"Proper" TDA3116 PCB

Status
Not open for further replies.
At power-up MUTE is high and SDZ comes up after the inital 10ms power-on-delay within the TPA. After the desired delay MUTE goes low, so amp is working.

FAULTZ condition latches the outputs to High-Z, SDZ is pulled low to clear the event.

At power-off, SDZ is pulled low before the minimum PVCC is reached.

SDZ can also be pulled low by an external input.

Minumum board.current should be +40mA.
 
Or with an led-indicator..
 

Attachments

  • antipop_new_led.png
    antipop_new_led.png
    13.5 KB · Views: 456
Well, better go this way. 🙂

Almost same behaviour as "final answer" with the benefits of an led-indicator.
 

Attachments

  • antipop_new_led_pnp_final.png
    antipop_new_led_pnp_final.png
    15.3 KB · Views: 429
  • faultz_20V_pnp_final.png
    faultz_20V_pnp_final.png
    7.9 KB · Views: 433
  • ponoff_cycle_10V_pnp_final.png
    ponoff_cycle_10V_pnp_final.png
    8 KB · Views: 421
  • ponoff_cycle_20V_pnp_final.png
    ponoff_cycle_20V_pnp_final.png
    9.3 KB · Views: 425
Last edited:
SDZ-mute on power-on does not work properly to knowledge.

Edit: The Problem at PFI is, that the trip-poin scales with PVCC, so this input works only for one voltage atm. As the z-diode keeps VDD at ~the same level, undervoltage at this pin gets only detected when PVCC is at ~4.5V which is the minimum PVCC, SDZ needs to be pulled low before.

To my understanding, it wont work this way.

The "only" thing you have then, is 200ms startup delay and FAULTZ clearing (which also works with SDZ/FAULTZ tied together).
 
Last edited:
The TPS3707 is fully functional at Vdd=2.0V. So assuming the time constant of the decoupling cap and the PVCC-to-VDD resistor is short compared to the rise time of PVCC, so that VDD reaches 2.0V before PVCC reaches 4.5V, you'll be alright.

And provided PVCC plateaus before the 200mS unmute time expires, you should be OK there too. If it doesn't, PVCC is probably rising slowly enough that the TPA's own PSRR should prevent any bad sounds.
 
Gmarsh what are the apps and requirements for this amplifier anyhow? eg typical sources and speakers?
my guess or most information I can gather, since you decided to use TDA3118 and other top sided cooling isn't designed in place

inputs > stereo SE floating > ipod or mp3 player
outputs 20Wx2 (6-8ohms) > 4" full range not 4 ohm car stereo speakers
power 16-24 V independent power source > (2 wire laptop bricks or 2x12V battery no charger)
 
Last edited:
package = Hammond 1455L1202
.079 slot for 0.062" PCB X 100 mm X xxx mm
using lowest slot allows 20 mm max component height

input 2X RCA panel mount connectors ?
output 4 x ?
power 1 x ?
volume pot ?
led indicator tricolor panel mount
power switch ?
fuse ?
 
Gmarsh what are the apps and requirements for this amplifier anyhow? eg typical sources and speakers?
my guess or most information I can gather, since you decided to use TDA3118 and other top sided cooling isn't designed in place

inputs > stereo SE floating > ipod or mp3 player
outputs 20Wx2 (6-8ohms) > 4" full range not 4 ohm car stereo speakers
power 16-24 V independent power source > (2 wire laptop bricks or 2x12V battery no charger)
This card will be able to drive music into 4 ohms easily.

TPA3116/TPA3118 ICs share the same die, provided the die is cooled sufficiently to keep it below Tj(max) then you're OK. TPA3116 has a Phi-jb of 5.7 deg C/watt, TPA3118 is 4.8 deg C/watt... this is the thermal resistance of the part to the thermal pad on the device, so there's no package limitation.

Pretty much every chinese amp out there uses a "DIP14/16" sized stick-on heatsink, which has a thermal resistance of around 50 degrees C/watt in natural convection. Example heatsink:

V5619A Assmann WSW Components | AE10812-ND | DigiKey

I rarely hear reports of these cards overheating when people are using them for ordinary music listening. Based on all of that, the TPA seems to be getting by just fine as an audio amplifier with an effective Rja of 55 degrees C/watt.

My PCB design is at the very least comparable to the "old board" mentioned in section 4.2 of this document, which has a ~45 degrees/watt Rja, or say, a 40 degC/watt case to ambient assuming a 5 degC/watt junction to ambient.
http://www.ti.com/lit/an/snva183b/snva183b.pdf

The "old board" only has 2sq inches of bottom side copper, I've got 7. So I'm already better than most of the chinese TPA3116 boards out there at getting heat out of the chip.

Now if you really want to do good, buy one of these guys and stick it to the PCB underneath the TPA3118, then bolt the card to a metal chassis:

H48-6-45-25-4-1A t-Global Technology | 1168-1701-ND | DigiKey

This is what I did on the red Boominator card where I was experimenting with a 12->24V DC/DC and "class D/G" operation. I had the TPA3118 outputting 50Wrms sine into power resistors, out of both channels, with the DC/DC converter also dumping heat into the PCB, using just the Hammond 1590N case as the heatsink. It ran overnight with no problems.

package = Hammond 1455L1202
.079 slot for 0.062" PCB X 100 mm X xxx mm
using lowest slot allows 20 mm max component height

input 2X RCA panel mount connectors ?
output 4 x ?
power 1 x ?
volume pot ?
led indicator tricolor panel mount
power switch ?
fuse ?
The card won't fit the card guides of a 1455L case, bolt it to the removable bottom with 1/8" standoffs. Interior height is 27.5mm, subtract 3.2mm for the standoffs and 1.6mm for the PCB thickness and you're at 22.7mm from top of PCB to bottom of the roof. So a 20mm cap is fine, and has 2.7mm of vent space above it.

Power switch? fuse? connectors? use whatever you want. This card is intended to drop in place of an EVM or a YJ card, not to be a full-featured finished product.
 
TPA3116 has a Phi-jb of 5.7 deg C/watt, TPA3118 is 4.8 deg C/watt... this is the thermal resistance of the part to the thermal pad on the device

thanks ,this is the key that escaped me. In conclusion , the TPA3118 is superior in every way if done "properly". edit> err , what is the estimated Theta J for package tru vias to the bottom PCB ground plane, and is the solder bonded to the package in our build?
I was thinking of using this amp PCB and DAC using some +/-12 V common supply
for working to some PC stuff, so id need a true differential interface located off on a cleaner ground.
 
Last edited:
This card will be able to drive music into 4 ohms easily.

TPA3116/TPA3118 ICs share the same die, provided the die is cooled sufficiently to keep it below Tj(max) then you're OK. TPA3116 has a Phi-jb of 5.7 deg C/watt, TPA3118 is 4.8 deg C/watt... this is the thermal resistance of the part to the thermal pad on the device, so there's no package limitation.

Pretty much every chinese amp out there uses a "DIP14/16" sized stick-on heatsink, which has a thermal resistance of around 50 degrees C/watt in natural convection. Example heatsink:

I rarely hear reports of these cards overheating when people are using them for ordinary music listening. Based on all of that, the TPA seems to be getting by just fine as an audio amplifier with an effective Rja of 55 degrees C/watt.

My PCB design is at the very least comparable to the "old board" mentioned in section 4.2 of this document, which has a ~45 degrees/watt Rja, or say, a 40 degC/watt case to ambient assuming a 5 degC/watt junction to ambient.
http://www.ti.com/lit/an/snva183b/snva183b.pdf

The "old board" only has 2sq inches of bottom side copper, I've got 7. So I'm already better than most of the chinese TPA3116 boards out there at getting heat out of the chip.

We ran some tests with our small TPA3132D2 board, which gives solid 50W (2x25W) continous output power at 8R wih PVCC=23.5V. Measured with 30Hz sine. This was measured near the thermal shutdown tripping point. 🙂

To quote myself from here:

http://www.diyaudio.com/forums/class-d/251859-building-tpa3132-amp-board-6.html#post4213925

So lets see:

Ideal:
TPA3132 = 8R/(8R+0R12+0R12) = 0.97 = 97%

TA = 18°C
Pout = 50W
Efficiency: 👎 = 97%
Rth_JA = 31K/W

Pdiss = (50W / 0.97) - 50W = 1.54W
TC = TA + Pdiss * Rth_JA = 18°C + 1.54W * 31K/W = 66°C

Well, we wont have that 97% in practice, so lets assume 92%:

Pdiss = (50W / 0.92) - 50W = 4.34W
TC = TA + Pdiss * Rth_JA = 18°C + 4.34W * 31K/W = 152.54°C

For 140°C:
Pout(TC)=[n*(TC-TA)] / [Rth_JA - Rth_JA*n]
Pout(140°C) = [0.92*(140°C-18°C)] / [31K/W - 31K/W*0.92] = 45.25W

So having the correct values from the test:

VCC:23.5V
Iin:2.2A
n=0.92%

Pout: 23V5*2.2A*0.92 = 47.56W

😛

So there wont be any trouble with the TPA3116/3118 on such a big pcb.
 
ugh filterless why increase more RF spray don't you care for the environment?
We already PV micro inverters sprouting up here in my neighborhood like mushrooms.
anyway how would I know what yer doing Im in this thread not yours.

BTW doesn't the speaker network act as a filter, so that must figure in the calcs
 
Status
Not open for further replies.