• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

"The Wiener" TPA3118 amplifier, group buy #3 + "Wiener Pro" prototypes.

NICKEL, audioillness and initials CFC (I think you're themixmusic, maybe), and lacro: payment received, order spreadsheet updated.

Group buy is now closed. Here's the grand totals:

2 bare stereo cards
9 4-ohm stereo cards, 8 with filter kits
8 6-ohm stereo cards, 3 with filter kits
26 8-ohm stereo cards, 7 with filter kits
1 16-ohm stereo card

8 bare PBTL cards
12 4-ohm PBTL cards, 8 with filter kits
6 8-ohm PBTL cards, 2 with filter kits

9 bare Pro cards
3 4-ohm Pros, 2 with filter kits
4 6-ohm Pros
18 8-ohm Pros, 4 with filter kits
1 16-ohm Pro

36 mounting kits.

Got a lot of assembly/packaging work ahead of me 🙂

Just transferred out my Paypal balance. Going to do one last design review on the Pro card (double check all my pinouts and all that), get my PCB files all packaged up and pull the trigger at Elecrow as soon as the Paypal transfer completes, which will hopefully be Friday.
 
Gmarsh, I notice the huge power trace goes directly under the inductors. I wonder if those inductors will radiate out into the power trace for such close proximity.
They're shielded inductors, they don't have much of an external magnetic field to begin with and the field they do have isn't in the ideal axis to couple into the power plane underneath.

There probably is a tiny amount of coupling present, but I doubt you'll be even able to measure it - the TPA chip itself is pulling upwards of 10A through the same trace under the inductor, if inductor coupling adds or subtracts a uA from that, it's hardly an issue.

I've run plenty of traces under power inductors in SMPS designs, including high impedance signals carrying voltage feedback and what have you, never had an issue.
 
Just ordered $500 worth of PCBs and stencils from Elecrow. It'll be about a week of manufacturing, followed by a week of shipping, followed by "bend over, it's canadian customs time..."

Fingers crossed I got the design right. The classic Wiener cards are a proven design and I've made practically no changes to them, but the Pro is an all new design. I've double checked everything but I can't help be a bit nervous I f'ed something up 🙂

Went with lead HASL for the classic cards, and ENIG for the Pro. Black soldermask all around.
 
Just ordered $500 worth of PCBs and stencils from Elecrow. It'll be about a week of manufacturing, followed by a week of shipping, followed by "bend over, it's canadian customs time..."

Fingers crossed I got the design right. The classic Wiener cards are a proven design and I've made practically no changes to them, but the Pro is an all new design. I've double checked everything but I can't help be a bit nervous I f'ed something up 🙂

Went with lead HASL for the classic cards, and ENIG for the Pro. Black soldermask all around.

Thanks for the update. Many people here are very eager to get their hands on the Wiener.
 
Haven't made the Mouser/Digikey orders yet (happening this week), I'll still take your order. Still a 4 ohm stereo standard Wiener with a filter kit?

PCBs shipped this morning from Elecrow, I'll have them by the end of the week or Monday hopefully.
 
So my TSSOP programming socket arrived in the mail for programming the ATTiny20's.

Open the package... it's a SOIC socket. sigh.

Didn't plan on populating the 6-pin ISP header on all of the boards, but that may end up happening...
 
Oh the joys of ordering crap from China on eBay.

I threw together a quick/simple PCB for programing bare Atmel AVR chips, which I ordered alongside the Wiener cards. I also ordered matching "test sockets" for 150 and 208 mil SOIC-8 packages and 4.4mm TSSOP-28.

150 mil SOIC socket came last week, exactly as described.
TSSOP28 socket came yesterday, received a SOIC-28 socket instead.
208 mil SOIC socket came today, open it up, received another 150 mil SOIC socket.

Just got refunded on the TSSOP socket, time to order another one and cross my fingers...
 
The card has differential inputs, with +/-/GND pins. (GND was missing on the previous stereo card, it's in there now.)

To hook up single ended input sources, connect input shield to - and input signal to +. And either connect the input shield to GND also, or ensure that the source ground and the Wiener power supply ground are the same by some other means.

For differential sources, input +/- goes to card +/- and cable/signal ground ("pin 1" on an XLR) goes to GND.

This is from way back. But am I correct in thinking the 3118 Stereo Wiener can handle balanced or unbalanced sources through the same 1/4" TRS jack? 1/4" unbalanced plugs would short - and shield, right?
 
Boards and stencils are here. And the VER2923 footprint worked, I was a bit nervous the board shop would screw it up 🙂

4MATbeih.jpg


Making the Mouser order tonight. Parts should be here Thursday or Friday, but my brother's in town over the weekend so it'll be next week when I start building.

Got the ATTiny20 code written for the Pro, but untested on hardware.
 
Looks great!! Looks like Au plate. I guess you mean the plated slots for the VER2923 footprint. They must be able to read a fab drawing or a drill schedule or milling data 🙂
Now try to get a slot with no plating 🙂
Good luck with your bring up effort!!
Cheers have a beer for me
 
That's Elecrow's ENIG service. Normally I get Pb HASL because the pre-tinned pads solder very easily with the Pb solder paste I'm using, but figured I'd go for gold this time, I just love the look of ENIG and black solder mask together.

There's no standardized way to specify a plated slot versus a non-plated one. You can fake one by putting a bunch of overlapping drills in a row, but sometimes you get ugly results and sometimes the board house rejects it. So you draw a slot on the dimension layer, and hope they cut the slot before plating. I mentioned this to Elecrow when I made the order, and they never acknowledged it, so I was a bit worried 🙂

Anyway, found my first bug, turns out PB1 on the ATTiny20 was TPIDATA, not PB2. Thankfully PB1 is non-connected so I can fix the board with a wire link when I'm bringing up the first board, and I'm preprogramming my microcontrollers before installing them on the production board. Hopefully that programming socket gets here soon and they send me the right one this time...
 
For the non pcb designers that is Electrolysis Nickel Immersion Gold plating 🙂 we used the process many years ago when we were doing BGA pkgs. i agree I always liked gold plate. At one time HP was one of the few that had the process. There was a fab in germany, boeblingen that HP bought, state of the art at that time. Anything that was 4/4 and high aspect ratio was done there.
For the slots we always identified it on the drill drawings and drill schedule. made sure we noted it in the readme file too. The drill hit was the diameter of the slot and at the centroid of the slot. If you want non-plated we provided both plated and non-plated drill files.
Lots of times they do not read the docs, it can be hit and miss as you have probably experienced. Glad to see they figured it out.
Lets hope you can keep it too one barnacle.
cheers
 
Up to three barnacles. I haven't put a single part on the new PCBs yet, unfortunately I'm discovering all of these while designing the ATTiny20 into something else.

#1: Wrong programming pin hooked up, mentioned above.
#2: Silkscreen error on the bottom side, instructing the user to remove the wrong caps when using PBTL mode. I'll sharpie over that 😉
#3: Apparently the Tiny20 needs a 5V VCC supply to program. I'm running it at 3.3V, so in-circuit programming will be out of spec. Absolute maximum voltage of the reset/fault/whatever pins on the TPA is 4.2V, which necessitated the 3.3V rail.

Here's hoping the chip actually programs at 3.3V so I can proof the design... or off 4V fed from a bench supply or something.