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

[antcollinet]

@antcollinet >​

I confirmed my order. As I am French, I always go through Audiophonics, I want to have a good guarantee and efficient after-sales service.

Enclosure > round corners 272x212x60mm (I used it for many projects, It is of very good quality , easy to work with and nice design)
Module > 3E Audio 260-2-29a
Binding post > Acrylic isolated terminal blocks
PSU > MP-H250S48 with PFC (Seems to be the best alternative for consistent value for money, I got good feedbacks about it)
For binding post cables, connectors, IEC, AC filter etc, I use 2.5mm2 good quality copper cables + XLR Neutrik, (from my old Hypex amp)


See you for the next step)

BOM so far = EUR 280. What are you planning for line level connectors? RCA, or XLR?

 

[daniboun]

BOM so far = EUR 280. What are you planning for line level connectors? RCA, or XLR?

XLR =)

if you have looked at the module user guide. The SE mode (RCA) is not supported by default, you can however activate it but you need a good skills in soldering.
The best performances are achieved in full balanced mode of course

280€ is the high price because I go through Audiophonics, you can save 20% through Aliexpress but I wouldn't take the risk)
That being said, at €280 you have a nice amp which should last over time with a Sinad which would place it in the top 5 ASR

 

[S=klogW]

Please tell me what changes you made to the board to reduce the gain.

UAN,

I replaced the two Rin (200ohm)and the two Rf (1240 ohm) around the OPA1612s with new Rin =300ohm and new Rf = 1000ohm (used 1% 50ppm MELF 0204 resistors). Also the 4.7nF capacitor that's in parallel with the Rf need to be changed to 5.6nF, c0g(np0) MLCC capacitor to maintain the original filter bandwidth.

 

[S=klogW]

XLR =)

if you have looked at the module user guide. The SE mode (RCA) is not supported by default, you can however activate it but you need a good skills in soldering.
The best performances are achieved in full balanced mode of course

280€ is the high price because I go through Audiophonics, you can save 20% through Aliexpress but I wouldn't take the risk)
That being said, at €280 you have a nice amp which should last over time with a Sinad which would place it in the top 5 ASR

I agree with your top 5 because, apart from Topping LA90 and Benchmark AHB-2 which are Class AB, all the other amps in the top tier are simply re-badged, re-clothed Hypex/Purifi clones.

 

[S=klogW]

Perhaps we should email Mr. 3eAudio and warn him that when word gets around ASR about the Bang-per-Buck he's offering, he's going to need to up his monthly production quota substantially! Don't want him to be in the same boat as Mr. Okto Research was with his uber-desirable DACs.

So I ask the first question to the enthusiasts :

PSU choice : LLC or PFC ? 36V or 48V ?

1) Connex LLC SMPS300RS (V3.3) 36V spec > SPEC HERE

2) MP-H250S48 SMPS with PFC 48V spec > SPEC HERE (NB, Voltage can be set from 45,5V to 51V)

3) Meanwell RPS-500 PFC 36V spec > SPEC HERE

4) Meanwell LRS350-36 spec > SPEC HERE

I settled on the Meanwell LRS psu (it's one of many I had already in my "Garden (shed) of Earthly Delights"!), because I temporarily wired in a low noise LM317, 1.5amp regulator at 34V dc output for the 5watt test to reduce the noise output of the Meanwell and saw no consequent improvement in amp noise power. So the Power Supply Rejection Ratio (PSRR) of the TPA3255 must be high enough already and PSU noise (save for crazy-bad psu's) likely isn't a big concern with the TPA3255.

Credit: Hieronymus Bish-Bash-Bosch

 

[daniboun]

I agree with your top 5 because, apart from Topping LA90 and Benchmark AHB-2 which are Class AB, all the other amps in the top tier are simply re-badged, re-clothed Hypex/Purifi clones.

I agree ) thanks for the reminder

 

[daniboun]

Perhaps we should email Mr. 3eAudio and warn him that when word gets around ASR about the Bang-per-Buck he's offering, he's going to need to up his monthly production quota substantially! Don't want him to be in the same boat as Mr. Okto Research was with his uber-desirable DACs.

I settled on the Meanwell LRS psu (it's one of many I had already in my "Garden (shed) of Earthly Delights"!), because I temporarily wired in a low noise LM317, 1.5amp regulator at 34V dc output for the 5watt test to reduce the noise output of the Meanwell and saw no consequent improvement in amp noise power. So the Power Supply Rejection Ratio (PSRR) of the TPA3255 must be high enough already and PSU noise (save for crazy-bad psu's) likely isn't a big concern with the TPA3255.


Credit: Hieronymus Bish-Bash-Bosch

I opted for a PSU with an active PFC for the reasons explained here : https://hardwaresfera.com/en/articulos/tutoriales/pfc-activo-fuente-alimentacion/
In any case, the 4 PSUs mentioned above are entirely appropriate for this project.

 

[daniboun]

Perhaps we should email Mr. 3eAudio and warn him that when word gets around ASR about the Bang-per-Buck he's offering, he's going to need to up his monthly production quota substantially! Don't want him to be in the same boat as Mr. Okto Research was with his uber-desirable DACs.

For the record... it seems to me that 3E Audio had sent a copy to Amir a few months ago, but there was a mishap...
I don't know what happened, hoping that this thread can revive 3E Audio's' motivations )

 

[daniboun]

UAN,

I replaced the two Rin (200ohm)and the two Rf (1240 ohm) around the OPA1612s with new Rin =300ohm and new Rf = 1000ohm (used 1% 50ppm MELF 0204 resistors). Also the 4.7nF capacitor that's in parallel with the Rf need to be changed to 5.6nF, c0g(np0) MLCC capacitor to maintain the original filter bandwidth.

Kudos for the mod. Do you have some pciture to share ?

 

[daniboun]

Found this graph wtith APG2923 Coilcraft inductors: this model is a special customized version from 3E Audio.
(PSU :Morsun LOF550-20B48)

 

[S=klogW]

I opted for a PSU with an active PFC for the reasons explained here : https://hardwaresfera.com/en/articulos/tutoriales/pfc-activo-fuente-alimentacion/
In any case, the 4 PSUs mentioned above are entirely appropriate for this project.

As for the choice between 36V and 48V psu; running the TPA3255 at 48V will give you 220watts into 4ohms at 1% thd+n. The 36V psu will allow 112watts into 4ohms at 1% thd+n: twice as much power from the 48V psu, but our ears work in decibels and that's only 3dB more. Also, we don't listen at these power levels (MegaDeaf fans excepted), a comfortable level in the home environment is 5 -25watts rms per channel. That's why I prefer the 36volts psu choice: lower thd+n from using the lower voltage and things run cooler, hence last longer.

 

[daniboun]

As for the choice between 36V and 48V psu; running the TPA3255 at 48V will give you 220watts into 4ohms at 1% thd+n. The 36V psu will allow 112watts into 4ohms at 1% thd+n: twice as much power from the 48V psu, but our ears work in decibels and that's only 3dB more. Also, we don't listen at these power levels (MegaDeaf fans excepted), a comfortable level in the home environment is 5 -25watts rms per channel. That's why I prefer the 36volts psu choice: lower thd+n from using the lower voltage and things run cooler, hence last longer.

Right. Here a power chart for the TPA3255 to help :
Max Output power is achieved with a 480W / 48V PSU but best THD+N figure will done with a 36/40V PSU

 

[UAN]

I now have a cheap power supply from Aliexpress 38 V, 9A. Shelf noise 140 dB at 5 watts, the amplifier sounds very good.

 

[daniboun]

I now have a cheap power supply from Aliexpress 38 V, 9A. Shelf noise 140 dB at 5 watts, the amplifier sounds very good.

Nice thanks for that picture )
I used the same PSU in a previous project but its voltage was too variable with the multimeter...I imagine it must have been defective (
I see you are using TRS plugs, any reason not to ground it ?

 

[UAN]

This is an old photo, it's better to ground it.

 

[S=klogW]

Found this graph wtith APG2923 Coilcraft inductors: this model is a special customized version from 3E Audio.

View attachment 310039

108dB sinad at 5 watts -SWEET! Does it say if this is using a 4ohm or an 8ohm dummy load. At 8ohms, my mod hits 108.3dB sinad (because 5watts at 8ohms needs less current c.f. Amir's standard test of 5watts at 4ohms).

The other way to get these better figures is to drop the gain of the OPA1612 first stage op amp. I did this with my mod because I only need 100watts rms into 4ohms output (i.e. 20volts rms) and I only use a balanced DAC as my source at 4.22V rms max output. So I only need my amp to give a 4.75 times overall gain (13.5dB).

Going beyond this level of gain reduction is a bit of a cheat (we love you Benchmark AHB-2!) because at super-low gains (only 9.2dB Benchmark AHB-2), you get super-low thd+n (-113dB Benchmark AHB-2) for the unit under test , in isolation, but all you are doing is hiding the overall increase in thd+n that the complete audio 'chain' will inevitably show, by off-loading the bad news into the up-front pre-amp or DAC. So, I contend that this 260-2-29A is honest and isn't appreciably far from AHB-2 standards and a great choice if you don't want 500W rms in Bridge mode (and you only need that if you've got hugely inefficient transformers (electrostatic speakers) or passive crossovers in your monkey-coffin speakers swallowing gobs of power before the drivers get a sniff (Revel).

 

[S=klogW]

Right. Here a power chart for the TPA3255 to help :
Max Output power is achieved with a 480W / 48V PSU but best THD+N figure will done with a 36/40V PSU

View attachment 310052

I've seen this table somewhere before and at the time thought 'useful but not entirely accurate (was the marketing department the source?)'. I've not checked every entry, but for example, the entry 36V, 10A, 4ohm giving 190watts: 190W rms into 4ohms requires 27.6Vrms (if you believe Ohm's Law), that's 39V peak required (pk = rms x sqrt2), plus the TPA3255 needs another 5.5V for signal headroom (unless you want your lovely low distortion amp to produce 10% THdistortion+N, or more, in order to claim more output power, as suggested in the TI TPA3255, marketing department-influenced, data sheet - "Engineers don't get it, we need much bigger numbers!!").

So the 36V psu actually needs to produce 44.5volts - these type of 36V psu's I call 44.5V psu's!

 

[daniboun]

108dB sinad at 5 watts -SWEET! Does it say if this is using a 4ohm or an 8ohm dummy load. At 8ohms, my mod hits 108.3dB sinad (because 5watts at 8ohms needs less current c.f. Amir's standard test of 5watts at 4ohms).

The other way to get these better figures is to drop the gain of the OPA1612 first stage op amp. I did this with my mod because I only need 100watts rms into 4ohms output (i.e. 20volts rms) and I only use a balanced DAC as my source at 4.22V rms max output. So I only need my amp to give a 4.75 times overall gain (13.5dB).

Going beyond this level of gain reduction is a bit of a cheat (we love you Benchmark AHB-2!) because at super-low gains (only 9.2dB Benchmark AHB-2), you get super-low thd+n (-113dB Benchmark AHB-2) for the unit under test , in isolation, but all you are doing is hiding the overall increase in thd+n that the complete audio 'chain' will inevitably show, by off-loading the bad news into the up-front pre-amp or DAC. So, I contend that this 260-2-29A is honest and isn't appreciably far from AHB-2 standards and a great choice if you don't want 500W rms in Bridge mode (and you only need that if you've got hugely inefficient transformers (electrostatic speakers) or passive crossovers in your monkey-coffin speakers swallowing gobs of power before the drivers get a sniff (Revel).

If I am not wrong 4 ohms sorry ) As I said on page 1, the reference PSU @ 3E Audio is the Morsun LOF350-20B48-C with PFC
I completely agree with your observation for the AHB2 Benchmark. Let's say that the 3E Audio module is better suited for moderate listening.
But we must put it into perspective) The AHB2 costs 16 times more expensive)

 

[daniboun]

I've seen this table somewhere before and at the time thought 'useful but not entirely accurate (was the marketing department the source?)'. I've not checked every entry, but for example, the entry 36V, 10A, 4ohm giving 190watts: 190W rms into 4ohms requires 27.6Vrms (if you believe Ohm's Law), that's 39V peak required (pk = rms x sqrt2), plus the TPA3255 needs another 5.5V for signal headroom (unless you want your lovely low distortion amp to produce 10% THdistortion+N, or more, in order to claim more output power, as suggested in the TI TPA3255, marketing department-influenced, data sheet - "Engineers don't get it, we need much bigger numbers!!").

So the 36V psu actually needs to produce 44.5volts - these type of 36V psu's I call 44.5V psu's!

You re probably right.
Source : http://archimago.blogspot.com/ well known also for their serious reviews and measurements.

 

[antcollinet]

XLR =)

I have the balanced MiniDSP flex, so if I build this thing, I'll probably put some TRS connectors in.