pcb amp...

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

Post Reply
User avatar
chief mushroom cloud
Posts: 429
Joined: Tue Jan 18, 2005 5:42 pm
Location: Peenemunde CA

pcb amp...

Post by chief mushroom cloud »

I really need some advice from the experts (someone who has done this). If I did a pcb tube amp preamp...how wide should the signal traces be, and what thickness copper should I ask for? Should I use a ground plane or dedicated traces? Also, I need engineering explanation of why specific trace widths and copper thickness. My math is good...throw it at me.
TIA guys
Don't overthink it. Just drink it.
User avatar
jelle
Posts: 2391
Joined: Sat Feb 04, 2006 7:55 pm
Location: New Jersey

Re: pcb amp...

Post by jelle »

Equivalent of 22awg solid core wire worked for me, you do the math. And thanks again for the books! :D
User avatar
JazzGuitarGimp
Posts: 2357
Joined: Mon Jul 23, 2012 4:54 pm
Location: Northern CA

Re: pcb amp...

Post by JazzGuitarGimp »

Chief,

I haven't tried it yet, but it's on my list if things to do. I like the idea of ground plane on the side of the board farthest from the chassis, with signal traces on the side closest to the chassis. That way the signal traces have shielding on both sides. I think I would still keep each ground node (ie, the negative connection of each filter cap in the PS string) as a separate ground plane, so that they can all be grounded at one point on the chassis (or two points, if that is your preference). So, each ground plane is situated beneath only it's associated components, with perhaps 25- to 50-mils of space between planes. Also, my thinking is that all the preamp tubes should use standard chassis-mount sockets, with an oblong hole in the board to allow access to the nuts that secure the sockets to the chassis, as well as the pin connection terminals. Then use short wires to wire the sockets to pads on the board - grid resistors could be flown between a pad on the board, and the socket terminal. These are just some things I've ruminated on in the past. I'm not sure trace impedance (ie, width and thickness) is going to be critical. I like Jelle's idea of sizing them to the same cross-sectional area of a 22-awg solid copper wire. I imagine 1oz copper is more than adequate.

If you jump in with both feet, please let us know how it turns out.

Cheers,
Lou
Lou Rossi Designs
Printed Circuit Design & Layout,
and Schematic Capture
TUBEDUDE
Posts: 1864
Joined: Thu Jan 04, 2007 5:23 pm
Location: Mastersville

Re: pcb amp...

Post by TUBEDUDE »

One ounce copper is mighty thin. The equivalent trace width would be quite wide. The thicker the trace, the more thermal abuse it will endure, as you drill or swap out components repairing or modifying the circuitry. I'd go as thick as possible, although I haven't seen more than 2 oz offered by the companies I've seen. Chassis mounting of sockets is just common sense for a durable design. But then again I'm based against PCB amps anyway.
Tube junkie that aspires to become a tri-state bidirectional buss driver.
User avatar
chief mushroom cloud
Posts: 429
Joined: Tue Jan 18, 2005 5:42 pm
Location: Peenemunde CA

Re: pcb amp...

Post by chief mushroom cloud »

2 Oz trace > 22awg wire is .194in (194mils)
but according to IPC-2152, a 194mil trace can conduct 2.5A w/o breaking a sweat. Using stripline calc, this amounts to ~33ohms impedance (FAAARRR lower than necessary for tube circuit).
Has anyone here done a blind A/B listening test (apple to apples...exact same circuit and components, pcb vs ptp)?
I've been approached to do this for an amp co, and they want metrics.
Don't overthink it. Just drink it.
User avatar
chief mushroom cloud
Posts: 429
Joined: Tue Jan 18, 2005 5:42 pm
Location: Peenemunde CA

Re: pcb amp...

Post by chief mushroom cloud »

jelle wrote:Equivalent of 22awg solid core wire worked for me, you do the math. And thanks again for the books! :D
no prob Jelle....;-)
Don't overthink it. Just drink it.
User avatar
JazzGuitarGimp
Posts: 2357
Joined: Mon Jul 23, 2012 4:54 pm
Location: Northern CA

Re: pcb amp...

Post by JazzGuitarGimp »

TUBEDUDE wrote:I'd go as thick as possible, although I haven't seen more than 2 oz offered by the companies I've seen.
I'm pretty sure most board houses will plate up to 3oz. But since the current levels in tube circuits are so low, I wouldn't go beyond 1oz. A 25-mil-wide, 1oz trace will get you 3-Amp capacity with a 10º rise in trace temperature, which is way more ampacity than needed.

I like the promise of repeatable, predicable "lead dress" afforded by a pcb, though it may take a handful of layout tweaks to get the circuit to behave the way you want it to. I would suggest laying in copper pour on the signal side to shield grid circuit traces from other traces / pads that lie next to the protected (grid circuit) traces.
Lou Rossi Designs
Printed Circuit Design & Layout,
and Schematic Capture
User avatar
jelle
Posts: 2391
Joined: Sat Feb 04, 2006 7:55 pm
Location: New Jersey

Re: pcb amp...

Post by jelle »

Tracy, if something catastrophic happens in the amp, would you like the board, or some resistor to be toast?

I know what I prefer. :wink:

Jelle
TUBEDUDE
Posts: 1864
Joined: Thu Jan 04, 2007 5:23 pm
Location: Mastersville

Re: pcb amp...

Post by TUBEDUDE »

The current capacity of a thin trace may be ample, but again, that is not the only quality to consider.
Tube junkie that aspires to become a tri-state bidirectional buss driver.
User avatar
chief mushroom cloud
Posts: 429
Joined: Tue Jan 18, 2005 5:42 pm
Location: Peenemunde CA

Re: pcb amp...

Post by chief mushroom cloud »

jelle wrote:Tracy, if something catastrophic happens in the amp, would you like the board, or some resistor to be toast?

I know what I prefer. :wink:

Jelle
:lol: :lol: Yes Jelle...for my own builds, I do ptp...but this is info I need to explain to a potential contract job offer....
:wink:
Don't overthink it. Just drink it.
User avatar
chief mushroom cloud
Posts: 429
Joined: Tue Jan 18, 2005 5:42 pm
Location: Peenemunde CA

Re: pcb amp...

Post by chief mushroom cloud »

TUBEDUDE wrote:The current capacity of a thin trace may be ample, but again, that is not the only quality to consider.
Go ooon? Do tell...... :D
Don't overthink it. Just drink it.
Post Reply