If I understand these buffers correctly they need to be in a fb loop with the input or gain opamp (like the buff634's) for dc coupled output.
I was listening to my TOL Asus soundcard analog out into the O2 to get a taste of S-D (Ak4396's), I could literally hear the jitter shifting like sand between the channels, exactly like the famous 2+nS jitter sound sample have, quite dissapointing.
So I guess I should build this software controlled wolfson kit first then, hopefully the wm8805 will help matters or I'll I will find a good hirez asynch usb.
Am I the only one building this for headphone only duty (probably 1 or 2 buffers per channel.) It seems it should offer exceptional ss performance for headphones. But I may splurge for the full buffer set to see how it does with my fostex 127E's, the designers description that it reminded him of an SET amp without the drawbacks cought my attention.
There is a decent chance I will be using them with headphones. My speakers are not very efficient, 87 dB, so I don't know how loud they will get. My space is quite limited as well so I won't be buying/building a new pair of speakers soon. I don't listen very loud and the room isn't huge, so the 16 watts may still be sufficient in which case I will use the amp with the speakers. Otherwise they will be powering headphones until my speaker situation changes.
nope, like i said in my reply to your post, they can be used open loop, in fact opc tried it both ways in the original BAL-SE wire headamp and found not a lot of difference in performance. the offset issues with the buffer are down to bad layout/implimentation, not the chip. in fact the reference design had issues with offset in the buffer due to the servo, not the buffer, but the buffer had a bad name for offset because of it for a while.
but and this is a big but, this was just open loop from the input VAS to the buffers, not at the end of a long cable and a couple of connectors. it might be nothing to worry about, but it would certainly be something i checked for oscillation
maybe, wolfsin has that in mind too I think, i'll be playing with it like that for ***** and giggles, but its aimed squarely at my tweeters. using it with only 1 or possibly 2 buffers you have basically just build the bal-se headamp, so there wouldnt be many no, perhaps a couple who missed the headamp GB
Last edited:
wolfsin izIn HogHeaven: too many toyz
He has a wonderful pair of heatsinks that would call for mounting the pcbs vertically so is considering a range of alternatives. Being able to mount varying buffer counts is very attractive for driving headphones even 'hungrier' than Senns.
there might be a couple, we are still waiting on custom trannies, heatsinks etc and about to put the passives GB through this weekend. i've worked out shipping etc wanted to confirm with DK i was all covered for duty etc and just when i thought i had a lul, anyway the passives will be here before trannies and sinks are ready so no delay. keen to get going, not sure how much additional benefit you'll get with 20 buffers though TBH, will post/PM/email later this afternoon.
AAK, I think this post from opc here might be of interest to you. Just wondering why you're going for more devices? opc stopped at 8/ch for a reason ... I don't mean to discourage, just would hate to see you unnecessarily pouring money into buffers if there is no advantage.
I think a lot of people building in the near future will be waiting on GB.
jazzm and opc are the only people I know of with them built, I haven't seen any comments from jazzm after he finished the 2nd channel either ...
I think a lot of people building in the near future will be waiting on GB.
jazzm and opc are the only people I know of with them built, I haven't seen any comments from jazzm after he finished the 2nd channel either ...
erm yeah after the PCBs are done and in people's hands along with some already built; is a little late yes.....
how would make it more convenient? non-standard mounting of the buffers probably involving lead bending + plus 8 heatsinks per channel or minimum 2 heatsinks and epoxy or clamp each package and a taller profile vs a single top mounted 4 dollar heatsink on top that just screws down....hmmmm doesnt seem like an improvement to me, particularly since it really isnt needed at all in many cases as they use the PCB copper pour for heatsinking, the option for which is lost with your scenario. I certainly would have tried to avoid that scenario even if it wasnt too late. aside from the performance the ability to use the PCB and no heatsink is one of the main advantages of this device
how would make it more convenient? non-standard mounting of the buffers probably involving lead bending + plus 8 heatsinks per channel or minimum 2 heatsinks and epoxy or clamp each package and a taller profile vs a single top mounted 4 dollar heatsink on top that just screws down....hmmmm doesnt seem like an improvement to me, particularly since it really isnt needed at all in many cases as they use the PCB copper pour for heatsinking, the option for which is lost with your scenario. I certainly would have tried to avoid that scenario even if it wasnt too late. aside from the performance the ability to use the PCB and no heatsink is one of the main advantages of this device
Last edited:
I can't quite let you get away with your arguments against upright mounting, as the same could be expressed about the PSU PCBs (where TO-263 voltage regulators would have been perfectly appropriate), but:
That is indeed true, it's very convenient to just have the devices soldered properly and be done.
Cheers,
Sebastian.
That is indeed true, it's very convenient to just have the devices soldered properly and be done.
Cheers,
Sebastian.
The problem is with the PSU PCB (assuming you're talking about the headphone amp psu pcb and not the psu integrated on the lpuhp pcb), the PCB size limits the copper pour and hence thermal dissipation. 2 regs on a tiny pcb ... things are gunna get toasty! In the suggestion from Avro Arrow we're actually bending legs on a perfectly good buffer and rigging some other heatsink to it. It certainly won't be as tidy as the design opc produced. Nor would it be as easy to make a custom CNC enclosure to act as heatsink!
Those PSU pcb as they are now can take a couple of amps with a few simple mods, certainly wouldn't be doing that with a smt reg on a tiny pcb and no heatsink.
Though I'm sure if I take just this quote out of context qusp may wish to re-word his original post.
Anyways, Avro, what were you thinking would be the advantage of vertical mounting the LME49600 buffers?
Those PSU pcb as they are now can take a couple of amps with a few simple mods, certainly wouldn't be doing that with a smt reg on a tiny pcb and no heatsink.
Though I'm sure if I take just this quote out of context qusp may wish to re-word his original post.
Anyways, Avro, what were you thinking would be the advantage of vertical mounting the LME49600 buffers?
Last edited:
This is really a spurious and wasteful discussion. Forethought and planning. Simply work out how much heat dissipation is required (or build a prototype and measure it). Chose the heat sink accordingly. If the PCB is inadequate then glue or clamp the row of devices to a suitable sink, standing them if that's easiest. Maybe solder them to one or two thin copper sheets then clamp that to a larger piece.
I've not bought into this unit and have not been paying attention so I'm just a spectator here but it's seems that some people might want to use it a bit beyond the original design plan which was without additional sinks? Those people might perhaps have been better do it in a simple and different way rather that trying to create a cleverly designed intricate additional sink?
Done though now isn't it?
I've not bought into this unit and have not been paying attention so I'm just a spectator here but it's seems that some people might want to use it a bit beyond the original design plan which was without additional sinks? Those people might perhaps have been better do it in a simple and different way rather that trying to create a cleverly designed intricate additional sink?
Done though now isn't it?
Last edited:
Maybe my eyes have gotten bad or maybe I have just forgotten how to read, but
I was sure there had been much discussion in this thread about how to add a heat
sink to this design.
My idea has only one advantage...it puts the thermal pad of the buffer in direct
contact with a heat sink. Other than that it's a pain in the butt to do.
The vertical space is already being used by the power supply part of the board so
it is of no real disadvantage to have another vertical heat sink. It is just one idea
among many. Like it or don't. I'm not here to argue.
I was sure there had been much discussion in this thread about how to add a heat
sink to this design.
My idea has only one advantage...it puts the thermal pad of the buffer in direct
contact with a heat sink. Other than that it's a pain in the butt to do.
The vertical space is already being used by the power supply part of the board so
it is of no real disadvantage to have another vertical heat sink. It is just one idea
among many. Like it or don't. I'm not here to argue.