It looks very tempting, but I really don't have a clue about Arduino or uC programming... I could manage to get all the IC soldered with the proper stencil but the programming is way over my head without an step-by-step guide.
Also, great comments on the SQ! It looks like a really nice amp! Which SMPS are you using to drive them?
Also, great comments on the SQ! It looks like a really nice amp! Which SMPS are you using to drive them?
As of today, I'm using the power supplies of my FX-Audio D802. A Meanwell LRS350-36V is on the way.
About the usage, I tried to record some information in Home * jmf13/NeatAmpTAS3251 Wiki * GitHub
I could work on step by step guide for the programming, based on other tutorials. However, I agree that it is the kind of process where small differences can raise issues not easy to spot for the beginner. It still makes sence if it is a sort of investment for other future projects.
JMF
About the usage, I tried to record some information in Home * jmf13/NeatAmpTAS3251 Wiki * GitHub
I could work on step by step guide for the programming, based on other tutorials. However, I agree that it is the kind of process where small differences can raise issues not easy to spot for the beginner. It still makes sence if it is a sort of investment for other future projects.
JMF
NeatAMP software
Hi,
if you're interested, I can send you the binaries of my software. So no need to get into C programming: just flash the board and use it.
To flash the board, you’ll need a $2 USB programmer and a free ST software. Easy.
To use the software, you’ll need a serial link with you PC, thus an USB to serial adapter will also be required: $2 too…
Way to use the software is described is it’s GitHub repo: GitHub - AIM65/NeatAMP-Test-Dev-software: Software aimed at easing tests and setup of Class D audio amplifier NeatAMP based on TI TAS3251 chip.
Hi,
if you're interested, I can send you the binaries of my software. So no need to get into C programming: just flash the board and use it.
To flash the board, you’ll need a $2 USB programmer and a free ST software. Easy.
To use the software, you’ll need a serial link with you PC, thus an USB to serial adapter will also be required: $2 too…
Way to use the software is described is it’s GitHub repo: GitHub - AIM65/NeatAMP-Test-Dev-software: Software aimed at easing tests and setup of Class D audio amplifier NeatAMP based on TI TAS3251 chip.
Hi
I am about to complete the TAS3251 test, but I found a question with the output:
I input a 1KHz sine wave to I2S of TAS3251 with no EQ mode.
The output A/B of TAS3151 has a 600KHz PWM carrier on 1KHz sine wave.
How do I filter out 600KHz carrier to restore 1KHz sine wave
Attachment: 121923.jpg is 1K sine wave with 600KHz carrier
Attachment: FG-1K-sine.jpg is standard sine wave.
Thanks.
I am about to complete the TAS3251 test, but I found a question with the output:
I input a 1KHz sine wave to I2S of TAS3251 with no EQ mode.
The output A/B of TAS3151 has a 600KHz PWM carrier on 1KHz sine wave.
How do I filter out 600KHz carrier to restore 1KHz sine wave
Attachment: 121923.jpg is 1K sine wave with 600KHz carrier
Attachment: FG-1K-sine.jpg is standard sine wave.
Thanks.
Attachments
Output filter are mandatory.
In stereo mode, TAS3251 operate in BTL : two amplifiers are bridged to drive one speaker; as those are two class D amp you need two output filters per speaker.
Picture is from TAS3251 AVM Application note, you'll see on LC filter per Amp output, speaker connected between two A/B output. Same thing for JMF11 NeatAMP.
Do not connect your scope on speaker two terminals as you scope is ground referenced : you need a differential probe.
Waveform you currently have, referenced to ground, is OK.
Read datasheet and AN in order to choose output self, there are also many thread here about TPA3255 and TPA3251 which will give you good insight about which self to choose.
Chris
In stereo mode, TAS3251 operate in BTL : two amplifiers are bridged to drive one speaker; as those are two class D amp you need two output filters per speaker.
Picture is from TAS3251 AVM Application note, you'll see on LC filter per Amp output, speaker connected between two A/B output. Same thing for JMF11 NeatAMP.
Do not connect your scope on speaker two terminals as you scope is ground referenced : you need a differential probe.
Waveform you currently have, referenced to ground, is OK.
Read datasheet and AN in order to choose output self, there are also many thread here about TPA3255 and TPA3251 which will give you good insight about which self to choose.
Chris
Attachments
Hello AIM65,
Thank you for your reply, but I have a bit confused.
I have added L/C filter on the output A/B , which is the same as the reference design, so does 600KHz carrier on 1KHz signal is a feature of TAS3251 (BTL CLASS D)?
Don’t I need to add a filter to filter out this 600KHz carrier at OUT A/B again?
Thank you again.
Thank you for your reply, but I have a bit confused.
I have added L/C filter on the output A/B , which is the same as the reference design, so does 600KHz carrier on 1KHz signal is a feature of TAS3251 (BTL CLASS D)?
Don’t I need to add a filter to filter out this 600KHz carrier at OUT A/B again?
Thank you again.
Hi Knee266,
Short answer is : no you do not need an additional filter, the one in place is the one to have.
But maybe it’s not a temperature issue, but a thermometer issue !. I explain :
How to do you make your measurement ?
Amp has to be loaded, no need to make tests at high power level, you can install a few Watts resistor of 8Ω (or something between 4 and 40, if the purpose is only to confirm the output waveform).
Connect scope channel I to speaker terminal A and scope channel II to speaker terminal. Have channel I and II set to the same sensitivity and same 0V location on the screen (traces superposed) then set you scope to display only Channel I minus Channel II. That will display what your load see as the output of the amplifier.
Forgot to mention : scope gnd has to stay connected to amp gnd !.
Chris
Short answer is : no you do not need an additional filter, the one in place is the one to have.
But maybe it’s not a temperature issue, but a thermometer issue !. I explain :
How to do you make your measurement ?
Amp has to be loaded, no need to make tests at high power level, you can install a few Watts resistor of 8Ω (or something between 4 and 40, if the purpose is only to confirm the output waveform).
Connect scope channel I to speaker terminal A and scope channel II to speaker terminal. Have channel I and II set to the same sensitivity and same 0V location on the screen (traces superposed) then set you scope to display only Channel I minus Channel II. That will display what your load see as the output of the amplifier.
Forgot to mention : scope gnd has to stay connected to amp gnd !.
Chris
Last edited:
Hello AIM65,
For currently, I added a 4 ohm speaker on the output A for testing.
I think, I need to add load and do the test again.
Thank you for your suggestion, I will update the results after the test.
----EDIT---
My friend mentioned that 600KHz needs to be filter out, and the output should be a complete sine wave, which caused my confuse again.
Does the SNR, THD or other values in TI TAS3251 specification include a 600KHz carrier?
Thank you.
For currently, I added a 4 ohm speaker on the output A for testing.
I think, I need to add load and do the test again.
Thank you for your suggestion, I will update the results after the test.
----EDIT---
My friend mentioned that 600KHz needs to be filter out, and the output should be a complete sine wave, which caused my confuse again.
Does the SNR, THD or other values in TI TAS3251 specification include a 600KHz carrier?
Thank you.
Last edited:
Hi
I'm lost : as said before you don't need additional filtering as you already have one.
In post #367 you write that you have an LC filter at output, then you ask if you should add another. Answer is no, you only need one LC filter on each TAS output as described in the EVM schematic. This LC filter is mandatory, values are in the EVM schematic (post #366).
There's a misunderstanding, a schematic is required to clarify the situation.
Could you post schematic of your TAS output (from TAS to speaker) ? and highlight on this schematic where you put the scope probe and gnd in order to get waveform shown in post #365?
Could you make the measurement described in post #368 and post results ?
Chris
I'm lost : as said before you don't need additional filtering as you already have one.
In post #367 you write that you have an LC filter at output, then you ask if you should add another. Answer is no, you only need one LC filter on each TAS output as described in the EVM schematic. This LC filter is mandatory, values are in the EVM schematic (post #366).
There's a misunderstanding, a schematic is required to clarify the situation.
Could you post schematic of your TAS output (from TAS to speaker) ? and highlight on this schematic where you put the scope probe and gnd in order to get waveform shown in post #365?
Could you make the measurement described in post #368 and post results ?
Chris
Hello AIM65,
#365, Please refer to the attached probe.png for the position of the probe
About #368,
I can already measure a 1KHz sine wave, but because of the 600KHz carrier, the line of sine wave on the oscilloscope is very thick.
If I turn down the volume, the sine wave will be more thicker. I upload the picture for yo reference. (vol30.jpg and vol50.jpg).
I think the output waveform has some room for improvement.
How about your PCBA status?
Thank you.
#365, Please refer to the attached probe.png for the position of the probe
About #368,
I can already measure a 1KHz sine wave, but because of the 600KHz carrier, the line of sine wave on the oscilloscope is very thick.
If I turn down the volume, the sine wave will be more thicker. I upload the picture for yo reference. (vol30.jpg and vol50.jpg).
I think the output waveform has some room for improvement.
How about your PCBA status?
Thank you.
Attachments
Last edited:
Thanks,
Where is your speaker connected to ?, what are L and C values of your filter ?, what is L model and manufacturer ?, can you post a picture of the pcb ?
About vol30 and vol50 pictures, sorry to insist, but those are not measurement made according post #368. Could you connect scope channel 1 probe to OUT_A+ signal, scope channel 2 to OUT_A- signal and scope gnd to map gnd; then align trace zero level then display channel 1 minus channel 2 ?
Chris
Where is your speaker connected to ?, what are L and C values of your filter ?, what is L model and manufacturer ?, can you post a picture of the pcb ?
About vol30 and vol50 pictures, sorry to insist, but those are not measurement made according post #368. Could you connect scope channel 1 probe to OUT_A+ signal, scope channel 2 to OUT_A- signal and scope gnd to map gnd; then align trace zero level then display channel 1 minus channel 2 ?
Chris
Hello AIM65
The L/C value is 7uH(EPCOS Inc) and 0.68uF (Panasonic) same as TI recommended.
For the waveform of CH1 minus CH2, please refer to the attachment, I connected a 20W/4ohm cement resistor between OUT_A+(CH1) and OUT_A-(CH2),
The PCBA should be no problem, I think that the 600KHz carrier amplitude is the main reason. What is your 600KHz Vpp amplitude?
Thank you.
The L/C value is 7uH(EPCOS Inc) and 0.68uF (Panasonic) same as TI recommended.
For the waveform of CH1 minus CH2, please refer to the attachment, I connected a 20W/4ohm cement resistor between OUT_A+(CH1) and OUT_A-(CH2),
The PCBA should be no problem, I think that the 600KHz carrier amplitude is the main reason. What is your 600KHz Vpp amplitude?
Thank you.
Attachments
@knee266:
I think that there isn't any problem. That's totally normal output of a class-d amp.
1. 600 kHz is far out of the hearable frequency range.
2. the connected speaker acts as additional filter for the 600 kHz carrier.
3. To verify if the output LC-filter is working correctly, calculate the attenuation of the LC-filter and measure the ripple with the scope...
I think that there isn't any problem. That's totally normal output of a class-d amp.
1. 600 kHz is far out of the hearable frequency range.
2. the connected speaker acts as additional filter for the 600 kHz carrier.
3. To verify if the output LC-filter is working correctly, calculate the attenuation of the LC-filter and measure the ripple with the scope...
Hi Knee266,
What matters is the voltage seen by the load resistor, this voltage is (VoutA+ minus VoutA-). You should set your scope to display the results of the function (channel1 - channel2). We expect a single trace, not two. Do not forget to align the zero of the two channel, both should be AC coupled.
All scopes do this basic math function, usually you invert channel 2 then add the two channel.
Correct me if I'm wrong, but on the screen, I see channel 1 and channel 2, not their difference. They are phase reversed, which is as expected as amp is bridged (BTL).
Once you'll have the voltage seen by the load on the scope, you'll be able to measure de HF ripple and confirm your filter is ok or not. It's fc is close 73kHz
Chris
Hi knee266
Please find here two videos related to differential measurements, they are certainly much more useful than my written explanations...
The second one use the same scope brand as you, it's a shame the guy didn't mention that phone lines and his scope are both gnd related. Video may introduce confusion about grounding.
I've a day off tomorrow, I'll be able to post some measurements in order to compare with yours.
Chris
#35: Using the ADD/INVERT mode on an analog scope to view differential voltages - YouTube
Differential Signal Measurement Using A Digital Oscilloscope [ADSL2+] - YouTube
Please find here two videos related to differential measurements, they are certainly much more useful than my written explanations...
The second one use the same scope brand as you, it's a shame the guy didn't mention that phone lines and his scope are both gnd related. Video may introduce confusion about grounding.
I've a day off tomorrow, I'll be able to post some measurements in order to compare with yours.
Chris
#35: Using the ADD/INVERT mode on an analog scope to view differential voltages - YouTube
Differential Signal Measurement Using A Digital Oscilloscope [ADSL2+] - YouTube