Bronze heatsinks, split from advanced GainClone thread

[dfdye]

Originally posted by jneutron
Copper at 4.5 Kelvin has incredibly high diffusivity, for example. This is because it's thermal conductivity remains the same (relatively) as it is at room temp, but it's heat capacity drops 2 to 3 orders of magnitude..heat flies through it at liquid helium temperatures.

I'll vouch for that first hand. We run experiments in copper blocks at 1.8K (evaporative helium cooling), and the cooler they get the faster they cool. Really neat phenomenon.

Getting back to the question of the thread. . . .

The question I have for the materials guys here is what is the best material at room temp with regards to heat "diffusivity"? I know we all use aluminum and copper becuase they are cheap/available, but what is best? Where does brass rank on the scale? I know there are tables and tables of heat capacities around, but can someone point me toward heat flow tables? I know I have seen them, but I can't seem to find one now.

[jleaman]

Originally posted by Peter Daniel
Here's the link where I described my preference: http://www.diyaudio.com/forums/showt...680#post343680

The binding posts will not be rotated.

The patented posts are an option; regular posts are available as well and one can always use banana connectors.

But you cant use banana connectors with them Cardas binding post's. I do like them thus why im going to use them. They worked well on the amps that Brian had. I just wanted to know why they were oriented that way

[BrianGT]

Originally posted by jleaman



But you cant use banana connectors with them Cardas binding post's. I do like them thus why im going to use them. They worked well on the amps that Brian had. I just wanted to know why they were oriented that way

Don't worry about it They work fine with my other cables with the cardas spades on them.

--
Brian

[jleaman]

Originally posted by BrianGT


Don't worry about it They work fine with my other cables with the cardas spades on them.

--
Brian

I'm not worrying. But just wondered why they were put that way that's all.

How do you like the cables that you are using now brian ?

[jneutron]

Originally posted by jneutron
Indeed, proximity of the heatsink to the silicon die changes the system inductance as well as the coupling between the output currents and the input gain stages.

In my confusion, I didn't realize that I was asked about this statement.

For the bulk of silicon chips fabricated with N type material, the substrate, which is the bottom of the chip, is soldered or brazed directly onto a conductive surface. Some chips use oxide wells for isolation, but the bulk of the chips are fabricated with reverse biased pn junctions to isolate the active stuff from the backside.

If you look at the device before the plastic molding step, you can see the wirebonds from the pads on the chip, to the copper formed leads that end up exiting the package at the edge.

These wirebonds are loops of wire, and they have inductance. One of those wires is the output wire, and one is the V+ wire, they are both typically aluminum, and 5 to 15 mils in diameter. Also present are the wirebonds to all the signal aspects of the device. They can be gold, or aluminum. I believe they are all 1 mil aluminum, it eliminates purple plague, a purple colored alloy of gold and aluminum.

What my statement made reference to was the fact that all the wires are inductive, all form loops to the silicon, and they are all inductively coupled to one another.

If you all recall your third grade physics, you will remember that for non magnetic conductive materials (mu = 1), all eddy currents generated will attempt to prevent the magnetic flux from entering the material..(Lenz's law). This means that the better the conductor, the LESS the flux will enter, therefore, the LOWER the inductance of the system...hence, the lower the coupling between wires.

As you raise the frequency, the inductance and coupling, lower.

I've run into this problem on occasion, where the test frequency has to be specified..especially if the coil form is copper...stainless is not so bad, and aluminum is in the middle.

If you build a circuit which can be affected by the amount of coupling caused by the wirebond inductances, then obviously, changing the conductivity of the mounting platform (heatsink) will affect the circuit.

So the real question is not can it affect the circuit, but by how much, and is it audible..

Cheers, John

[poobah]

Wallowing in the minutia... Who's gonna help me design new drivers... altogether new???

[dfdye]

Originally posted by poobah
Who's gonna help me design new drivers... altogether new???

Start a new thread! I'll offer my ignorance!!!

Piezo tweeters sounded like such a great idea, too!

Now, with drivers, I agree that brass cones will make a huge difference than alum cones!

[jh6you]

Very psychical . . .

[Bobken]

Hi Peter,

I was interested to read your comments relating to bronze heatsinks here, which come as no real surrprise to me, as it happens.
What has been disappointing, though, is the subsequent rash of 'contesters' who haven't any idea of what they are talking about in this instance, simply because they have not tried out what you have done. What is it about human nature which encourages so many people (who otherwise seem to consider themselves to be intelligent) to not only consider it to be their right, but even often some kind of obligation, to 'put the world to rights' here, even though this is entirely based on just their thoughts and suppositions?

Mostly, they appear to waste theirs and other peoples' time in such juvenile ridiculing of what someone has discovered and openly shared in good faith, and I thought I had escaped from demonstrations of such intellect as shown when doing this, when I left junior school nearly 60 yrs ago. What a world we live in when someones' sport is based on antagonising others, just for fun!

Whilst I have not (yet!) found any realistic method of proving this hypothesis, I believe that this phenomenon is caused (certainly in the main) by alterations to vibrations affecting the semi-conductors attached to the heatsinks. More than a few audio designers (and those involved in more 'consequential' spheres of electronics, like life-support and space-exploration etc.) are well aware that unwanted vibration effects can badly prejudice the operation (and reliability) of any electronic circuits and components. It is simply a way of life for them, and they don't even question these effects, whether they can be satisfactorily explained by anyone, or not. For many years I have believed this vibration-damping to be the main reason for differences heard with alternative metals being used for heatsinks, rather than anything electrical or thermal, as has hitherto been suggested.

I first became aware of some similarly odd and quite unintuitive 'influences', when improving a Spectral pre-amp, over 20 yrs. ago. This was rather ahead of its time in some ways, having entire ground-planes on both sides of the full-size PCB which were also gold-plated throughout, and this appearance together with a very neat and careful component layout, caused to me as an engineer to appreciate the aesthetics of the populated PCB when the case was open. Having seen at some audio show, a similar unit with a clear acrylic top which was intended to show-off this rather better than average internal appearance, I wished to emulate this 'exposed internals' effect, so I duly made up an 8mm Perspex (Lexan) top for my own pre-amp.

However, and especially as this pre had massive gain to suit even very low-output MC cartridges without any added step-up device, I was quite concerned about the possibility of RFI, or whatever, affecting the overall sound, and probably detrimentally. Accordingly, before finally screwing down this new cover more permanently, I considered it wise to conduct some listening trials to determine this potential, one way or the other.

Somewhat to my surprise (and quite the opposite to all expectations) not only did the new cover not make the sound worse, it actually sounded marginally better than with no lid at all. Being very puzzled by then, but interested in getting to the bottom of this anomaly, I experimented further for several more days, and there was no doubt at all that the preferred order of 'sonic goodness' was: 1st. the new cover, 2nd. no cover, and last (by a more noticeable margin), the original metal top. The entire enclosure was very shallow on these Spectral pre-amps, so the proximity of the covering material was quite close to the circuit board, and almost touching the tops of the taller components.

The conclusion I reached after a couple of weeks playing around with these alternatives, was that placing substantial areas of metal near to sensitive electronic circuits was detrimental to the overall sound, and that damping effects from the dense and quite heavy acrylic top, actually benefitted the overall sound, as a bonus.

Since that time, I discovered that these phenomena are quite well known of course (for example in DNM amps where they specifically avoid any metal they possibly can in their designs) but more significantly to your bronze heatsink findings, that damping of all components, PCBs etc., is commonly used by many audio manufacturers to improve their sonic results.

Following on from these earlier discoveries, I experimented with all kinds of damping materials in every location one could envisage, and inevitably there were noticeable changes (not always to the good!), but to keep things in perspective, I am not talking about 'chalk and cheese' changes here. They are subtle, but nevertheless obvious to a diligent listener with a reasonably revealing system, and a half-decent pair of ears.

Acrylic, lead, and bitumen-based materials, together with some extremely dense sound-deadening polymer materials used in vehicle sound-proofing and Blu-Tac all seem to work well, as do some of the more proprietary materials like Microsorb (excellent, clean and easy to apply), Isodamp, Navcom, and Sorbothane which are effective too. The last thing anyone should ideally encourage for the best possible sound appears to be uncontrolled 'ringing' particularly in narrow bands in any constructional materials used, with steel being the worst, to my knowledge, followed by aluminium and copper-based alloys, out of all the commonly available metals.

I haven't (yet!) tried bronze for myself, but most bronze (it varies with type) has a highish constituent of tin in its alloy, up to 40%, so I would anticipate this being very good here, as a lump of tin is quite similar to lead in many physical respects. The fact that Bell Metal (being a form of bronze) seems to be counter to the general 'non-ringing' desire, is probably taken care of largely because it is known to store less energy than most metals, and because it is useful in making bells having different musical pitches, in a similar way to brass instruments, and suggests that any inherent resonance here is wider bandwidth, and less 'peaky'.

Incidentally, I can readily spoil the sound of my favourite (DIY) amps which I have spent many years in 'voicing' to my satisfaction by simply sticking some lumps of Blu-Tac on each of the output devices. I have frequently demonstated this to visitors, and the effect is not very subtle at all. It is just like someone has thrown a heavy sack over the front of the speakers, with most of the 'loss' being in the higher frequencies, and this gives an extremely lifeless un-involving result which I detest.

Similarly, whilst varnishing some wood-work in my house a few years ago, I tried out some (new to me then) more eco-friendly acrylic water-based varnish. The way it changed during drying out from opaque to a clear hard finish, fascinated me, as did the fact that after full polymerisation , even though it was water-based, its properties had completely changed (and irreversibly so) to subsequently being completely water-proof. Shortly before this time, I had read some comments about J.Peter Moncrieff's 'compound', the name of which escapes me for now, but which is apparently used by brushing on to components and supposedly has some sonic effect on them, so on a whim, I tried some varnish on my output transistors, merely to see what (if anything) might happen.

I simply daubed a quite thick layer on to the fronts of the plasic packaged mosfets (TO3P) and let it air dry for a few minutes until it was touch-dry. When I then fired up the system again, the sound was quite appalling to my ears, and to me was completely unlistenable at that point, as far as any musical enjoyment went, rather disappointingly. I stuck it out with continually listening for the ensuing perhaps 5 or 6 hours, during which time the sound gradually returned more towards normal, presumably as the heat from the devices toasted off the varnish and polymerisation was taking place. ( No, it was certainly not because I was becoming accustomed to this) By that time, I had quite enough of that experience, and duly scraped off the clear coating to immediately regain the original quality of sound I enjoyed before I had commenced this trial, I might add with some relief. Incidentally, the varnish was carefully only applied to the plastic case, and was not covering any electrical parts like the leads, so I don't believe that any dielectric/capacitive effects could have been responsible for what changes I heard

I believe it was simply the initially greater damping effects of the relatively soft varnish in the early stages, which then reduced somewhat as the varnish hardened over time, and nothing else.

Whilst this was one of my aborted trials connected with audio, of which there have been many others, similarly unsuccessful, the fact that this varnish (in different states of hardening) had such an obvious effect on the sonic results when applied merely to the faces of output devices, made me realise how much effect can be realised here, even if I didn't happen to enjoy the outcome on that particular occasion.

****Continued in next post, due to character over-run.****

[Bobken]

Finally, back on heatsinks, I can readily hear the effects through my audio system from squeezing some damping material between the fins of the amps' heatsinks, which then prevents them from making the usual tinkling sound when fingers are brushed across these fins, and this is another example which reinforces my view that these changes, if and where they occur, are mainly vibration-related. Similarly, as more than one audio manufacturer has found to his benefit (e.g. Naim Audio) tightness of output device mounting screws can and does have an effect on the sound. After much experimentation, and I have used a torque driver here for some years to ensure consistency, I found that overtightening these screws does have an adverse effect
which is also related to the type of pad used (mica & grease, sil. rubber thermal pads, almn. oxide etc.) and to the type of fixing i.e. TO3, TO3P etc.

Regards,