[00940]

Quote:

Originally Posted by kevin gilmore High end? For high impedance phones, sound is amazing for such a little squirt. Low impedance phones, seriously lacking.

Great news, I'm wondering how it will power my K340 (400ohm). It'd be perfect with in a small box with an usb dac.

 

[was ist los?]

Kevin, you mention that the amp is lacking in sound for low impedances, but do you think it's still a better option than an opamp/buffer setup? Also, have you tried it with Ety's?

 

[kevin gilmore]

Yes it certainly sounds better than the opamp/buffer solution.

My intended use is with a pair of ety 4s. And is sounds much better
than my previous unit.

realy anchient
http://gilmore.chem.northwestern.edu/dynporta.jpg

a few years old
http://gilmore.chem.northwestern.edu/ety1.jpg

 

[Garbz]

i still don't see why a DC/DC converter is needed. THe tpa6120 runs from +- 5-15v, two 9v batteries should supply ample power.
There seems to be a lot of discussion on this forum regarding how bad DC/DC converters sound in the PSU.

/EDIT: Sorry i missed Kevin's post

 

[was ist los?]

A DC-DC converter is not required, however, i would rather use 3-4 AA batteries than 2 9V batteries. AAs will have longer runtime too.

 

[jasonhanjk]

Yes, AA NimH has gone up to 2300mAH here. It does not seems stoping at this point.

Kevin, which PCB layout did you use for drawing the amp? Cadsoft eagle?

 

[doobooloo]

Quote:

Originally Posted by kevin gilmore High end? For high impedance phones, sound is amazing for such a little squirt. Low impedance phones, seriously lacking.

Thanks Dr. Gilmore for your pics and info. It's interesting that the chip drives high impedance well and low impedance not so well - in that case, would it be helpful to add some sort of an output impedance mode for low impedance phones (maybe the 120ohm output adapter)?

 

[kevin gilmore]

I'm using Proteus Ares at the moment for board layout. I don't really
like it, but it is way better than what i was using. Expresspcb works
well, but then you are forced to pay their very overpriced prices.

Many people will find this little amplifer extremely acceptable for all
but k1000's. But something that pulls 30ma cannot possibly compete
with a dynahi that pulls 600ma. However with enough heat sink it
can actually put out more than 500ma per channel, so with a big
enough power supply, this sounds REALLY good. With a gain of 10
the output impedance is about 1.3 ohms. The dynahi impedance
is one order of magnitude better at .12 ohms. So for 32 ohm cans
one is a whole lot more tight with a lot more slamm than the other.
Notice that 9 volt batteries will sag severly under loud conditions
with low impedance cans.

A lot of people are going to like this. It is however surface mount
only, you have to be careful with how you do the heatsinking and
those with standard soldering irons will not be able to put this together.

 

[Garbz]

on the point of trying to get this chip on the board, is it absolutly necessary for that pad to be soldiered down?

I'm thinking along the lines of a large gob of artic silver on the bottom. Also i have plenty of epoxy heatsink resin leftover from my computer work so is it possible to just glue a heatsink to the top?

If that thermal pad only needs to transfer heat then the actic silver should do it, if not then there's plenty of conductive stuff too.

/EDIT: Actually i think artic silver is conductive from memory

/EDIT2: The other thing is would it be possible to put mount the heatsink underneath? Just cut a hole into the board where the thing would sit and mount a bent piece of aluminium in.

(excuse drawing it's late and i can't finda ruler)

 

[jasonhanjk]

In the semiconductor firm what they do is solder the chip to a large area PCB copper pad for heatsink. They use in line oven that is capable of heating up to 250 deg C.

For us poor fella, we need to get solder paste and maybe 100W soldering iron to do the same thing. Actually I like the idea of thermal paste heat sink.

 

[kevin gilmore]

what i did for my proto was solder the back of the chip to a piece
of copper foil cut to the right size.

 

[Garbz]

would it be too much trouble for pictures? That sounds like it's worth seeing!

 

[was ist los?]

I plan on trying what Kevin did on his proto.

 

[kevin gilmore]

I'll post some photo's when it is done, hopefully next week.
Still trying to finish other stuff, plus i'm busy at work.

 

[10SNE1]

From p. 24 of http://focus.ti.com/lit/an/slma002/slma002.pdf

An alternative to attaching the thermal pad of the package to the
thermal land of the PCB with solder is to use thermally conductive
epoxy for the attachment. This epoxy can either be dispensed from
the liquid form with a material that will cure during the reflow cycle,
or a “B” staged preform that will receive the final cure during the
reflow cycle. These materials can be the same as normally used
with externally applied heat sinks. When epoxy is used as the
attachment mechanism, then the effective attachment area is 100%
of the die pad area, and there is some added benefit as thermal
transfer to the PCB can occur, even with no copper thermal land at
the surface of the PCB.