Hi,
@Russ / Brian.. I have another one of my boring posts here, but I know you're busy, so there's no hurry in replying...
I finally got around to putting a basic amp together using a pair of RevC modules and the Buffalo DAC / LCDPS. I've actually been using the Buffalo via my Denon amp (yes, I know) since the kits arrived. The LCDPS and Volumite were (carefully) built in around two hours and all worked perfectly as expected, so thanks once again Russ / Brian. (btw, the kits arrived around the 18th Dec IIRC).
It's sounding awesome already, but I've just got a few questions...
OK, I certainly don't have golden ears as you probably know, but here is a brief summary of the sound so far...
The Buffalo does actually sound very much like I was hoping. Pretty much all of the HF blurring I was getting before seems to have gone (probably jitter related) and all the individual sounds are extremely well separated. The mid range has the biggest improvement. It has just the right amount of warmth and is VERY dynamic and punchy, which I love. Even the sub (M&K MX-350) sounds so much better, and seems to integrate with the main speakers better too.
The Buffalo is maybe a bit bright at the very top end, but I think there is an obvious reason for this is my case... I've got the Buffalo outputs connected directly to the RevC modules (using GND and +), so of course I don't have an active I/V stage as yet. Am I correct in saying that there is a fair bit of HF noise from the Buffalo which should ideally be filtered? I know from diyaudio a while back that you now issue different caps / resistors for the IVY depending on the DAC used?
I'm convinced that some of the brightness I'm hearing sounds like a very faint "ghost" of the main output at a very high frequency. Could this possibly be due to my lack of filtering, or is the HF noise from the Buffalo not really audible in the normal sense? Or, do you think it's mainly due to my lack of an output stage, which I know can cause a lean sound (sloping upwards at the top end??). I'm using the TP toslink module directly into D1 on the Buffalo (comparator switch open). My source is the PC and the Audiotrak Prodigy HiFi optical output. And yes, I'm using MP3's
for the time being, so I realize I could just be hearing more of the "perceptual coding" nasties now that the jitter isn't masking it.
(btw, I'm not good with impedance matching stuff - I did some AC theory stuff at college, but it bored me to death!)
To tell you the truth, although I'm definitely considering getting an IVY module at some point, the main factor is the size of each module. The chassis I'm using is fairly big and I plan to put everything in there, but when I eventually go to the 7.1 route I think I'll quickly run out of space. (I do plan to put a DSP board in there too, so I can do the crossover filtering and basic room correction directly before the Buffalo.) Maybe I'm trying to fit too much into one case, but it's all part of the challenge I suppose!
The other issue I'm having is overheating of the LM3886's... The chassis does have quite big heatsinks, which I thought would be more than adequate for the cooling? When playing at moderate levels the temperature of the chips rise very quickly until they're too hot to touch. This happens on my main speakers, and on my older (smaller) speakers, which are unfortunately 4-ohms nominal impedance. I realize I don't have any crossover filtering yet either, so the speakers could be presenting some bad impedances at the lower frequencies (which they weren't really designed to play).
The RevC's are bolted directly to their own section of heatsink with a thin layer of compound between them. Each heatsink "section" is 100mm x 95mm, and the fins are 25mm tall (30mm total with the 5mm base part.) The heatsink sections have 9 fins each and are separated from the next section by a very thin strip. You can just see the heatsinks in the right-hand side of this posted image...
http://www.diyaudio.com/...stid=1650256#post1650256The LM3886's are the plastic "TF" type. The compound is Artic Silver 5, and is spread thinly (not see-through, but almost). I had to drill the heatsink holes myself with a hand drill, which wasn't much fun but looks OK. I'm confident the holes were well de-burred after drilling and that the chips are sitting flat to the heatsinks, but it still feels like there's hardly any heat being transferred to the heatsinks? Do the TF chips often run quite hot to touch at higher volumes anyway? And is it generally necessary to expose the bare aluminium of anodized heatsinks where the chips are attached?
The supply toroidal is a 300VA 25-0-25 Nuvotem. Am I tempting fate with being at the preferred supply limit and having 4-ohm speakers? I've probably answered a lot of my own questions here, but I would have hoped the heat transfer to be much better? btw, what does the SPIKE protection "sound" like?
Oh, just to confuse things further - I originally bought the RevC kits from a very helpful forum member, but some of the 100nf caps C7, C19 and C20 (from across the LM318 rails) were missing. I haven't got around to ordering some more and only have a single cap across the plus and minus (LM318) rails on each monoblock - could this be a source of oscillation which could feasibly cause the LM3886's to heat up quicker?
I love the Volumite btw. I'm always a bit paranoid about volume control as I don't want it to go full volume and fry my speakers, but the Volumite is fantastic. There is also very little noise from the speakers when near the minimum volume (as long as all of the connections are short). I will be adding remote control in the near future, so I'll make sure it works as well as the Volumite does before I use it long term. ;)
Cheers,
Ash.
EDIT: Great, for some reason I've got into the habit of spelling "Buffalo" with one "F" and two "L's"!
When I was young, I remember once asking my brother what a Yak looks like - he said "They're basically a bit like a Buffallo, but with pointy horns and Dreadlocks"! You be the judge....
http://www.yak-kamel-sti...ng.de/pictures/yak_1.jpghttp://www.curly-hair-st...s-whoopi-goldberg-05.jpgEdited by user Thursday, January 8, 2009 4:41:20 AM(UTC)
| Reason: Not specified
