Back on topic... does anyone have suggestions for the default low and high gains?
I've heard 2X and 8X, 3X and 7X, 3X and 6X... obviously for those building the board from scratch they can use whatever gain resistors they want in. But I have to put some sort of default resistor values on the revised BOM and schematic when I release the artwork hopefully this weekend.
Right now the low gain defaults to 3.1X (10 dB) which is about as much as the input gain stage can handle from typical home (2 V RMS) line outputs.
Some want a pre-assembled PC board to make building the O2 easier. And at least one vendor has expressed an interest in offering such a board. So those without soldering irons won't have an easy way to change the resistors (besides clipping one pair for 1X gain). That makes getting the default values right more important.
I've heard 2X and 8X, 3X and 7X, 3X and 6X... obviously for those building the board from scratch they can use whatever gain resistors they want in. But I have to put some sort of default resistor values on the revised BOM and schematic when I release the artwork hopefully this weekend.
Right now the low gain defaults to 3.1X (10 dB) which is about as much as the input gain stage can handle from typical home (2 V RMS) line outputs.
Some want a pre-assembled PC board to make building the O2 easier. And at least one vendor has expressed an interest in offering such a board. So those without soldering irons won't have an easy way to change the resistors (besides clipping one pair for 1X gain). That makes getting the default values right more important.
I think 3 and 6 is appropriate for 99% of the phones out there. On the other special cases where you have special needs they could possibly order them with the resistor locations unpopulated. Perhaps 4/8.. I am using 6 right now for 32-ohm Sennheiser cans and even with my cheapo Zen Stone as an input, it's still too hot. When plugged into my PC it's way too hot. I go to maybe 30% on the pot.
It think it's time you tested some worst case scenarios like 15 Ohm mag planer and 600+ Ohm Sennheiser. That is something your scope isn't going to tell you. You get a good idea, but it's really not plainly obvious how much pot travel a particular set of phones is going to have on any one gain setting. You also want it to fit nicely on the log curve. Usually the last 3/4 if the pot is the target area as the curve will be such that the volume change is less. Most commercial pots are designed with not a true log curve. It's usually something slightly more tuned.
It think it's time you tested some worst case scenarios like 15 Ohm mag planer and 600+ Ohm Sennheiser. That is something your scope isn't going to tell you. You get a good idea, but it's really not plainly obvious how much pot travel a particular set of phones is going to have on any one gain setting. You also want it to fit nicely on the log curve. Usually the last 3/4 if the pot is the target area as the curve will be such that the volume change is less. Most commercial pots are designed with not a true log curve. It's usually something slightly more tuned.
Thanks Rembrant. I'm not aware of any 15 ohm planars, are you? The HiFiMan planars seem to represent the worst case loads in terms of low impedance/low efficiency and hence are the ones I designed the amp aroundy. Ditto for the Beyer DT880-600 I benchmarked at the high impedance end of the spectrum. I agree real world listening tests with those headphones is a good idea.
I don't own those cans, but I know people who do so it's on the (eventual) agenda. Headphones at the extremes, obviously, may well have different gain requirements. But that's fine. I'm trying to follow the 80/20 rule here and cover the biggest majority possible.
I don't own those cans, but I know people who do so it's on the (eventual) agenda. Headphones at the extremes, obviously, may well have different gain requirements. But that's fine. I'm trying to follow the 80/20 rule here and cover the biggest majority possible.
Since your SuperFi 5s (If I found the right model anyway...) are sort of on the low end of impedance and the high end of sensitivity for even 2 and 3 way BA IEMs you might want to see how high the "low" gain mode could go without being too loud from an average or slightly "hot" source as soon as the pot begins to track properly and set it around there.
If its not too low it would probably be a step towards giving it the broadest range of compatibility.
I don't think there's necessarily any "right" answer as optimizing gain across headphones of wildly differing efficiencies and different source levels is probably more than you can do with only 2 different settings unless the pot is perfect. I've seen the measurements you did for the tracking but can't figure out a decent way to figure out how that applies to my IEMs.
I may be a fringe case though. My two favorite headphones (Shure SE530 and Fostex T50RP) have more than a 20dB difference in efficiency between them. After factoring in the occasional use of 'phones with efficiencies in between that I haven't figured out a good set of gains for mine yet.
I might have to buy all the different ones and experiment since I can't figure out a good way to get and SPL reading on my IEMs...
If its not too low it would probably be a step towards giving it the broadest range of compatibility.
I don't think there's necessarily any "right" answer as optimizing gain across headphones of wildly differing efficiencies and different source levels is probably more than you can do with only 2 different settings unless the pot is perfect. I've seen the measurements you did for the tracking but can't figure out a decent way to figure out how that applies to my IEMs.
I may be a fringe case though. My two favorite headphones (Shure SE530 and Fostex T50RP) have more than a 20dB difference in efficiency between them. After factoring in the occasional use of 'phones with efficiencies in between that I haven't figured out a good set of gains for mine yet.
I might have to buy all the different ones and experiment since I can't figure out a good way to get and SPL reading on my IEMs...
LOL about the 15-Ohm planers. I plan on making a set one day and they will certainly be low efficiency and probably low Z, maybe 15. We will see. It all depends on what I can be done with the stator and magnets. People were using those planer tweeters but I don't think those can be had in the higher Ohms any more. I have searched and all that are left are very low Z. So basically useless.
Maverick, it's true the math only gets you so far--especially if you're using the mfg. specs. Tyll's numbers at InnerFidelity are probably at least a bit more consistent and realistic. I have more than a half a dozen full size cans and more than a dozen in-ear headphones but those quickly boil down to the ones that are closest to the extremes. Those are the five I used in the blind test with the O2 vs Benchmark DAC1 Pre. But even those don't cover nearly the range that's out there.
I'm told my critics are attacking the O2 on the basis the input stage can be overloaded. But they're stretching the truth. They're talking about the O2 running on BATTERY power but using an AC powered home source. How likely is that?
In reality, the O2 can handle 99% of all home sources when it's running on AC power at the default 3X gain. And it can also handle 100% of portable battery powered sources I know of when running on batteries. So where's the problem?
And for the 1% fringe case, it's very easy to modify the gain of the O2 to suit any unusual needs. The benefits of the O2's gain structure easily outweigh the 1% that might encounter a problem and need to tweak the gain.
The O2 is very much designed for the real world. Everything is a trade off involving performance, cost, complexity, etc. The O2 is designed for optimal performance for 99% of the applications. If my critics want to argue the 1% case, that seems fairly desperate.
I've also always made the gain issues, including input stage overload, very clear on my blog. The information is there for anyone concerned about the issue.
In reality, the O2 can handle 99% of all home sources when it's running on AC power at the default 3X gain. And it can also handle 100% of portable battery powered sources I know of when running on batteries. So where's the problem?
And for the 1% fringe case, it's very easy to modify the gain of the O2 to suit any unusual needs. The benefits of the O2's gain structure easily outweigh the 1% that might encounter a problem and need to tweak the gain.
The O2 is very much designed for the real world. Everything is a trade off involving performance, cost, complexity, etc. The O2 is designed for optimal performance for 99% of the applications. If my critics want to argue the 1% case, that seems fairly desperate.
I've also always made the gain issues, including input stage overload, very clear on my blog. The information is there for anyone concerned about the issue.
Hi
I would be running the O2 from one of my AC powered desktop dacs.... either a North Star, Stello or Constantine and I would be running the O2 from AC power as well.
I'm sure you've read the accusations so I won't repost them, but is it true that if I turn the gain up for my T-1s or 650s the O2 is going to clip?
USG
USG,
Basic math and reading will answer your question.
The O2 runs off dual 9V supplies. Assuming you had the gain at 3X, that means the input can take no more than 9/3 = 3Vpk before it clips.
The so-called "accusations" are all about basic gain matching. There's even an article in this site about it. Plenty of good preamps can put out more than 8V easy. Plenty of power amps clip at no more than 3V input. People who buy them don't ask your type of question because they know what to do.
The O2 might be beyond your technical level if you attempt to DIY it. Better off buying a commercial product with warranty.
Basic math and reading will answer your question.
The O2 runs off dual 9V supplies. Assuming you had the gain at 3X, that means the input can take no more than 9/3 = 3Vpk before it clips.
The so-called "accusations" are all about basic gain matching. There's even an article in this site about it. Plenty of good preamps can put out more than 8V easy. Plenty of power amps clip at no more than 3V input. People who buy them don't ask your type of question because they know what to do.
The O2 might be beyond your technical level if you attempt to DIY it. Better off buying a commercial product with warranty.
I'm not sure I could build this then.
Tell me if I have this right: The O2 will have issues with extremely hard to drive headphones because the high gain necessary will cause clipping? But can this be fixed easily if you modify it?
It might be a valid criticism because it is supposed to be a "do it all" amp, but the critics (well, amb) is also saying the input will always clip, which makes no sense at all using the standard gain and a typical input right? He seems to be making a mountain out of a molehill just because NwAvGuy pissed him off.
Maybe a std. gain of 2x might be good just to be safe, but I wouldn't really know.
I'm sorry if us non-technical people are coming across as idiots, but we are just trying to understand, and I think this amp interests a lot of people who aren't that familiar with DIY.
Etilletas = Satellite (no DIY experience)
Tell me if I have this right: The O2 will have issues with extremely hard to drive headphones because the high gain necessary will cause clipping? But can this be fixed easily if you modify it?
It might be a valid criticism because it is supposed to be a "do it all" amp, but the critics (well, amb) is also saying the input will always clip, which makes no sense at all using the standard gain and a typical input right? He seems to be making a mountain out of a molehill just because NwAvGuy pissed him off.
Maybe a std. gain of 2x might be good just to be safe, but I wouldn't really know.
I'm sorry if us non-technical people are coming across as idiots, but we are just trying to understand, and I think this amp interests a lot of people who aren't that familiar with DIY.
Etilletas = Satellite (no DIY experience)
Last edited:
The amp was made to be useful for sources that don't have a preamp out. Like portable devices and such. These guys are on about something because they can't blowtorch the input with 4vp from a preamp that they could either remove from the signal chain altogether, or simply attenuate. The loudness of this amp, at the right gain setting, would not require you to overdrive the input with any set of phones I can think of.
Why this happens.
The volume control pot was placed between the gain and buffer stages of the amp for various reasons. One of them being the opamp that was chosen for the gain stage. Ultimately, the attenuation could be moved to its own stage with buffers flanking but that would require at least two extra dual opamps and more board space etc. They could blowtorch that attenuation stage all they wish. There are always trade offs. This amps claim is superior performance at reduced cost. Not that one can't overload the input with a ridiculous amount of signal.
Just to be clear, the O2 on AC power runs off 12 volt rails. The maximum input is 7/X where X is the gain factor. So with 3X gain it can handle 2.3 V RMS input which is more than most anything without a volume control can produce. If your source has a volume control, there's no issue at all.
It's also worth noting the input stage and output stage (at max volume) clip at the same level. And there are hardly any headphones on the planet that need 7 V RMS.
It's also worth noting the input stage and output stage (at max volume) clip at the same level. And there are hardly any headphones on the planet that need 7 V RMS.
I don't fully understand it either but I've decided to lower the default gain to 2.5X (8 dB) to provide some extra margin for rare sources. This allows up to 2.8 V RMS of input. I've already revised the article and the new resistor values will be in the revised design documents.
If anyone knows of realistic O2 sources, without volume controls, that have more than 2.8 V RMS of output, please post a link to the specifications here. And if there are such sources, the gain resistors can be simply clipped dropping the O2 to 1X gain and you could still drive the HD600s beyond an ear splitting 110 dB SPL with such a source.
dacmagic? 4.2vrms in balanced mode: Specifications / DacMagic Digital to Analogue Converter / Digital Music / Products / Cambridge Audio
CMIIW
Last edited:
But that's 2.1 V RMS in unbalanced mode and would work great with the O2's unbalanced input.
I just looked up my Stello DA100 and it has a 2.4Vrms output signal level. I've never thought of the Stello as being a "hot" source but it seems to be very close to the maximum with a 2.5 gain.
My Pico portable dac is listed as having an output of 2.84Vrms on the HeadAmp web site (maybe I'm reading it wrong). Link I never thought of the Pico as being a particularly "hot" source either and have plugged it into all my amps, both portable and desktop, with no problems.
My feeling is that there might be a few more 2+V sources out there then we realize, but I'll leave it to the EEs here to sort it out.
yeah but personally, i always use the balanced output using xlr-rca converter and feed it to my single ended amp since the balanced out sounds better than the single ended one
and i'm pretty sure some members here are doing the same thing since it's covered in dacmagic mod thread somewhere in the Digital Line Level section in this forumanyway, i was just trying to give some input
Last edited: