Hi, I just bought a few Texas Instruments TPA3118 boards to test out for a mono boom box. I bought the very plain type with only the chip and in/mute/power/out solder points.
The other components in my test bench are a Dayton Audio PS65LP-4 (4 Ohm, 99 dB sensitivity), a multi-purpose laptop charger set at 19V output, and a 100K log pot on the signal. The input jack has the two channel positives shorted to mono. All components are brand new.
After putting everything together I noticed a few obvious issues with the sound that I had not heard about in any of the forums dealing with this chip:
1. The sound is extremely low, with no bass. With the pot at max level I get a sound that's similar to a budget laptop speaker, even worse in terms of bass (I am regulating the volume from my cell phone). With a 99dB speaker and a 60W amp, I should have blown up my neighbor's ears by now...
2. The pot seems to do very little difference. Even at minimum I have a signal, and it's clearly lower than at max, but not what I expected from a 100K pot. I even soldered an additional 100K resistor in series, which made little difference.
3. Even changing the input voltage from 19 to 15V and back makes very little difference.
OK, the speaker has no enclosure and the system has not been broken in, but this is way worse than any other brand new system I have put together before.
See attached photo of the assembled system. I double checked input and speaker polarity, and input signals. Can anybody think of any obvious flaw (except for a defective board)?
Thanks.
gm
The other components in my test bench are a Dayton Audio PS65LP-4 (4 Ohm, 99 dB sensitivity), a multi-purpose laptop charger set at 19V output, and a 100K log pot on the signal. The input jack has the two channel positives shorted to mono. All components are brand new.
After putting everything together I noticed a few obvious issues with the sound that I had not heard about in any of the forums dealing with this chip:
1. The sound is extremely low, with no bass. With the pot at max level I get a sound that's similar to a budget laptop speaker, even worse in terms of bass (I am regulating the volume from my cell phone). With a 99dB speaker and a 60W amp, I should have blown up my neighbor's ears by now...
2. The pot seems to do very little difference. Even at minimum I have a signal, and it's clearly lower than at max, but not what I expected from a 100K pot. I even soldered an additional 100K resistor in series, which made little difference.
3. Even changing the input voltage from 19 to 15V and back makes very little difference.
OK, the speaker has no enclosure and the system has not been broken in, but this is way worse than any other brand new system I have put together before.
See attached photo of the assembled system. I double checked input and speaker polarity, and input signals. Can anybody think of any obvious flaw (except for a defective board)?
Thanks.
gm
Attachments
1 - post a link to the actual board.
2 - sketch a schematic of how you have connected everything.
3 - it looks that you have connected red input wire to L & R from your phone - might be better to put a resistor from L, a resistor from R then join after the resistors - the phone might not like the outputs being joined.
4 - the vol pot looks to be wired strangely. Connect one end to source signal +. other end to signal -, wiper to amplifier input +.
5 - Is this a monoboard or a stereo board wired for mono? (your response to 1 will reveal the answer).
2 - sketch a schematic of how you have connected everything.
3 - it looks that you have connected red input wire to L & R from your phone - might be better to put a resistor from L, a resistor from R then join after the resistors - the phone might not like the outputs being joined.
4 - the vol pot looks to be wired strangely. Connect one end to source signal +. other end to signal -, wiper to amplifier input +.
5 - Is this a monoboard or a stereo board wired for mono? (your response to 1 will reveal the answer).
It appears to be the mono (paralleled) version which means the input load varies between 9K and 60K per
channel or 5-30K (mono) depending on the gain setting. I believe 26 dB is considered 'standard' so a 100 K
pot is not a good choice into a 15K input. Adding a 100 K resistor on series with the pot knocks off several
more dB. Read the fine print in the data sheet. The mono version doubles the power because it doubles the
current so you CAN use a 2 ohm load but you're using 4 so you'll get no benefit but that is OK. "No Bass"
from a free standing diver means nothing to me. They always have no bass without a baffle. Changing the
power supply will do nothing to the response, just the maximum power output. I also have no idea what a
phone puts out. The MUTE input may need to be shorted to run. The data sheet says operate is mute pin to 0.
Lose the 100 K resistor, change to 100 K put to something between 10 and 25K. Be cautious about measuring
the output. The outputs should be 1/2 the power supply Voltage. Being a Bridge Tied Load (BTL) amp BOTH outputs
move relative to ground.
I would use Audacity (free) to make some test signals and use the analog output of the computer as your
source as it's much more predictable. BTW I use TPA3116 (close cousin to the TPA3118) for TV and there is no
low end response issue at all.
G²
channel or 5-30K (mono) depending on the gain setting. I believe 26 dB is considered 'standard' so a 100 K
pot is not a good choice into a 15K input. Adding a 100 K resistor on series with the pot knocks off several
more dB. Read the fine print in the data sheet. The mono version doubles the power because it doubles the
current so you CAN use a 2 ohm load but you're using 4 so you'll get no benefit but that is OK. "No Bass"
from a free standing diver means nothing to me. They always have no bass without a baffle. Changing the
power supply will do nothing to the response, just the maximum power output. I also have no idea what a
phone puts out. The MUTE input may need to be shorted to run. The data sheet says operate is mute pin to 0.
Lose the 100 K resistor, change to 100 K put to something between 10 and 25K. Be cautious about measuring
the output. The outputs should be 1/2 the power supply Voltage. Being a Bridge Tied Load (BTL) amp BOTH outputs
move relative to ground.
I would use Audacity (free) to make some test signals and use the analog output of the computer as your
source as it's much more predictable. BTW I use TPA3116 (close cousin to the TPA3118) for TV and there is no
low end response issue at all.
G²
The pot wiring was the problem! thank you. As to your questions:
1. the item I bought on eBay: Audio Amplifier Mono Digital tpa3118 PBTL 60w 8v to 24v 45x3 Car 392 | eBay Chip data page from the manufacturer: TPA3118D2 data sheet, product information and support | TI.com
2. See attached.
3. What is the recommended resistor value?
4. See diagram. I connected the wiper to the jack +, the end with the lowest resistance at max to the amp input +, and the - straight from the jack to the amp. This was clearly wrong.
5. Mono.
What remains to fix is the bass quality, which I will evaluate in a bass reflex enclosure after some break-in and maybe by improving the capacitors as I saw suggested somewhere else.
Thanks.
gm
Attachments
Yeah, the other 100K resistor was just a test. I'm actually thinking to use a 250K pot because the 100K is very sensitive after wiring it correctly. I foresaw that I would need a higher value to play the box low indoors and pump it up in loud, outdoors scenarios.
It seems like MUTE cuts the output when shorted, so the normal operation is with MUTE open.
Thanks.
gm
Near the bottom of the Wiki page it shows how a pot should be wired which is not what you did in your drawing.
Potentiometer - Wikipedia
If you're mono mixing the output of the phone you should have a separate resistor for each channel into the
top of the pot. 100 ohms would be good into a 10K pot.
G²
Potentiometer - Wikipedia
If you're mono mixing the output of the phone you should have a separate resistor for each channel into the
top of the pot. 100 ohms would be good into a 10K pot.
G²
Are you serious? Judging bass response without a speaker enclosure?
Btw - "breaking in" is that kind of audiofool babble you can ignore completely. There is no break in. Period.
Re check the data sheet. The 3116 and 3118 are both rated for 7.5 Amps for the over current trip point. The
3130 is 4.5 Amps. The The 4 ohm load makes no difference except possibly with continuous sine wave because
the 3118 does have as much heatsink capacity due to the power pad being on the bottom rather than the top
like the 3116.
G²
3130 is 4.5 Amps. The The 4 ohm load makes no difference except possibly with continuous sine wave because
the 3118 does have as much heatsink capacity due to the power pad being on the bottom rather than the top
like the 3116.
G²
Get rid of the pot for testing, make it as simple as possible.
Try one input channel, forget about the other one. Undo the short of the input channels.
Are you sure you've the correct cable from the phone. Lots have "3 ring" connectors having L,R,GND,MIC connections. A stereo mini-jack will have unpredictable results.
Try one input channel, forget about the other one. Undo the short of the input channels.
Are you sure you've the correct cable from the phone. Lots have "3 ring" connectors having L,R,GND,MIC connections. A stereo mini-jack will have unpredictable results.