pucho812
Well-known member
Anyone using these yet? Seems like the audiophiles like them. I'm thinking about a project for them but wanted to see if anyone has tried them yet.
tpa6120
- Thread starter pucho812
- Start date
Help Support GroupDIY Audio Forum:
rock soderstrom
Tour de France
Bo Deadly
Well-known member
That chip has been discussed here extensively. Google "site:groupdiy.com tpa6120". My end impression was that TPA6120 was way overkill. In most cases a simple parallel NJM4556 circuit like that NwAvGuy analyzed in some blog way back when is perfectly great for most cases (esp. headphone amp).
pucho812
Well-known member
rock soderstrom said:
wow a complete solution for around 28 dollars. interesting...
might work well to just pick this up and put it into my existing idea and done.
Bo Deadly
Well-known member
I don't know about that link. It describes a ADC from Toshiba?
pucho812
Well-known member
squarewave said:
I saw that
pucho812
Well-known member
and then we have....
https://www.amazon.com/Areyourshop-CJMCU-612-Headphone-Amplifier-Fidelity/dp/B074391GTT/ref=sr_1_3?keywords=tpa6120&qid=1569031584&sr=8-3
https://www.amazon.com/Areyourshop-CJMCU-612-Headphone-Amplifier-Fidelity/dp/B074391GTT/ref=sr_1_3?keywords=tpa6120&qid=1569031584&sr=8-3
Pucho, I was the marketing guy at TI for the TPA6120 for quite some time, and at the time, it really was a kick ass part.
However, to say it was like using farming tractor to crack a small nut doesn't even begin to describe what it was.
My former colleagues at TI have released some *really* nice, well thought out, opamps more recently, designed for use with the high end DAC's in Pro Audio.
INA1620 looks lovely.
However, to say it was like using farming tractor to crack a small nut doesn't even begin to describe what it was.
My former colleagues at TI have released some *really* nice, well thought out, opamps more recently, designed for use with the high end DAC's in Pro Audio.
INA1620 looks lovely.
pucho812
Well-known member
I'm just looking to do headphone amps for some private cue systems at the studio. 8)
Yes it is..... ;D I need to soak this in beer, but that looks like it might solve a design issue I dealt with a few decades ago, but revisited only several years ago.Rochey said:
Back in the late 70s I came up with a lower noise summing strategy using synthesized current sources (variant Howland current pump) in place of the bus feed resistors into a virtual earth sum amp. In theory summing current sources eliminates the N+1 noise gain term. The output impedance of these synthesized current sources depends on the matching of precision resistors, so in production using 1% resistors I could reduce the noise gain 20+ dB. .
Decades later I had discussions with THAT Corp engineers about repurposing the resistors on their precision line receivers to make a very good synthesized current source (long story short, they had enough resistors to do what I wanted but my design required five I/O connections and they only had four I/O protection circuits on the chip substrate).
The INA1620 has 8 precision resistors and 12 I/O pads (presumably all with ESD protection). If I sacrifice one of the two internal op amps, I could turn the other one into a kick ass precision current source.
Its a shame that large analog summing structures are all but obsolete... :
JR
[edit I attached a image of improved Howland current pump (current source).
In my consoles I used 1% 20k resistors but this should work with the precision 1k resistors on the chip.
Use one pair of 1ks for R12 and R13, another pair in series for R15, another pair in series for R14, another pair in series for R11
connect the current output node to a virtual earth sum bus.
This stage will look like unity gain differential with 1K bus feed resistor.
This would not be cheap since these would have to be after the pan pot so two per channel.
This concept only makes sense for very large structures. Last console I used this in was over 100 feeds to L/R bus... Not worth the troubles or expense for only 16-24 feeds. [/edit]
Attachments
Bo Deadly
Well-known member
Unfortunately it's a tiny 4mm square QFN package which could be difficult to solder. But not impossible. A 2mm plated hole and a lot of flux might be enough to get to the thermal pad. Then immerse the pins in solder and wick off the excess.
yeah my bad....squarewave said:
JR
Monte McGuire
Well-known member
- Joined
- Jul 26, 2013
- Messages
- 384
I've been getting small test PCBs from PCBway recently, buying framed stencils and having a good time with pure surface mount soldering. 0.5 mm lead pitch parts are a little tricky, but 0.5 mm pitch on a 4mm x 4mm DFN with a power pad isn't such an awful thing with a stencil. Yes, it can be complex to gear up for surface mount reflow, but stencils are pretty cheap now and can generally be ordered along with PCBs for not all that much additional cost.
I use an infrared pre-heater and a hot air wand to do the actual soldering. Others have used small ovens as well and had good results. I use leaded 63/37 paste for simplicity - lead free makes things more difficult for little benefit IMHO.
Just wanted to point out that a stencil isn't an impossibility today, and the advantage is that you can use some really nice modern parts.
The TPA6120 is also available in the DWP package, which is a gull wing with a power pad, and should be easier to solder by hand. Still, the DFN will have better thermal properties. Over time, more and more parts will move to leadless packages, so the writing's on the wall. It is a steep ramp, but once you get surface mount reflow down, the results are quite beautiful, and you can do some really neat thermal things with only a PC board.
I use an infrared pre-heater and a hot air wand to do the actual soldering. Others have used small ovens as well and had good results. I use leaded 63/37 paste for simplicity - lead free makes things more difficult for little benefit IMHO.
Just wanted to point out that a stencil isn't an impossibility today, and the advantage is that you can use some really nice modern parts.
The TPA6120 is also available in the DWP package, which is a gull wing with a power pad, and should be easier to solder by hand. Still, the DFN will have better thermal properties. Over time, more and more parts will move to leadless packages, so the writing's on the wall. It is a steep ramp, but once you get surface mount reflow down, the results are quite beautiful, and you can do some really neat thermal things with only a PC board.
Bo Deadly
Well-known member
Just to be clear, the last few comments were about INA1620 which is only 4mm QFN.
and I was joking.. SMD is the future and better parts are often only available in SMD.squarewave said:
JR
clintrubber
Well-known member
As I understood the Aphex HeadPod 454 uses the 6120 chip.
Possibly the later model HeadPod 4, which adds S/PDIF input, uses the 6120 as well.
I bought the 454 to be able to tackle high-Z headphones as well, and it does.
The serious metal enclosure & the big warning sign makes one a bit afraid of the capabilities of the 6120
https://www.soundonsound.com/reviews/aphex-headpod
Possibly the later model HeadPod 4, which adds S/PDIF input, uses the 6120 as well.
I bought the 454 to be able to tackle high-Z headphones as well, and it does.
The serious metal enclosure & the big warning sign makes one a bit afraid of the capabilities of the 6120
https://www.soundonsound.com/reviews/aphex-headpod
rock soderstrom
Tour de France
Rochey said:
This chip looks like a perfect solution for mobile use , as well. It can work surprisingly well with low voltages and quiescent current, check the datasheet, its a superchip! 8)
The downside is, my prehistoric soldering station and I are not able to weld this beast.
okay, a simpler SMD device wtihout the precision opamps...
http://www.ti.com/tool/TIPD177
http://www.ti.com/tool/TIPD177
Bo Deadly
Well-known member
That's just a reference design for optimal interfacing with a DAC using "differential current output".Rochey said:
The amplifier on the output is OPA1622 which is a good part and powerful but at 3mm square it's even smaller than INA1620.
Although the picture of the PCB on the web page shows what looks like an SOIC part which is strange.
pucho812
Well-known member
I picked up some premade, pre soldered tpa6120's circuit boards off amazon. They were less money per unit from overseas then it would be to buy a chip from mouser. Plus the PCB's are purple... ;D The boards look exactly like the Data sheet example of the TPA as a headphone amp. has the chip, caps, and resistors. All you do is connect it up to bipolar DC power and solder in any jacks, and whatever else you plan to build round it.
