sek: surely you jest?
the whole basis of your argument is flawed and the goal incomprehensible; did you have a brain explosion?
I take it it's your time of the month again, Jeremy.
We've been through this before.
I would say the same thing, yes it is familiar; weve been in this position where you argue with me pointlessly before.... nope nothing to do with the time of month, i'm just well sick of you doing this, picking something out of one of my posts to argue with. you even said as much when you said 'I cant quite let you get away with this' get away with what? get away with recommending against madness? (Sorry Avro Arrow yours was just a harmless but somewhat misguided suggestion and we havent heard your reasoning yet, so the vitriol isnt directed at you at all)
this time however Sebastian i'm totally at a loss; normally although our methodologies are totally different and i find your solutions unnecessarily complicated, i can respect the engineering and thought behind it and just remember there is different strokes for different folks. this however is totally baffling, there is not even a silly reason to have SMD regulators, or mount the buffers upright, no reason at all and no benefit at all. Doing this would result in a whole lot of extra work and some extra cost for no pay off whatsoever. In the case of the regulator it would render it unusable for the builds it was designed for and make it larger; not smaller. The loads it was designed for include swinging up to +/-100vdc for the power amp input stage and at start up this needs to cope with a burst of over 250ma! i dont think this would work even with a large and more expensive multilayer 2oz(or higher) copper pcb with thermal modelling needed to design it with the SMD regulators unless you added a heatsink that was at least as big as the upright version, because there is a larger thermal impedance to get heat out of the PCB.
then we have the mounting the buffers upright.... again far more work to build, at least as expensive, larger and you have to either glue or rig some way to clamp it to a sink. all of this for what? what benefit do you propose? there is no issue with how it is now, the engineers at National I presume did some testing too and it sure is convenient, only in very hard loads and hot climate will you even even need a heatsink at all and the buffers are already at the limits of their power output so improving dissipation will not allow you to get more power
so yes i'm annoyed, i'm annoyed at you arguing with me pointlessly again, there is a special word for people that do that on the internet.
I use SMD D2PAK regs extensively, i love them, but they are definitely power limited by their dissipation.
anyway lets stop this NOW, i kept my initial post short for a reason.
this time however Sebastian i'm totally at a loss; normally although our methodologies are totally different and i find your solutions unnecessarily complicated, i can respect the engineering and thought behind it and just remember there is different strokes for different folks. this however is totally baffling, there is not even a silly reason to have SMD regulators, or mount the buffers upright, no reason at all and no benefit at all. Doing this would result in a whole lot of extra work and some extra cost for no pay off whatsoever. In the case of the regulator it would render it unusable for the builds it was designed for and make it larger; not smaller. The loads it was designed for include swinging up to +/-100vdc for the power amp input stage and at start up this needs to cope with a burst of over 250ma! i dont think this would work even with a large and more expensive multilayer 2oz(or higher) copper pcb with thermal modelling needed to design it with the SMD regulators unless you added a heatsink that was at least as big as the upright version, because there is a larger thermal impedance to get heat out of the PCB.
then we have the mounting the buffers upright.... again far more work to build, at least as expensive, larger and you have to either glue or rig some way to clamp it to a sink. all of this for what? what benefit do you propose? there is no issue with how it is now, the engineers at National I presume did some testing too and it sure is convenient, only in very hard loads and hot climate will you even even need a heatsink at all and the buffers are already at the limits of their power output so improving dissipation will not allow you to get more power
so yes i'm annoyed, i'm annoyed at you arguing with me pointlessly again, there is a special word for people that do that on the internet.
I use SMD D2PAK regs extensively, i love them, but they are definitely power limited by their dissipation.
anyway lets stop this NOW, i kept my initial post short for a reason.
Last edited:
I think the confusion has come from not understanding the context of the posts a while back when discussion was about the custom heatsink design.
hypertune has access to a CNC so can create whatever heatsink he so desires, this allowed him to suggest a heatsink of similar size to the half brick heatsink that opc used in the build on page one but with a step in the base of the heatsink to allow it to directly couple to the PCB and buffer casings, this was to give an improvement over coupling only to a buffer case that does not have particularly good thermal conductivity and using an extra thick thermal pad to conduct heat from PCB to the half brick heat sink.
I understand what Avro is saying, heat through the thermal pad on the buffer is much better than through the case. The suggest heatsink design that was discussed though aims to take the heat from pcb more efficiently rather than using a complicated mounting method, since the heatsink is probably only absolutely essential when high ambient temperatures or extended high load on the amplifier this seems like the most reasonable solution.
hypertune has access to a CNC so can create whatever heatsink he so desires, this allowed him to suggest a heatsink of similar size to the half brick heatsink that opc used in the build on page one but with a step in the base of the heatsink to allow it to directly couple to the PCB and buffer casings, this was to give an improvement over coupling only to a buffer case that does not have particularly good thermal conductivity and using an extra thick thermal pad to conduct heat from PCB to the half brick heat sink.
I understand what Avro is saying, heat through the thermal pad on the buffer is much better than through the case. The suggest heatsink design that was discussed though aims to take the heat from pcb more efficiently rather than using a complicated mounting method, since the heatsink is probably only absolutely essential when high ambient temperatures or extended high load on the amplifier this seems like the most reasonable solution.
Last edited:
Jeremy, I find it surprising how we seem to be on such a difficult level of understanding. For the life of me, I can't see how my above post can provoke a reaktion from someone I have such a distant relation to. It never happens with anyone else but you. I find that revealing.
Remember that this is an opinion, not a a fact. And bad choice of wording on top.
The BUF634, a similar device, is in very wide use in it's TO-220 case. I'm not arguing against surface mounting or advocate the use of BUF634 here, I merely provide a form of counterexample against your somewhat flawed argument that upright heat sink mounting has no demand or justification with these buffers.
Again, an opinion, not a fact.
There's also a widespread judgment about those who feed such people on the internet, but that is not to imply either of this applies to any of us.
Thanks for allowing me to respond to an insult that I found unjustified...
Last edited:
I was referring to PCB and heatsink considerations in general, thus referring to both of the projects (and trying to consider the difference between the two).
But you're of course right in that the current lpuhp PCB would not sustain the thermal demands of SMD regulators.
Yes, I got that, and I find it a perfectly viable option in case the demand for such an increase in cooling is there. It just depends on how much you have to bend...
BTW, an alternative to bending the pins into an appropriate position is to bring the pads to the pins (as it were) by extending them with something appropriate (and short). People have wired transistors towards heat sinks all the time. Not as beautiful, but certainly eliminates the risk of bent pins breaking under vibration during use.
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!
But it would touch the buffers where they really need the dissipation...
You've lost me sek ... the pcb and thermal dissipation by the present pcb + heatsink is sufficient for the buffers as calculated and confirmed by opc's testing (IanAS were you really asking if opc had sent boards out without knowing the heatsinking arrangement was suitable?)
Basically I still don't understand what advantage is to be gained by adding complexity when the existing arrangement is known to be OK.
The design opc has provided gives the most compact arrangement possible.
This isn't going anywhere though ... pcbs are already in the hands of those who will build, I have already ordered and received over 70 buffers for the Australian/NZ lpuhp builders with a few of these buffers being used on various other headphone amps and a few spares just for myself for good measure and experimentation with low current low noise regulators. I totally understand the arrangement you're suggesting but see no situation where I'd ever consider utilising such an arrangement.
Basically I still don't understand what advantage is to be gained by adding complexity when the existing arrangement is known to be OK.
The design opc has provided gives the most compact arrangement possible.
This isn't going anywhere though ... pcbs are already in the hands of those who will build, I have already ordered and received over 70 buffers for the Australian/NZ lpuhp builders with a few of these buffers being used on various other headphone amps and a few spares just for myself for good measure and experimentation with low current low noise regulators. I totally understand the arrangement you're suggesting but see no situation where I'd ever consider utilising such an arrangement.
hahaha what is this if not arguing pointlessly? trolling was the word i was searching for...or would you prefer the alternative that means you are just a bit slow on the uptake?
The buffers are at FULL POWER as it is!!! and it HAS NOTHING TO DO WITH THE HEATSINK it has to do with low impedance, limited swing (limited by the LME49990 not the buffer) and current limiting.
the current limiting will cut in into such low impedance SPEAKER LOADS before they burn up. bringing other devices into it is meaningless, we are talking about the current design with the current parts and the current loads on the current PCB; which is perfectly OK, making some of the best measurements i've ever seen
you are hilarious, welcome to ignore. I have resisted doing that for a while as occasionally you have something interesting to offer, but now you have tipped the scales so that my frustration far outweighs that and these arguments all center on this needless and thankless complication you seem to love.
I have better things to do with my time; one of which includes building this amp..as is
Last edited:
yep it sure is!
Hey Ian, nah there really is no basis for the argument at all, for starters sek isnt even in the GB and neither is Avro Arrow, the amp was not designed to be used for a high power amp and nope nobody has raised any concerns about the power output, only 3 people including opc have built it so far and none have raised any issues about it. the heatsink as i said is really only a safety measure for hot climates and low impedance speakers (at the same time)
it was designed to be used in multiamped systems within its limits, which have nothing to do with the heatsinking. as i said above the power is limited by the swing on the input opamp due to limited rail voltage on the LME49990 and the current limiting on the buffer, which has nothing to do with the heatsink either.
hey man look its all good like i said in my note. increasing dissipation capability beyond what already exists provides no benefit, as its not the limiting factor. the conversation about them earlier in the thread was just for peace of mind due to my living in a hot climate without constant aircon; we were the only ones raising this concern. weve 'solved' it (me and hochopeper) by arranging to get fully CNC billet enclosures to act as a heatsink by contact with the PCB, as well as the buffer bodies, as i'm only driving tweeters (same with hochopeper I think) my concern is probably unwarranted anyway. the rest of the GB are getting custom CNC heatsinks (but not full enclosures) as well that also contact the PCB AND Buffers, so its a total non issue that was 'solved' many moons ago
the thermal pad of the buffer is already soldered to the heatsink, the additional heatsink in most cases could be left off entirely, but in the interests of overkill I took it a step further and i'm much happier with the outcome of the enclosure than ruining a clean design for dubious benefit.
another thing that hasnt been mentioned, the tabs on the buffer are LIVE, connected to the negative rail, so any other heatsink would have to be isolated
anyway since this is all of no value, only lip service and the parts are all about to be on their way to everyone, how about we save this thread for build queries hey?
Last edited:
I'm not saying that, this thread appears to be saying that there's been a bit of an issue over some additional dissipation requirement resulting in a subsequent clever custom design. I recall a few posts about searching for a suitable method to couple the component side of the board to the sink.
My limited experience is that you need identical amps to drive tweeter and bass. You hear the sound of a particular amp and the sound of another amp. At the crossover region you hear the two separate characters and in my experience they didn't get along. I and friends and probably cases I read in forums said the same, it sounded better with one amp than two different ones. I first tried this in about 1981 (yes 31 years ago!) so it's not new. Then is was an analogue active crossover and, IIRC, a Quad 405 mk2 and a Crimson Electric. The Crimson was better for either and both bass or treble. That's, of course, not to say it can't ever work if things happen, or are carefully selected, to synergise
I was watching this thread as I still consider adding another 4 chips to the BAL-BAL headphone driver, just to see if my results differ from Owens of too much noise and no benefit. The BAL-BAL is thin sounding with the HD800 and I desire to boost the bass a bit (without adding a tone control
).
Last edited:
Hi Guys,
I haven't read through all the above, but I think I get the gist of it.
Overall, it was a pretty simple decision. I did indeed consider upright mounting, and even sacrificed a few LME49600's trying to bend the leads into an offset 5 pin pattern so I could make a through-hole footprint and mount them onto the same heatsinks used in the PSU.
The end result would have looked like a bit of a hack job, and trust me, mounting 16 of these IC's onto heatsinks with clips and bolts, bending the leads to fit, and soldering everything in place would have been a very obnoxious task. Ask qusp what he thought of mounting the LME49830 on the heatsinks I chose for the power amp. That was easy compared to this!
I wanted a solution that didn't have to be custom, and mounting the IC's upright on bars would have required everyone to purchase or fabricate a custom bar, and most people don't want to do that.
Lastly, having super-effective heatsinking really isn't necessary for this amp. Like I mentioned before, I'm still running mine all day, every day on my PC, and they're just fine sitting at about 50-65C depending on use. You could probably solder a few pieces of copper sheet on each tab and call it a day.
The heatsink Hypertune has offered to machine is a perfect solution overall. It's not too difficult to machine, it fits perfectly with a lot of contact area, and will provide enough cooling to run these at full output for extended periods.
Looking back, the ideal way to heatsink these is to run filled micro-vias all through the pads to the back side of the board, and use a single TIM pad and a flat plate to heatsink directly to the bottom. The thermal resistance of the filled vias is very low, and would basically provide direct access to the tab of the chip, allowing for very effective heatsinking. If you look at SPMS designs where these packages are used as output rectifiers or fets, that's exactly how they do it.
Cheers,
Owen
I haven't read through all the above, but I think I get the gist of it.
Overall, it was a pretty simple decision. I did indeed consider upright mounting, and even sacrificed a few LME49600's trying to bend the leads into an offset 5 pin pattern so I could make a through-hole footprint and mount them onto the same heatsinks used in the PSU.
The end result would have looked like a bit of a hack job, and trust me, mounting 16 of these IC's onto heatsinks with clips and bolts, bending the leads to fit, and soldering everything in place would have been a very obnoxious task. Ask qusp what he thought of mounting the LME49830 on the heatsinks I chose for the power amp. That was easy compared to this!
I wanted a solution that didn't have to be custom, and mounting the IC's upright on bars would have required everyone to purchase or fabricate a custom bar, and most people don't want to do that.
Lastly, having super-effective heatsinking really isn't necessary for this amp. Like I mentioned before, I'm still running mine all day, every day on my PC, and they're just fine sitting at about 50-65C depending on use. You could probably solder a few pieces of copper sheet on each tab and call it a day.
The heatsink Hypertune has offered to machine is a perfect solution overall. It's not too difficult to machine, it fits perfectly with a lot of contact area, and will provide enough cooling to run these at full output for extended periods.
Looking back, the ideal way to heatsink these is to run filled micro-vias all through the pads to the back side of the board, and use a single TIM pad and a flat plate to heatsink directly to the bottom. The thermal resistance of the filled vias is very low, and would basically provide direct access to the tab of the chip, allowing for very effective heatsinking. If you look at SPMS designs where these packages are used as output rectifiers or fets, that's exactly how they do it.
Cheers,
Owen
@ Ian regarding using different amps in active XO
when the 2 amps are the wire and the bigger wire and the crossover corrects based on measurements I doubt there will be any of this effect. ive already got another 4 of the lateral fet amp on top of the 2 channels i plan to use with the LPUHP initially and i'll see how it goes. I had considered this possibility, but i'm not fully invested in using the LPUHP here, i might use it for a full range center speaker or something for the studio so i'm not overly concerned. also ATC would seem to disagree with you; some of their gear uses identical amps, but their flagship and other top line active studio monitors do not. the ATC active monitors are the best active speakers ive ever heard and many seem to agree
oh sure, that what I meant about doing some thermal modelling for thermal impedance solutions, it may have resulted in a solution that got the heat out quicker, but what benefit apart from an intellectual one would there have been since there are no issues that mean the current solution doesnt keep up with the needs of the amp?
when the 2 amps are the wire and the bigger wire and the crossover corrects based on measurements I doubt there will be any of this effect. ive already got another 4 of the lateral fet amp on top of the 2 channels i plan to use with the LPUHP initially and i'll see how it goes. I had considered this possibility, but i'm not fully invested in using the LPUHP here, i might use it for a full range center speaker or something for the studio so i'm not overly concerned. also ATC would seem to disagree with you; some of their gear uses identical amps, but their flagship and other top line active studio monitors do not. the ATC active monitors are the best active speakers ive ever heard and many seem to agree
opc said:
oh sure, that what I meant about doing some thermal modelling for thermal impedance solutions, it may have resulted in a solution that got the heat out quicker, but what benefit apart from an intellectual one would there have been since there are no issues that mean the current solution doesnt keep up with the needs of the amp?
Last edited:
P.S. - I think we should organize a large international DIY meet where everyone can meet everyone else in person. You guys can be pretty cut-throat with one another, and I find that goes away when you meet someone face to face.
I think in a fundamentally different way when I write to the Ottawa group compared to anyone else. There are a lot of subtle cues that you get when you're in the same room as someone, and they're lost here, making people hostile and short with one another for absolutely no reason. I know what all the Ottawa folks look like, I know their humor, attitudes and disposition, and it helps immensely in communicating.
Remember what you mother told you (or should have told you):
"if you don't have anything nice to say, don't say anything at all"
Cheers,
Owen
I think in a fundamentally different way when I write to the Ottawa group compared to anyone else. There are a lot of subtle cues that you get when you're in the same room as someone, and they're lost here, making people hostile and short with one another for absolutely no reason. I know what all the Ottawa folks look like, I know their humor, attitudes and disposition, and it helps immensely in communicating.
Remember what you mother told you (or should have told you):
"if you don't have anything nice to say, don't say anything at all"
Cheers,
Owen
Free plane tickets for everyone!
I would imagine that would be a bit of a hike for you guys on the other side of the planet, even if the flight were provided.
It also limits the number of massive amplifiers you can bring along. I once flew across Ontario with a Dynaco ST-35 clone packed in my carry-on, and things didn't go so well at airport security. Point to point wiring looks pretty creepy in an X-ray machine, and trying to explain that it's a home-made 1950's tube amp doesn't help either
Cheers,
Owen
I would imagine that would be a bit of a hike for you guys on the other side of the planet, even if the flight were provided.
It also limits the number of massive amplifiers you can bring along. I once flew across Ontario with a Dynaco ST-35 clone packed in my carry-on, and things didn't go so well at airport security. Point to point wiring looks pretty creepy in an X-ray machine, and trying to explain that it's a home-made 1950's tube amp doesn't help either
Cheers,
Owen
I already tried to cut through all of this with sek a while ago and approached him by PM, apologized, explained myself and what was getting to me and tried to move forward, he denied being provocative just as he did here by trying to make me think it was all a misunderstanding or something, but i just decided to drop it, take a deep breath and move forward, but here he is again, doing it again and honestly expecting me and you all to believe that its all innocent and i've just got some bad PMT going on; which is provocative in itself for christ sake
so no sorry; while i agree meets are great for that, I doubt i would even say hello, after today i would go out of my way to avoid him
Ah but they've spent decades experimenting, honing and refining, including making their own drive units, to achieve that synergy
Ah but they've spent decades experimenting, honing and refining, including making their own drive units, to achieve that synergy
that they have =) I knew even if you were some freak of nature that didnt like ATC monitors (SMC-150 dome mids... drool) the patriot in you would have to respect them.
whos to say this isnt the beginning of that journey with me? in fact it kinda is and these are all just tools to that end; luckily having multichannel digital XO enables much more rapid prototyping than building different amps
you guys saw that right? opc just offered plane trips to everyone! i'll bring some sunshine in a bottle and cascade pale ale
I need the bottled sunshine you see, because it sure doesnt come from my demeanor
lol
Last edited:
I'm sorry if I came across as asserting a flaw in Owen's design. I was not, instead I was commenting on what I considered a contradiction in the chain of argumentation. It's interesting how the idea of upright heat sinking got attached to me. I did not bring it up, nor did I justify it. I merely discussed it's viability after someone else showed demand.
Yay, Canada!
I wouldn't assume a professionally designed DIY SMD job to be unsuspicious, either!
That's also my experience. In the luckily very rare cases - like this one - there's an easy fix (see attachment).
The only other case I can remember was the good old Carlos Filipe, who later got banned from the forums for his attitude against a growing number of members.
Cheers,
Sebastian.
So it has come to this...
Attachments
In that case I'm even more lost.
Avro Arrow suggested it, I guess after a misunderstanding of the original heatsink discussion thinking that there was a thermal dissipation issue with the design.
It has now been pointed out (repeatedly) that there is no thermal issue to be solved and that the added complexity of mounting vertical is unnecessary, yet, you still seem fixated that vertical mounting is actually a good idea? I was not attaching the idea of upright mounting to you, you had already suggested it as still being viable, I and others (qusp and opc mainly) were just trying to point out that it is a needless complexity. Yet you continue to seem convinced otherwise. Even after opc has described trialling almost exactly this solution and deciding against it. IMHO about now is when you should be posting to say that you were wrong this time around ...
Yay, Canada!
At least there is one thing we can all agree on!
Though I hope no one judges us Aussies on qusp's choice of drink, seriously Cascade mate?