3E Audio 260-2-29A : TPA3255 / HA3588 CoilCraft / PFFB Fully differential.

[ngk275]

That's not what "The G Word" says. The 'Pin 1 Problem' is explicitly about differential inputs and the problem created when manufacturers do the wrong thing with ground, removing the immunity from ground related noise that differential inputs ought to have. "The G Word" explains how this happens and how designers can avoid it. AES48 is a standard created to try to get manufacturers to consistently do it properly.

The essence of it is to make sure stray ground currents from loops or leakage are directed away from your audio signal. The suggested way to do this is to connect pin 1 to the metal chassis, and not to connect it to the board inside the chassis. The chassis provides a path for the ground currents, and you don't create an alternative path through the board. Using wires instead of the chassis should work, and may be your only option if the chassis is insulating (wood, plastic).

Connecting Pin 1 to the board may or may not be a problem depending on how well the manufacturer has handled routing on the board. Note the bit in "The G Word" about workarounds to make the PCB routing software route a critical part of the circuit the way the designer intended so that it would not be affected by stray ground currents.

Measurements will only show a difference if you have a ground related noise issue when measuring. See section 3.3 of https://www.jensen-transformers.com/application-notes/ AN007 for a way of introducing ground noise so you can check whether what you have done is really immune. That app note has in depth coverage of the causes of ground related noise issues, and a systematic method for solving them.

Thanks for your explanation. More to read!

English is not my first language so maybe I wrongly understand this sentence from "The G Word". Or it could be very well that I don't understand what is meant by the "differential method". This 3e board is PFFB fully differential as said in this thread name. Not the same thing?

 

[antcollinet]

Thanks for your explanation. More to read!

English is not my first language so maybe I wrongly understand this sentence from "The G Word". Or it could be very well that I don't understand what is meant by the "differential method". This 3e board is PFFB fully differential as said in this thread name. Not the same thing?
View attachment 334438

It is likely some parts of the circuit are not differential.

For example - we don't know if the feedback is implemented as differential signalling. Just because the chip amp is differential doesn't mean all the circuits surrounding it are.

Further - insensitive doesn't mean completely immune.

 

[ngk275]

I think I've got it! So it's still best to follow the best practice:
- Pin 1 to chassis, shortest cable possible. Maybe using XLR connector with pin 1 connected to chassis internally is even better.
- 1 GND from amp to chassis as well, maybe to star point.

Moving to choosing a suitable PSU. I intend to pair this amp with Elac DBR 6.2 which has nominal impedance of 6 Ohms:
- With 36V10A PSU, I can get maximum I of Min (10/2 = 5A, 36/6 = 6A) = 5A per channel
- At 5A we have 5^2 * 6 = 150W for each channel which is already more than maximum input of this speak at 120W
So 36V10A should do fine, right?

Another speaker I consider is Wharfedale 12.1, 8 Ohms, maximum input at 100W, only 28 V needed to have 3.5A per channel.
36V10A is okay in this use case too.

Edited after realizing my mistake.

 

[somebodyelse]

Thanks for your explanation. More to read!

English is not my first language so maybe I wrongly understand this sentence from "The G Word". Or it could be very well that I don't understand what is meant by the "differential method". This 3e board is PFFB fully differential as said in this thread name. Not the same thing?
View attachment 334438

The key part you missed is the "method explained earlier" - the earlier parts of "The G Word" have been explaining how to design differential circuits and their connections so that 'ground' doesn't influence the signals, among other things. The 'Pin 1 Problem' section explains how some designers have not been following that method, and why it leads to problems severe enough for AES to write a standard to avoid them.

 

[daniboun]

Let's split that into two, and first ignore the frequency characteristic of a speaker. Lets first model it as a fixed and purely resistive load.


So if you have an 8 ohm speaker.

Power = IxV
Power also = Vsquared/R
And Power = Isquared*R

(These are all the same law = if you include V= I * R, and I = V/R and substitute those terms in for I or V in the first equation above)

Amps are first and foremost a voltage output device. But they can also only output a certain current.

So if you have an amp that can ouput 20V, then it can put P = Vsquared/R = 400/8 = 50W into the speaker.


In order for it do to this, it must be able to output I = 20/8 amps = 2.5A
Double check - 20V times 2.5A = 50W, yes everything checks out.

Lets say the amp is able to output 3A, so everything is fine. The power output is limited by the maximum voltage the amp can put out.


So we connect the same amp to a 4ohm speaker.

Now, from the voltage it would be possible for it to deliver 20*20/4 = 100W (double that into the 8 ohm speaker)

But then it would need 20/4 amps = 5A (double the current). It can't do that, it can only do 3.5 A. so with the 4ohm speaker we are current limited. When the amp reaches it's current limit it must reduce the voltage (or switch it off it goes into full protection)

Any questions so far before we start to think about speaker frequency characteristic?

What has been explained here is very interesting and verifiable.

Under 4R the current considerably influences the power
Under 8R we are more dependent on the voltage

 

[daniboun]

I think I've got it! So it's still best to follow the best practice:
- Pin 1 to chassis, shortest cable possible. Maybe using XLR connector with pin 1 connected to chassis internally is even better.
- 1 GND from amp to chassis as well, maybe to star point.

Moving to choosing a suitable PSU. I intend to pair this amp with Elac DBR 6.2 which has nominal impedance of 6 Ohms:
- With 36V10A PSU, I can get maximum I of Min (10/2 = 5A, 36/6 = 6A) = 5A per channel
- At 5A we have 5^2 * 6 = 150W for each channel which is already more than maximum input of this speak at 120W
So 36V10A should do fine, right?

Another speaker I consider is Wharfedale 12.1, 8 Ohms, maximum input at 100W, only 28 V needed to have 3.5A per channel.
36V10A is okay in this use case too.

Edited after realizing my mistake.

I don't know if you were able to read my comments above. I think the choice of PSU will actually depend on the impedance of your speakers.

BTW, take a look a this, if you are looking for amazing Speakers @ fair price :

Elac FS 247.4, a big slap in the face!

Hi amigos, After listening, I had the opportunity to purchase a pair of Elac FS 247.4. I might as well tell you that at €1000 per pair (new), I got a golden deal) And what a big slap in the face! These speakers are exceptional from the first listen. For instance : nn the complex title Grare...

www.audiosciencereview.com

Dynavoice DF-8 Measurements : need feedbacks )

Hi amigos, Found a pair of Dynavoice DF-8 for a great price (700€ / pair) but never had the opportunity to test it.... Seems they measure pretty well and the lowest frequency @-3 dB is around 40 Hz, amazing for the price. Here the full and detailled measurements (NB : you need to Translate via...

www.audiosciencereview.com

 

[ngk275]

I don't know if you were able to read my comments above. I think the choice of PSU will actually depend on the impedance of your speakers.

BTW, take a look a this, if you are looking for amazing Speakers @ fair price :

Elac FS 247.4, a big slap in the face!

Hi amigos, After listening, I had the opportunity to purchase a pair of Elac FS 247.4. I might as well tell you that at €1000 per pair (new), I got a golden deal) And what a big slap in the face! These speakers are exceptional from the first listen. For instance : nn the complex title Grare...

www.audiosciencereview.com

Dynavoice DF-8 Measurements : need feedbacks )

Hi amigos, Found a pair of Dynavoice DF-8 for a great price (700€ / pair) but never had the opportunity to test it.... Seems they measure pretty well and the lowest frequency @-3 dB is around 40 Hz, amazing for the price. Here the full and detailled measurements (NB : you need to Translate via...

www.audiosciencereview.com

I did consider speaker nominal impedance in my calculation and compared what PSU can produced vs. the maximum input of the speaker. Next level might be take 5.5V into the equation. Isn't this the right approach? If not, please point me to the right direction.

Thanks for widening the speaker universe to me. Before, I only know of speaker reviewed by Amir and Erin.

 

[daniboun]

I did consider speaker nominal impedance in my calculation and compared what PSU can produced vs. the maximum input of the speaker. Next level might be take 5.5V into the equation. Isn't this the right approach? If not, please point me to the right direction.

Thanks for widening the speaker universe to me. Before, I only know of speaker reviewed by Amir and Erin.

I have difficulty understanding your approach? Do you want to have an Audio module that best fits your next speakers or do you want to choose speakers that measure well and could be driven by any good module?

I sincerely think that when you are looking for a good speakers, you must consider their versatility, obviously taking into account the measurements. As enthusiasts, we all have a tendency to update our setups and I probably think more about our amps than our speakers )
BTW, do you listen to your speakers at high volume levels or do you generally listen to them at a moderate listening level? Last week end I pushed my volume all the way up with my dual Mono 3E Audio 'Bamboo project', I realized that it was more than enough to cover my satisfaction)

 

[ngk275]

I have difficulty understanding your approach? Do you want to have an Audio module that best fits your next speakers or do you want to choose speakers that measure well and could be driven by any good module?

I sincerely think that when you are looking for a good speakers, you must consider their versatility, obviously taking into account the measurements. As enthusiasts, we all have a tendency to update our setups and I probably think more about our amps than our speakers )

Ok, I understand your view now.

But this project is for my bedroom, given the space and the budget I have, as well as speakers available where I live, Elac DBR 6.2 and Wharfedale 12.1 are the 2 best candidates. That's why I'm calculating the PSU specs based on these 2 speakers. So "want to have an Audio module that best fits your next speakers" is what I'm after.

Just moderate to low-high level if that's make sense to you.
I set up a system like this for a friend living room: Wharfedale 12.1, Fosi V3 48V5A, SMSL C100. With maximum volume at SMCL C100, volume of the Fosi V3 stay at around middle. We tried to turned it all the way up but couldnt stand the loundness. I very much agree with other member that we only need 25V rsm for normal listening.
On the other hand, I cant resist the urge to try to get an amp that could bring maximum input power to a specific speaker ^^!

 

[daniboun]

Ok, I understand your view now.

But this project is for my bedroom, given the space and the budget I have, as well as speakers available where I live, Elac DBR 6.2 and Wharfedale 12.1 are the 2 best candidates. That's why I'm calculating the PSU specs based on these 2 speakers. So "want to have an Audio module that best fits your next speakers" is what I'm after.

I think you can also consider the TI datasheets , I guess that if they exist, they have meaning)

 

[antcollinet]

- With 36V10A PSU, I can get maximum I of Min (10/2 = 5A, 36/6 = 6A) = 5A per channel

you've missed that the PSU puts out 36V Dc - which can only be the peak of the sinusoidal waveform. The maximum RMS would be 36/Sqrt(2)

= 36/1.41 = 25.4V. In reality a little lower since there will be some voltage lost in the output stage of the amp.

So lets say 24V RMS output = 4A into 6ohm = 96W - pretty much matches the chart posted by @daniboun above.


Personally I'd be looking for a higher voltage power supply.

 

[ngk275]

you've missed that the PSU puts out 36V Dc - which can only be the peak of the sinusoidal waveform. The maximum RMS would be 36/Sqrt(2)

= 36/1.41 = 25.4V. In reality a little lower since there will be some voltage lost in the output stage of the amp.

So lets say 24V RMS output = 4A into 6ohm = 96W - pretty much matches the chart posted by @daniboun above.


Personally I'd be looking for a higher voltage power supply.

Got it, need to calculate base on RMS, not the peak value. Do I need to take into account 5.5V headroom mentioned by S=klogW?
The chart makes sense to make. After rereading ur discussion with S=klogW I don't understand how this amp could get 260W @ 4 Ohm per channel with the chosen PSU of 48V12A. Amir's review also showed that max power per channel at 1% THD+N (-20 db right?) is about 220 - 230W. Scratching my head!

Another that thing that I dont understand is higher Voltage PSU seems to give better performance, like at 48V @ 4 Ohm is 230W for 1% THD+N. So lower power, such as 96W should mean better THD+N. But 96W with 36V only gives 1% THD+N. Why both S=klogW and daniboun say that 36V is better for this amp?

 

[daniboun]

Got it, need to calculate base on RMS, not the peak value. Do I need to take into account 5.5V headroom mentioned by S=klogW?
The chart makes sense to make. After rereading ur discussion with S=klogW I don't understand how this amp could get 260W @ 4 Ohm per channel with the chosen PSU of 48V12A. Amir's review also showed that max power per channel at 1% THD+N (-20 db right?) is about 220 - 230W. Scratching my head!

Another that thing that I dont understand is higher Voltage PSU seems to give better performance, like at 48V @ 4 Ohm is 230W for 1% THD+N. So lower power, such as 96W should mean better THD+N. But 96W with 36V only gives 1% THD+N. Why both S=klogW and daniboun say that 36V is better for this amp?

As we have demonstrated several times and it is quite classic and known with the TPA3255 chip, it shows a better THD+N figure at lower voltages.

We discussed and measured this here :

3E Audio 260-2-29A : TPA3255 / HA3588 CoilCraft / PFFB Fully differential.

Hi amigos, I imagine some of you have had the opportunity to test the latest 3E Audio TPA3255 module with PFFB + HA3588 Coilcraft inductors I refer to version 260-2-29A dated July 2022 ) Designed with Fully differential signal chain from input to output. This module seems to be close behind...

audiosciencereview.com

After some research I found a table for the TPA3255 which seemed more coherent and above all more detailed to me and this could help you :
BTW > I had anticipated this data a little when I carried out my "Bamboo" project) No need to use a 48/10A in Mono mode )

 

[ngk275]

Thanks for sharing the table.

3e audio TPA3255 260-2-29A Amplifier Review

The new Wiim amp is based on this chip. Let's hope that the implementation was excellent as well. It likely has no PFFB, they would have said so as loud as they could. I suspect it's performance will be below this module, maybe somewhere like a Fosi V3.

audiosciencereview.com

This post might give me a clue. He mentioned that after clipping point, the amp run out of current. Still lower than 1% TDH+N at that point.
So maybe after that the data is just calculated by the testing device?
If so, at clipping point the amp can output 144W per channel with 48V12A SPU which makes sense

 

[MCH]

As we have demonstrated several times and it is quite classic and known with the TPA3255 chip, it shows a better THD+N figure at lower voltages.

We discussed and measured this here :

3E Audio 260-2-29A : TPA3255 / HA3588 CoilCraft / PFFB Fully differential.

Hi amigos, I imagine some of you have had the opportunity to test the latest 3E Audio TPA3255 module with PFFB + HA3588 Coilcraft inductors I refer to version 260-2-29A dated July 2022 ) Designed with Fully differential signal chain from input to output. This module seems to be close behind...

audiosciencereview.com

??
Amir 48V: 102 dB @5W
The link you posted 34V: 103 dB @5W

Does not seem like a huge difference to me.

 

[daniboun]

??
Amir 48V: 102 dB @5W
The link you posted 34V: 103 dB @5W

Does not seem like a huge difference to me.

I agree it is not a huge difference ) Keep in mind that 103 dB @5W is for the Stereo module @34V (from Amir we get 101db @5W @48V)
I was talking about the TPA3255 chip in general. Afterwards, if we rely on the different measurements and the different manufacturers who offer amps based on TPA3255, it's not always easy to see clearly, I admit.

 

[adLuke]

Thanks again to everyone for clarifying these basic concepts.
I would like to do an exercise to see if my understanding is correct.
The two constraints that have to be both simultaneously be respected are:
1) Volts^2/R=max power
2) Amp^2*R=max power
If I want to look for a power supply that goes perfectly with my amplifier board i get:
Nominal Voltage of SMPS (say 48v). Divide that by Sqrt(2) (so I get 48/1.414=33.95v).
So this voltage will be able to output a maximum Power of 33.95^2/R in watts. i.e. 1152/2=576w into 2Ohms; 1152/4=288w into 4Ohms; 1152/8= 144w into 8Ohms.
But there is also a limit in terms of current. So if my SMPS has 10 amperes of current for each speaker, then it will be able to output a maximum power of: 10^2*R, i.e. 100*2=200w into 2Ohms: 100*4=400w into 4Ohm; 100*8=800w into 8Ohm.
Considering the 2 limiting factors in conjunction will give me: 200w into 2Ohms speakers (current limited); 288w into 4Ohms (voltage limited); 144w into 8Ohm (voltage limited).
So I could draw my conclusion that I might as well get a lower current PSU if my speakers are not low impedance.
Any part of that is wrong?

 

[daniboun]

So if my SMPS has 10 amperes of current for each speaker,

As per TI spec the TPA3255 can admit max 14A.... Could be a bad idea to go over

 

[sarieri]

Hi daniboun,

I bought a 3eAudio 260-2-29A a couple of months ago. I am very impressed with the quality of their engineering and very much like their 'hot plate' heatsink approach vice the standard approach; even before you bolt the plate to your chassis, the 260-2-29A runs much cooler than the heat fin types. I removed the hot plate and cleaned the heat paste from the TPA3255 IC and I have great confidence this is a real TI device, unlike some chips I've seen on the $20 TPA3255 boards.
I measured the 260-2-29A THD+N at 5watts into a 4ohm dummy load:


View attachment 309909

Really good to see it is bang on spec at 103.2dB sinad. Right up there in the top tier in Amir's League table of power amps. If you can use a soldering iron and a screwdriver it gives the best bang for buck of any amplifier, I believe. I use it with a Meanwell LRS-350-36 turned down to 34V for best sinad.

I use this with a 2-way active loudspeaker (Siegfried linkwitz's LXmini) and as such have no use for more than 100watts into 8ohms maximum output - four channels with 100watts each, that's more than enough power for my ears! So. I reduced the 260-2-29A gain from 20.5dB to 16.5dB and now get 105.8dB sinad 5watts into 4ohms. Now I'm within 0.4dB of the sinad performance of the Trinnov 8m at US $9,500 for 8 channels vice US $130 for two channels (3e amp plus Meanwell psu plus an old biscuit tin to house its cool running self!). DIY audio is the best: hours of fun; exercises the brain; saves a fortune and the music sounds tip-top!

So my advice is that if you retain some handywork skills after the 'Great Global Outsourcing Event' then take the 3eAudio DIY route; it's got the quality, the performance and the price!

View attachment 309908

Have you tested what is the max power output when power supply is only 34V? According to Ti, it’s likely not great.

 

[adLuke]

As per TI spec the TPA3255 can admit max 14A.... Could be a bad idea to go over

Yes, Sorry I was picking numbers at random to make my example.
I did not consider the 3e specs.
However this gives me a chance for another question.
With household electrical appliances we care to make sure the voltage is right and that the socket can provide at least the amount of current required. More is better.
In your reply you mentioned that the 3e board can take only 14 amps: does it mean we cannot have a psu that could provide more current than that?