB out of my headphones, witch is plenty (I'm currently running the from my Lehmann BCL clone, and it gets loud enough
The K1000 is really designed to be driven by a speaker output, not a headphone output. I don't recommend the O2 for the K1000. Even on AC power, 7 Vrms just isn't enough unless you like to listen to compressed pop music at modest levels. Wide dynamic range music requires peaks of at least 105 dB SPL with 110 dB SPL being a better number. The average SPL will still be below 90 or even 85 dB when peaks are hitting 105 dB with wide dynamic range music. If you haven't seen it, you might want to check out my article on headphone and amp compatibility:
More Power?