[kevin gilmore]

How's this for nutty. Buy a bunch of the zero ohm surface mount
ultra tiny resistors. Then solder them to the board. Then solder
the chip to them. No bending of pins.

 

[DCameronMauch]

Wow. That is pretty nutty. But I would expect nothing less from you.

My PCB is coming along nicely. Hope to have it posted by end of day.

 

[Earwax]

1) On a scale of CMOY -> Meta42 -> PiMeta -> PPA where would this thing rate?

2) Would the chip fit on one of these BrownDog adapters?
http://cimarrontechnology.com/index....WPROD&ProdID=6
Assuming there's a way to deal with the heatsink issue...

 

[dip16amp]

Quote:

Originally Posted by chillysalsa Looks really interesting.. Dip16amp (or anyone else) - how would one rank the sound of this with your X16 Gilmore amp or RA-1 with a pair of Grados? If bass is so good, everything else fall into place as well?

I like my X1 amp (ra-1) for portable battery power use and X16 amp (gilmore) for home use. The TPA6120 bass is the most noticable thing over my X1, X3, X12 and everything else also sounds good so far. My X12 amp (ppa circuit) is using opa627 and double buf634 on each channel, so I can't compare it to a standard PPA. My concern of the TPA6120 is on battery power as it won't run as long as the X1 without a bigger battery pack. Runs good on my dual voltage power supply but not as good as my X16 amp.
The TPA6120 draws 21.7 mA on my +14 volt supply and 21.6 mA on my -14 volt supply.
The X1 draws 8.2 mA on each 9 volt battery.
The X16 draws 106 mA on each 14 volt supply.

 

[DCameronMauch]

Okay. Think I finished my PCB layer for this sucker. Using two TPA6120A2s for balanced output. Three AD8620s. One per channel and one for buffering ground for the non-inverting configuration feedback loop of each TPA. I like keeping absolutely all current off my ground planes. The circuit includes the Meier crossfeed circuit too. All opamps have their power JFET isolated. Here is the PCB layout in PDF format. I don't have a schematic drawn. For me, the PCB layout is the schematic. If anyone is interested, I can provide the Cadsoft Eagle format layout file. Also I guess I could draw up a schematic. The layout includes six large vias per TPA for the power pad. Should be empty space left inside to get solder through from the back side. Also didn't fill the ground plane under the critical areas of the TPA as the data sheet recommended. Whole board is 3" by 2.5". The the power, there are 3x 18mm caps and 10x 8mm caps. I planned on using Nichicon KZ series. The two big squares are Panasonic 470nF ECWF metalized polypropylene caps for the xfeed. I had to place a few resistors under them to keep the board small and get everything routed nicely. All resistors are SMD1206 variety. Attached is a tiny low resolution gif image of the board. The PDF is much high quality.

 

[aos]

It's hard to follow the circuit without schematics, but I don't understand what you did there - there's ground driver for one side of the balanced configuration but not for the other (it's grounded directly to ground plane it seems)? Also, unless you own $300 Cardas balanced cable, what good does it do - I mean, from practical perspective (the benefits of completely separate channels are obvious but without headphones that are wired in such fashion, and inability to get the connectors to make your own cables, it's useless)?

 

[DCameronMauch]

Attached is the schematic. That should make it a little more clear what I am doing with the grounds. Only the high impedance + terminal of any given amp chip is connected to ground. No current there. The feedback loops of the TPAs definitely have current flow. Thus the ground buffering. Amps 1-3 are AD8620s. Amps 4-5 are the TPAs.

I purchased the Cardas regular headphone cables. Both those and the stock HD600 cables were four conductor all the way to the plug. So I just chopped off the plug and put on a four conductor XLR. Easy.

 

[aos]

Why don't you drive the other TA with the same ground driver instead of through ground plane then? You're losing the symmetry. Your other opamp is driving the same combination anyway (non inverting plus inverting). Neat idea otherwise.

 

[DCameronMauch]

Both TDAs have their own seperate ground buffer for the non-inverting feedback loop. Those two ground buffers are done with an AD8620 dual channel opamp. Take a close look at the schematic. It is completely symmetrical.

 

[aos]

No, your non-inverting input is tied to ground, not buffered ground. That makes it non-symmetrical. Since the other opamp already drives inverting and non-inverting inputs anyway (signal driver) I don't see why can't the ground driver do the same and drive ground for both inverting and non-inverting.

 

[DCameronMauch]

I'm not following. I don't see any difference between the way the two TDAs are being driven. Do you mean the two opamps inside the TDA? The inverting one has 50 ohm to ground on the + pin. I didn't buffer that ground because there is no current flow there. The non-inverting has the AD8620 going through a 50 ohm to the + pin. And the feedback loop is going into the ground buffer. Could you download the image, circle what you mean, and post it back up?

Edit: Did you notice the opamp input pins are + on top on some and - on top for others? I did that to make the schematic visually symmetrical.

 

[aos]

Your 3's are driving the inverting pins of upper 4 and lower 5, but not the non-inverting pins of lower 4 and upper 5. There might not be (much) current there but it would still be symmetrical to connect them to 3's outputs. At the very least it will add to the offset, though in all probability it won't make much difference - but then you're already adding opamps to drive the ground which in itself probably won't make much difference.

 

[DCameronMauch]

Oh, okay. Now I'm following you. Sorry about that. I was thinking in terms of opamp with an inverting or non-inverting configuration. Not the input pins themselves. I had not considered the DC offset of the AD8620. This might be a case to pick the better grade. I checked to see if there was any easy way to route the other ground connection to the buffered ground. Unfortunately, there isn't. Looking at the data sheet for the TPA, it says the input impedance is 300k. Is that the + terminal? I would imagine for a current feedback design, the - terminal should be a great deal lower than that. Anyways, I was expecting something greater than a few meg. So there may be a tiny bit of current flow, but still quite minimal. Shouldn't disturb the ground plane too much. I'm not sure though if the DC offsets would cancel eachother out or double up at the headphones. In the second case, it might be better to only have one side subjected to the offset.

 

[null]

Sooooo.... anyone gonna take on building it for me? I'll pay them.

 

[Kenny12]

hm... how large a heatsink do you need?

also anyone have a simple layout for this thing?