WIKI's can be misconstrued as absolute references. Apparently, there are posts from me used as references on what works best. I have had no involvement with constructing the WIKI. The combination of 10uH inductors and 680nF film caps on the output LC is what works best in my system and I shared these experiences with others who can draw some help from them. Everyone should read the thread to get proper context of what is stated. For example, I tried 1,500uF Panasonic FM power supply decoupling caps and liked them more than the stock electrolytic caps, but later found that smaller capacitance caps closer in value to what is used on TI's evaluation boards and their suggested circuits work better. If you read my posts, I often mention what other equipment like speakers are in my system. What works for me may not work for others because of system differences.
Anyway, I suggest that anyone interested in these amps read the thread in its entirety and learn for themselves what might work for them. The search tools aren't ideal, but they are functional.
Anyway, I suggest that anyone interested in these amps read the thread in its entirety and learn for themselves what might work for them. The search tools aren't ideal, but they are functional.
So for 8 ohm i should get 22uH inductors then? Sorry, im still a noobie when it comes to this stuff.
Look for an impedancecurve for your speakers, a 8 ohm spec doesn't tell much. Generally you will need higher uH with higher ohms to avoid excessive peaking and filter the switching frequency.
1pc TPA3116 6 0 CH 50W 6 Amp Kit Amplifier Assembled Board | eBay
I bought this board a few months ago. Got it quickly and in tact. The board has been updated since then with better Nichicon electrolytiics. Mine had cheap Elite brand which I swapped out for Panasonic FR's. No issues with the board at all. Works great.
sabertooth:
that looks ideal for me, thanks for posting!..can I ask what you ended up with for power supply/avg output per channel?
that looks ideal for me, thanks for posting!..can I ask what you ended up with for power supply/avg output per channel?
I haven't finished the whole project yet. Haven't measured output but it gets plenty loud for my 8ohm 89db speakers. I used a Hammond 182S12 toroidal which gives about 17V 18A power to the board parallel wired. For rectifying - filter I used this board...
Heavy Duty Single Rail Power Supply PCB for OTL JHL | eBay ... Populated with Nichicon KW 35V 6800uf and a Vishay 1uf 100v. I left off the resistors. Totally unregulated linear PSU. Put in a soft start circuit as well.
I'm currently working out the preamp section (tube buffer more accurately stated) which will consist of 3 6N3 tube circuits which run off little 12vac 1a transformers. I've hooked up the circuit using a wall wart tranny and a couple different tubes to each 2 channel section and gave it a listen.
Sounded pretty nice to me. I built the pre's from bare pcb's Tube 6N3 Buffer Audio Preamplifier Pre Amp Amp DIY PCB Bare Board | eBay
Boards were only $4.99 when I bought them. They've gone up a lot since then. I sourced my own parts. It's more expensive that way but I used very good quality components .
Just need to put in the little transformers and get some 100k pots for the 6n3 circuits, wire it up and be done.
Hope to get it finished in a week or two. Not much time for my hobbies right now though.
Don't believe me, go Google, something like "amplifier input output impedance".
Its not a hard requirement rather a guideline, I have also seen 1:10 and 1:100 ratios stated. But its not straight forward and more important where there are interconnects (especially capacitative ones) between the 2 elements. Also the worst case only happens at full volume on the pot so if you are always below say 3/4 then a larger pot may work.
Personally I try to stick to 1:4 but as I say its a guideline and nothing is that black and white.
Where is the pot on your 6n3 board? A friend bought a different 6n3 board and the pot was after the active stage which kinda defeats the purpose. We are gonna remove that pot and put one up front.
irribeo, can you let us know how you arrived at those Q numbers? That might be useful info for the wiki, speaking of which...
You did make one tiny edit. 😉
Anyway, I take the credit/blame for most of what's on the Wiki. To be clear, I'm not talking about the actual "meat" of the content---I don't want to imply that I'm any kind of authority on this stuff! I mostly just copied tidbits from this and some other threads that I thought were relevant/helpful. In fact, what I did was go through this thread, and take my own personal notes, keeping references in case I ever wanted to go back and review. After that, I noticed the Wiki had only an introduction, and I thought others might find my notes useful.
So think of the Wiki like this: it's like missing a class in school, and copying your friend's notes. Better than nothing, but certainly not as useful as notes written in your own hand.
Much of that content is implicitly from Rhing, only because he's supplied a lot of useful content to these threads (usually with pics of his nice work). I hope my (excessive?) quoting of his posts doesn't discourage further contributions from him!
😀
L from BTL filter is L for SE
Load BTL is load/2 for SE
C from BTL is C*2 for SE
Read application note, Matt should too LOL
In his formula for Q for EVM he should have used: L=10u, C=1.36u and Rload = 2 (for 4 ohm)
😎 only Matt does make a mistake, which the application he refers to doesn't make. Q for EVM 4 ohm is what I posted, 0.73. His mistake is he uses the BTL filter values in SE formula, then you get Q~1 😀 Same reason why without translation to SE BTL values give totally different Q in RLC calculators then in all application reports 😀
L from BTL filter is L for SE
Load BTL is load/2 for SE
C from BTL is C*2 for SE
Read application note, Matt should too LOL
In his formula for Q for EVM he should have used: L=10u, C=1.36u and Rload = 2 (for 4 ohm)
Last edited:
Actually, some "experts" has criticized the "usual" approach of putting the pot at the input. The reason being why decreases the signal and then amplify it again. The output volume is controlled by cutting the input signal strength (using the pot)while the gain of the amplifier stage remained the same. This will decrease S/N and have detrimental effect on the sound, at least on paper.
Regards,
Alright, third time I will talk about this...
It is VERY important to have your ouput filters to match your speakers impedance as there are no UNIVERSAL output filter cuts that will serve all speakers and so one board may arrive with a filter optimized to 8 Ohms while the next board comes with 4 Ohms filters and so on.
The filter settings suggested by TI are not carved in stone, in fact these are the values that will be "OK" with 4 to 8 Ohms speakers but will better serve 6 Ohms speakers .
Bellow you have the resulting frequency cut, damping ratio and Q for the most common output filter values:
Output filter of 10uH, 0.68uF on 4 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.958706236059
Damping ratio
ζ = 0.521536192416 Overdamped - HF loss
Output filter of 10uH, 0.68uF on 6 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.639137490706
Damping ratio
ζ = 0.782304288624 Slightly underdamped
Output filter of 10uH, 0.68uF on 8 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.47935311803
Damping ratio
ζ = 1.04307238483 Underdamped - HF Peak
Output filter of 22uH, 0.68uF on 4 ohms speakers
Cut-off frequency
fc = 41148.5309373[Hz]
Quality factor
Q = 1.42199114749
Damping ratio
ζ = 0.351619629197 Very overdamped - HF Loss
Output filter of 10uH, 0.56uF on 6 ohms speakers
Cut-off frequency
fc = 67255.2386576[Hz]
Quality factor
Q = 0.704295212274
Damping ratio
ζ = 0.709929573972 Critically damped - Flat response
Output filter of 22uH, 0.68uF on 8 ohms speakers
Cut-off frequency
fc = 41148.5309373[Hz]
Quality factor
Q = 0.710995573743
Damping ratio
ζ = 0.703239258393 Critically damped - Flat response
If you have 8 ohms speakers and want a higher frequency cut like 86kHz then you can use 10uH and 0.33uF. I think this is the thid time I will post the calculator bellow:
Output Filter Calculator for class D amplifiers
Again, speaker impedance and frequency varies from one manufacturer to the other and so it is better to have a look at the REAL impedance curve of you speakers to avoid hit and miss. There are several speakers with nominal impedance of 8 Ohms that in fact are more likely to be 6 Ohms.
Harsh sound may be an indication of an underdamped filter.
You may use a underdamped filter to correct your speakers HF response.
It is up to you to decide if your speakers will be better served by an critically damped or an underdamped filter .
It is VERY important to have your ouput filters to match your speakers impedance as there are no UNIVERSAL output filter cuts that will serve all speakers and so one board may arrive with a filter optimized to 8 Ohms while the next board comes with 4 Ohms filters and so on.
The filter settings suggested by TI are not carved in stone, in fact these are the values that will be "OK" with 4 to 8 Ohms speakers but will better serve 6 Ohms speakers .
Bellow you have the resulting frequency cut, damping ratio and Q for the most common output filter values:
Output filter of 10uH, 0.68uF on 4 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.958706236059
Damping ratio
ζ = 0.521536192416 Overdamped - HF loss
Output filter of 10uH, 0.68uF on 6 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.639137490706
Damping ratio
ζ = 0.782304288624 Slightly underdamped
Output filter of 10uH, 0.68uF on 8 ohms speakers
Cut-off frequency
fc = 61033.1345767[Hz]
Quality factor
Q = 0.47935311803
Damping ratio
ζ = 1.04307238483 Underdamped - HF Peak
Output filter of 22uH, 0.68uF on 4 ohms speakers
Cut-off frequency
fc = 41148.5309373[Hz]
Quality factor
Q = 1.42199114749
Damping ratio
ζ = 0.351619629197 Very overdamped - HF Loss
Output filter of 10uH, 0.56uF on 6 ohms speakers
Cut-off frequency
fc = 67255.2386576[Hz]
Quality factor
Q = 0.704295212274
Damping ratio
ζ = 0.709929573972 Critically damped - Flat response
Output filter of 22uH, 0.68uF on 8 ohms speakers
Cut-off frequency
fc = 41148.5309373[Hz]
Quality factor
Q = 0.710995573743
Damping ratio
ζ = 0.703239258393 Critically damped - Flat response
If you have 8 ohms speakers and want a higher frequency cut like 86kHz then you can use 10uH and 0.33uF. I think this is the thid time I will post the calculator bellow:
Output Filter Calculator for class D amplifiers
Again, speaker impedance and frequency varies from one manufacturer to the other and so it is better to have a look at the REAL impedance curve of you speakers to avoid hit and miss. There are several speakers with nominal impedance of 8 Ohms that in fact are more likely to be 6 Ohms.
Harsh sound may be an indication of an underdamped filter.
You may use a underdamped filter to correct your speakers HF response.
It is up to you to decide if your speakers will be better served by an critically damped or an underdamped filter .
I think its more complicated than that, the pot is the input load for the preceding stage but for the power amp the input impedance it sees is the resistance of the pot that is going to earth which varies depending on volume setting - plenty to read up if you Google it. I have used passive preamps for many years and the 1:4 ratio was always the guideline.
Last edited:
BTL!!!!!!!!!!!! read section 4 of application sloa119a
correct calculation once again gives:
BTL filter parts: 10uH 680uF with an 4 ohm speakerload
Cut-off frequency
fc = 43156.9433363[Hz]
Quality factor
Q = 0.737563556583
Damping ratio
ζ = 0.677907680683
Breeze Audio 2.1 listening impressions
I am a happy camper with the Breeze Audio 2.1 amp (2.1 HIFI digital amplifier TPA3116 D2 50W+50W+100W 2.1 HIFI digital power amplifier /TPA3123/ TDA2030/ LM1875-in Amplifier from Electronics on Aliexpress.com) - I was given one by a generous diyAudio member as it was delivered to him DOA. The shipping caused the heavy toroidal chokes to sway back and forth and snap off a couple of the bootstrap caps. Anyhow, there is plenty of room on the underside where I added some 100 volt 0.1uF x qnty 2 as a fix for the 0.22 uF 63V caps that came stock. Once I did that, it works like a charm. Also, the puny aluminum heat sinks were double sided taped on had come off. I soldered a couple of 0.5 in dia copper rings that I hack sawed from water piping as heat sinks to the top thermal pads. I also added some hot melt glue around the inductors to stabilize them to each other and the filter cap bank.
I am listening to it with the stereo channels driving a 3.5 in Vifa TC9FD fullrange in a 0.4x scale Karlsonator cabinet and the subs driving dual 6.5 in woofers in a Nautaloss sub cabinet. Powering it all with a 90 watt Fujitsu laptop 19volt smps brick. It sounds fantastic! 😀
The Breeze Audio 2.1 is one of the few (if not the only one) with a built in variable frequency XO for the sub with independent volume for main and sub. This feature works very well and allowed me to integrate the sub to perfectly balance out the sound.
I am really impressed by how compact this amp is - it is about half as short as I imagined it would be. It is 3.5 in wide x 4.3 in long - smaller than a standard photograph.
The build construction, apart from the beefy looking but floppy inductors, indicates pretty good neat soldering. Not the optimal layout like the Ybdz 2.0 in terms of short distances between all the bypass caps, etc but fairly good quality parts throughout. I am surprised they are using electrolytic 3.3 uF Elna caps on the inputs? Would non-electrolytics sound better here? Also, the main DC power rail caps are 220 uF Elite brand - probably could use an upgrade here to Panasonic OSCON 330uF's.
Thanks again to my benefactor for giving me this nice Father's day present! 🙂
Here is a photo of the amp with the copper ring heatsinks and hot glue stabilized inductors. Great sounding amp!
I am a happy camper with the Breeze Audio 2.1 amp (2.1 HIFI digital amplifier TPA3116 D2 50W+50W+100W 2.1 HIFI digital power amplifier /TPA3123/ TDA2030/ LM1875-in Amplifier from Electronics on Aliexpress.com) - I was given one by a generous diyAudio member as it was delivered to him DOA. The shipping caused the heavy toroidal chokes to sway back and forth and snap off a couple of the bootstrap caps. Anyhow, there is plenty of room on the underside where I added some 100 volt 0.1uF x qnty 2 as a fix for the 0.22 uF 63V caps that came stock. Once I did that, it works like a charm. Also, the puny aluminum heat sinks were double sided taped on had come off. I soldered a couple of 0.5 in dia copper rings that I hack sawed from water piping as heat sinks to the top thermal pads. I also added some hot melt glue around the inductors to stabilize them to each other and the filter cap bank.
I am listening to it with the stereo channels driving a 3.5 in Vifa TC9FD fullrange in a 0.4x scale Karlsonator cabinet and the subs driving dual 6.5 in woofers in a Nautaloss sub cabinet. Powering it all with a 90 watt Fujitsu laptop 19volt smps brick. It sounds fantastic! 😀
The Breeze Audio 2.1 is one of the few (if not the only one) with a built in variable frequency XO for the sub with independent volume for main and sub. This feature works very well and allowed me to integrate the sub to perfectly balance out the sound.
I am really impressed by how compact this amp is - it is about half as short as I imagined it would be. It is 3.5 in wide x 4.3 in long - smaller than a standard photograph.
The build construction, apart from the beefy looking but floppy inductors, indicates pretty good neat soldering. Not the optimal layout like the Ybdz 2.0 in terms of short distances between all the bypass caps, etc but fairly good quality parts throughout. I am surprised they are using electrolytic 3.3 uF Elna caps on the inputs? Would non-electrolytics sound better here? Also, the main DC power rail caps are 220 uF Elite brand - probably could use an upgrade here to Panasonic OSCON 330uF's.
Thanks again to my benefactor for giving me this nice Father's day present! 🙂
Here is a photo of the amp with the copper ring heatsinks and hot glue stabilized inductors. Great sounding amp!
