@Jelle, good point on the TRRS ground issue when using TRS plugs. You would have to switch that extra ring contact somehow for TRS operation.
The Clip+ is amazing for what it is, but there are all sorts of headphones that are way beyond its voltage capability--including such favorites as the HD600/650 (at least if you want anything approaching realistic listening levels). The Clip+ is only good for around 0.5 Vrms and many headphones need 1 - 5 Vrms. I already have a Clip Zip on order and will be testing it (the Clip+ is officially history).
@Orpheus, the O2 board already has a 4 pin header on it for a 4 pin headphone connector (off-board). I was planning to do the same with the desktop amp. If by some miracle I can get the connector footprints to "nest" I could also try for an on-board 4 pin that would replace the 3.5mm or 1/4" jack. Otherwise I'm not sure there will be enough space but I can try.
The main benefit is improved crosstalk and I really doubt it will make any sort of audible difference to get rid of a few milliohms of shared impedance. But I also understand many audiophiles like going that extra mile and, in this case, it certainly can't hurt. It would also be useful for someone who's already re-wired their headphones for another amp.
Does anyone know if there's a "most popular" 4-wire headphone plug (i.e. what's already on the most headphones)? What's most used with the "fully balanced" amps?
The Clip+ is amazing for what it is, but there are all sorts of headphones that are way beyond its voltage capability--including such favorites as the HD600/650 (at least if you want anything approaching realistic listening levels). The Clip+ is only good for around 0.5 Vrms and many headphones need 1 - 5 Vrms. I already have a Clip Zip on order and will be testing it (the Clip+ is officially history).
@Orpheus, the O2 board already has a 4 pin header on it for a 4 pin headphone connector (off-board). I was planning to do the same with the desktop amp. If by some miracle I can get the connector footprints to "nest" I could also try for an on-board 4 pin that would replace the 3.5mm or 1/4" jack. Otherwise I'm not sure there will be enough space but I can try.
The main benefit is improved crosstalk and I really doubt it will make any sort of audible difference to get rid of a few milliohms of shared impedance. But I also understand many audiophiles like going that extra mile and, in this case, it certainly can't hurt. It would also be useful for someone who's already re-wired their headphones for another amp.
Does anyone know if there's a "most popular" 4-wire headphone plug (i.e. what's already on the most headphones)? What's most used with the "fully balanced" amps?
Here is a post from qusp on 4 wire connectors in the "Wire" thread. He does a lot with balanced stuff
http://www.diyaudio.com/forums/head...e-headphone-amplifier-pcbs-9.html#post2581996
I was about to suggest mini XLR on the desktop version for a different reason. TRS jacks always short by nature on insert and removal. I'm kinda skeptical the NJM4556s are going to last over the long term with just 1R in series and a TRS jack. It would be interesting to make a reliability test jig with a relay hooked to a computer. Try shorting the O2 outputs to ground for 500mS each time, to simulate a TRS insertion/removal, a total of 1000 - 10,000 times at random intervals. Output running at the full 7Vrms on +/- 12V rails into a 600R load.
With miniXLR that issue goes away, as long as someone keeps their TRS-to-miniXLR adaptor connected first to the headphone.
http://www.diyaudio.com/forums/head...e-headphone-amplifier-pcbs-9.html#post2581996
I was about to suggest mini XLR on the desktop version for a different reason. TRS jacks always short by nature on insert and removal. I'm kinda skeptical the NJM4556s are going to last over the long term with just 1R in series and a TRS jack. It would be interesting to make a reliability test jig with a relay hooked to a computer. Try shorting the O2 outputs to ground for 500mS each time, to simulate a TRS insertion/removal, a total of 1000 - 10,000 times at random intervals. Output running at the full 7Vrms on +/- 12V rails into a 600R load.
With miniXLR that issue goes away, as long as someone keeps their TRS-to-miniXLR adaptor connected first to the headphone.
Last edited:
@agdr, you make a good point about 4 pin connectors being immune to the shorting problem. But I think few headphone fans are in the habit of plugging their headphones in, or unplugging them, with the volume cranked up. It's sort of Headphone Use 101 and is in most every owner's manual.
It's good to be skeptical, but honestly, how realistic is what you propose? The O2 isn't being sent into outer space. And it's not a cost-is-no-object design.
Who's going to be plugging and unplugging their headphones thousands of times at an output level all but a very few headphones can even deal with? Few will ever get the O2 past 2 or 3 Vrms and even they're unlikely to be plugging and unplugging headphones regularly at that level. Typical listening levels are much lower.
I've shorted the 4556 under a variety of conditions. It has active current limiting and limits the current to about 100 mA peak per op amp. I've not had a single failure despite plenty of abuse. I think the big issue is heat and I've discussed that in the dissipation section of the Circuit Description. If the shorts are brief heat shouldn't be a problem. The big issue is the unknown SOA of the output transistors and secondary breakdown characteristics. But if that were a big problem it should have shown up in my testing.
NJM supposedly designed the 4556 for headphone duty. It's been used in thousands of Grado RA-1 commercial amps and countless thousands more clones and Cmoys. Is anyone aware of 4556 failures being a problem?
NJR doesn't specify it needing a series output resistor to protect it from TRS jack shorts. In fact they don't specify any cautions regarding short circuits or maximum current levels.
There are countless other headphone amps, including some very expensive commercial and DIY designs, that have zero current limiting and are indeed prone to blowing up if you do what you suggest. Their designers claim it was a worthwhile compromise for various reasons. I don't agree with having an amp that fragile and believe the O2 strikes a fair compromise for a minimalist low cost amp--as hopefully will the desktop version.
The alternatives to solve your concern are not very attractive. They either involve significantly compromising the performance of the amp--in audible ways--by adding series output resistance. Or using expensive output buffers which still might be prone to short circuit damage according to their datasheets. There are locations on the board to add extra series resistance if anyone wants. You can also increase the value if the 1 ohm isolation resistors.
I've said all along, if someone wants more of an "ultimate" design, there are several to choose from such as The Wire and the National LME49600 reference design. I'm trying to offer something a bit more modest and think it's only fair the judge it accordingly.
Last edited:
how exactly is mini XLR superior to RCA for single ended signals i wonder? other than size, i think you will find the contact/connection is better in every way. TRS i'll give you, its crap; anything that shorts the outputs to ground on the way in or out is easily bettered now, as is mini XLR, but the O2 is going for mass appeal as much as high performance. i also find that many of todays minijacks are prone to damage over time from leverage
Last edited:
@qusp, I'm certainly not complaining about RCAs. Like you said, you have to go with what the masses want. Between 3.5mm and RCA inputs you've covered 99+% of the users.
But I'm still unclear as to the "standard", if there is one, for 4-wire headphone plugs? I've seen full size XLR and mini-XLR. Which is more popular and are all the pin-outs the same? I would think the guys running around Can Jam would want some sort of standard so they can try different gear?
But I'm still unclear as to the "standard", if there is one, for 4-wire headphone plugs? I've seen full size XLR and mini-XLR. Which is more popular and are all the pin-outs the same? I would think the guys running around Can Jam would want some sort of standard so they can try different gear?
hirose would be my choice for middle of the road and a few amps and sources are using it, please oh please dont use that stupid protector connector, it should stay on cameras where it belongs. mini XLR hasnt really caught on, possibly due to not having a cable mounted male/female inline pair that i'm aware of for adapters
as i said in that post, lemo rules, but for a price and nobody except me is using them for headphones AFAIK hehe, so mass appeal isnt quite there
as i said in that post, lemo rules, but for a price and nobody except me is using them for headphones AFAIK hehe, so mass appeal isnt quite there
Last edited:
no, what we have is a mass of different half assed solutions, i have been campaigning quite vocally for a standard connector that was a practical size for years. full size 4 pin xlr is becoming more popular for home, but its still pretty rare, as far as portables there is like 5 different proprietary connectors in use by different companies with everyone refusing to enter into dialogue on a standard, hirose HR10 being the best of them imo (but lemo is better) and at least more than one company is using it, though its mostly in asia its used.
******* stupid, i gave up and made my own lemo 'standard'
my RCA comment was in reply to a post on the previous page
******* stupid, i gave up and made my own lemo 'standard'
my RCA comment was in reply to a post on the previous page
Last edited:
RocketScientist - good thoughts. I'm just a reliability freak and always think years down the road. 
I assume you are referring to the max output voltage swing vs. output current graph in the 4556 data sheet about active current limiting in the 4556. I don't see any text mention of that in the sheet, but I may be missing it. Just from past experiences it seems like about 1/2 the time active current limiting actually has to *used* on a repetitive basis in anything that is the beginning of the end. But that's a design problem of course. The 4556 is probably fine, since as you mention it has been used in headamps already.
Well a big a statistical test is about to happen that will confirm it for certain, with 200+ pcbs about to be out in the wild. It is no big deal. The chip is in a socket and you've provided a place for additional series resistance if it ever becomes and issue for someone.
I claim to know exactly zero about balanced systems and parts. I think qusp has 1000x my experience on that one.
I'll go crawl back under my rock.
I assume you are referring to the max output voltage swing vs. output current graph in the 4556 data sheet about active current limiting in the 4556. I don't see any text mention of that in the sheet, but I may be missing it. Just from past experiences it seems like about 1/2 the time active current limiting actually has to *used* on a repetitive basis in anything that is the beginning of the end. But that's a design problem of course. The 4556 is probably fine, since as you mention it has been used in headamps already.
Well a big a statistical test is about to happen that will confirm it for certain, with 200+ pcbs about to be out in the wild.
It is no big deal. The chip is in a socket and you've provided a place for additional series resistance if it ever becomes and issue for someone.I claim to know exactly zero about balanced systems and parts.
I think qusp has 1000x my experience on that one.I'll go crawl back under my rock.
Last edited:
Yeah, but why on earth did they opt for the 6 pin rather than 4 pin?
I see so many retarded decisions being made in the headphone market.
se
Thanks all for the added info on connectors. I guess I shouldn't be surprised by the lack of a clear standard. But I would think both the amp manufactures, and the ones selling headphone cable upgrades, could make more money if the round peg of one went into a compatible round hole of the other (to borrow from qusp's tagline).
So, just to be clear, there isn't even an obvious commercial 4 pin standard? Or perhaps there is one but many ignore it?
@steve perhaps that's a product niche you can fill--high quality 4 pin adapter harnesses? Or do such things already exist somewhere? I have to admit I know next to nothing about the world of balanced headphones.
@agdr, I agree the O2 is going to have a much bigger initial trial-by-fire than I anticipated. I may have to be the one who crawls under a rock! As for current limiting, I ran other tests, but this graph, if you do the math for the 15 and 33 ohm results, clearly shows it in action:
So, just to be clear, there isn't even an obvious commercial 4 pin standard? Or perhaps there is one but many ignore it?
@steve perhaps that's a product niche you can fill--high quality 4 pin adapter harnesses? Or do such things already exist somewhere? I have to admit I know next to nothing about the world of balanced headphones.
@agdr, I agree the O2 is going to have a much bigger initial trial-by-fire than I anticipated. I may have to be the one who crawls under a rock! As for current limiting, I ran other tests, but this graph, if you do the math for the 15 and 33 ohm results, clearly shows it in action:
Actually I already provide adapter cables for those who purchase my company's headphone cables. I'm not so sure I want to separate the two though. And there are already other companies out there who do make and sell adapter cables separately.
se
It's hardly exciting, some geeky updates...
I've finished updating the O2 Measurements where applicable with the V1.1 latest board, at the new gains. About a third of them changed enough to warrant updating and the other two thirds were so close (within normal measurement tolerances) I didn't bother. I also found (as usual) and fixed several typos, formatting issues, etc. along the way.
I've switched to the industry standard dBv for noise measurements. Someone emailed me and accused me of faking the noise measurements by using a different reference. By switching the dScope to dBv there is no user adjustable reference so the numbers can't be doctored. And, given all my critics, I'd have to be really foolish to fake any of the measurements as I'm sure the O2 will be measured by others.
The unit dBv is always referenced to 1 volt RMS. The reason I used to use dBr referenced to 400 mV is some portable devices can't manage 1 volt. But it's easy to convert between the two. The details are in the article.
Bottom Line: Most things are slightly better at the lower gain and I think only one measurement, the wideband THD at 1 Khz at 15 ohms, was barely worse.
I've finished updating the O2 Measurements where applicable with the V1.1 latest board, at the new gains. About a third of them changed enough to warrant updating and the other two thirds were so close (within normal measurement tolerances) I didn't bother. I also found (as usual) and fixed several typos, formatting issues, etc. along the way.
I've switched to the industry standard dBv for noise measurements. Someone emailed me and accused me of faking the noise measurements by using a different reference. By switching the dScope to dBv there is no user adjustable reference so the numbers can't be doctored. And, given all my critics, I'd have to be really foolish to fake any of the measurements as I'm sure the O2 will be measured by others.
The unit dBv is always referenced to 1 volt RMS. The reason I used to use dBr referenced to 400 mV is some portable devices can't manage 1 volt. But it's easy to convert between the two. The details are in the article.
Bottom Line: Most things are slightly better at the lower gain and I think only one measurement, the wideband THD at 1 Khz at 15 ohms, was barely worse.
Our standard adapter cables are all multi-pin to TRS, but we'll also do multi-pin to multi-pin on special order.
se
actually theres probably more money in it while people need to buy a cable or adapter for each device they own. for full disclosure i also run a high end cable company and i also offer adapters to go with my cables including a swiss army knife set to cope with multiple standards and using the lemo for the interface (female inline on headphone side so theres only one, as its considerably more expensive than the male), but I like Steve have been pushing for a sane standard. 4 pin XLR is becoming the defacto standard slowly taking over from 2 x 3 pin connectors, slowly, but its of no use for portable
the existence and market for such a swiss army set of adapters says it all
Last edited:
yeah Steve (regarding 6 pin hirose) i wondered about that too, they also make a portable balanced dac to match the amp, perhaps they needed to either share a reference of some sort, or allow for a shield and wanted to just use the one connector for both headphones and interconnect. i'm prepared to forgive them (ibasso) for that given its still the best realistic choice of the lot, i prefer lemo 0S size push/pull locking connector, but its a bit pricey, though they could have used the polymer housing versions which are about half the price and compared to audiophile connector and equipment prices its nothing, the dielectric is excellent if you care for that, its reasonably easy to solder, sexy as hell and comes in cable, panel, pcb mount in male and female. basically it ticks boxes on all sides.
retarded choices indeed
retarded choices indeed
Last edited:
Could be.
I'd just prefer a 4 pin as a cable maker as it makes it a bit easier to solder four leads to a 4 pin connector than a 6 pin connector.
Oh, you mean the Redels?
se
Thanks for the added info. I fear a full size XLR isn't likely to find a home on the PCB of even the desktop board I'm working on. But a mini-XLR might. The least expensive PCB option at Mouser is still relatively expensive (as in 3 of them cost more than all the other 80+ parts on the board). Would the (don't hate me) $8 Switchcraft TRAPC4MS1X work if I can find room on the PCB? Anyone would obviously still be free to use a panel mounted anything. And if it will work, what's the pinout?
On the other hand, if I can fit a full size XLR, they're half the price or less from Neutrik.
On the other hand, if I can fit a full size XLR, they're half the price or less from Neutrik.
Last edited:
nah i think your best option is just to have a 2.54mm pinheader/strip of PTH in parallel to whatever connector you supply as pcb mount phone connector, so that its easy for people to choose what they want
the rean/neutrik 4 pin panel gold plate are slightly cheaper at mouser than the tin plated brass one you linked and are more common, these are becoming popular on the LCD2 cables (at the headphone end) as the cable connectors are pretty reasonably priced (much cheaper than the switchcraft) and decent quality. but there is no male cable mount, only male panel mount and female cable mount.
yeah Steve the redels, but i just checked mouser and they arent as cheap as i thought, only ~ 10 dollars less for size 1 than my favored alloy version size 0S (19aud vs 29 for 4 pin male cable mount and 22 vs 29 for the panel mount) and only available in shell size 1 so its not a direct comparison. 1 is probably a decent size, i prefer the next one down, size 0S but possibly You would have a little trouble with your cotton sleeved braid fitting? although if you can fit it in a mini your all good, but certainly some manufacturers might struggle. so perhaps size 1 is a good compromise and could be used for home and portable systems, although the alloy size 1 is another 10-15 dollar jump from Lemo 0S.
the HR10 looks better for price and flexibility really, only the alloy lemo have all the options as far as gender and mounting options. of course buying in bulk from the manufacturer would lower the price considerably, i even went as far as entering dialogue with Lemo about that, but our silly community (mostly manufacturers) wont agree or even enter into a dialogue on the matter
the rean/neutrik 4 pin panel gold plate are slightly cheaper at mouser than the tin plated brass one you linked and are more common, these are becoming popular on the LCD2 cables (at the headphone end) as the cable connectors are pretty reasonably priced (much cheaper than the switchcraft) and decent quality. but there is no male cable mount, only male panel mount and female cable mount.
yeah Steve the redels, but i just checked mouser and they arent as cheap as i thought, only ~ 10 dollars less for size 1 than my favored alloy version size 0S (19aud vs 29 for 4 pin male cable mount and 22 vs 29 for the panel mount) and only available in shell size 1 so its not a direct comparison. 1 is probably a decent size, i prefer the next one down, size 0S but possibly You would have a little trouble with your cotton sleeved braid fitting? although if you can fit it in a mini your all good, but certainly some manufacturers might struggle. so perhaps size 1 is a good compromise and could be used for home and portable systems, although the alloy size 1 is another 10-15 dollar jump from Lemo 0S.
the HR10 looks better for price and flexibility really, only the alloy lemo have all the options as far as gender and mounting options. of course buying in bulk from the manufacturer would lower the price considerably, i even went as far as entering dialogue with Lemo about that, but our silly community (mostly manufacturers) wont agree or even enter into a dialogue on the matter
Last edited: