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

[antcollinet]

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!

Bear in mind that 190W is already with 1%THD - I've no idea how flat the top is of a 1% clipped sine wave. It might be enough to create the extra power.

 

[S=klogW]

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

Not the neatest work, but I'm sixty-one and both my eyes have had cataract operations - still happy, still hand soldering (albeit with a USB microscope).

 

[daniboun]

Not the neatest work, but I'm sixty-one and both my eyes have had cataract operations - still happy, still hand soldering (albeit with a USB microscope).

thanks for sharing
Hot air welding would help a lot in this case)

 

[daniboun]

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

You can also use XLR > TRS balanced cables, I am using some Cordial ones
If one day you intend to sell it, XLR connectors are more standard, it is just my opinion)

BTW : What about the Mini Flex DSP Balanced via Bluetooth input (LDAC) ? Is Dirac Live really a substantial option?
I'd like to know if we stream via BT from the Mini Flex would there be a major interest compared to a very good DAC like the X26 Pro?

 

[MCH]

cool projects guys,

Is anyone selling finished amps with this modules? if not, is it known why? price, reliability, too new...? somehow i find it suspicious if none of the known assemblers is using such a good performing bargain

 

[S=klogW]

Bear in mind that 190W is already with 1%THD - I've no idea how flat the top is of a 1% clipped sine wave. It might be enough to create the extra power.

Just pointing out that if you follow this table to the letter, you may end up disappointed if you measure its performance at high power levels. Would we, the truth-seekers of the ASR want an amp with so much potential to be compromised by over-driving it into the 1 percent to 'square wave' percent distortion region merely to chase a diminishing returns power output number? I say "No, brothers and sisters!" (forgive me channeling my father who was a Workers' Union Rep.) Better, I think, to honour the 5.5volt headroom and plug that into the P=Vsquared/R and P=Isquared.R equations and always have sinad less than or equal to 0.01% up to the stated maximum power. Specs that reflect the core qualities of a product, not Spec-manship, is what I enjoy seeing. Texas Instruments' 10percent distortion spec. -Poof! In my Best Hagrid voice.

 

[daniboun]

cool projects guys,

Is anyone selling finished amps with this modules? if not, is it known why? price, reliability, too new...? somehow i find it suspicious if none of the known assemblers is using such a goo performing bargain

I think the most obvious answer would be that this is a product for the DIY market, 3E Audio targets an informed audience here. Besides, I think it's more profitable for them to sell modules in this segment rather than finished products.
When it comes to reliability, 3E Audio has a pretty solid reputation. I have mounted several of their old modules and never had a problem. I can't say the same for Topping, a breakdown on my PA5 and another on my LA90....

 

[S=klogW]

cool projects guys,

Is anyone selling finished amps with this modules? if not, is it known why? price, reliability, too new...? somehow i find it suspicious if none of the known assemblers is using such a goo performing bargain

Good question, MCH(annel? - 7.1.4; 9.4.6? If so, you'll appreciate why I don't need more than 100watts rms per channel).
Every one should have a go at answering this, but I think that Hypex et al have got the reliable supply chain that the re-badgers need. I get the impression that Mr3eAudio, like me, works out of his garden shed as I watch him run his stock down on Aliexpress to nothing before he returns a week of so later with a new stock of 100 units. Maybe he needs a V.C. cash injection or a big group-buy order from ASR ...

 

[S=klogW]

Bear in mind that 190W is already with 1%THD - I've no idea how flat the top is of a 1% clipped sine wave. It might be enough to create the extra power.

Hobby horse time - I don't know why we still use a 1970's-applicable 1% THD+N (*) as the bar we set for maximum power output in this Brave New World of 0.001% THD+N or better power amplifiers. The 'knee' in these performant amplifiers can clearly be seen at less than 0.003% and the graph y-axis doesn't need to extend beyond 0.01% to tell us everything we need to know - that vertical take-off says the amp saturates to all intents and purposes at some 120watts (so likely this is an 8ohm load graph. Of the two figures attached, Fig.7 is the title to use is my guess). 1% and 10% THD+N limits are simply not appropriate for these high quality amps.

And by way of a value comparison: $90 3eAudio boards compared with what is considered stellar and desirable by some audiophiles, I recently repaired a guy's Spectral Audio DMA-100S stereo power amp - nasty circuit boards with tracks that lifted as soon as you bring a soldering iron near them; a pair of (obsolete) custom toroid transformers designed to fit the chassis height rather than the rated current required, hence one had burnt out (had to get seriously creative to fix that). After hours of work and frustration I got it working and measured it at 80dB sinad 5watts into 4ohms: back within Spectral's spec, but 20dB worse than if I could have just dropped a replacement TPA3255 board and a cheap smps into the pretty Spectral chassis instead! Price on eBay for a 30 year old Spectral Audio DMA-100S? Typically 3000 to 3500 Euros. Go figure!

______________
(*) Unless it's because without the 1% THD+N limit valve amps wouldn't achieve qualifying here on ASR.
________________

 

[daniboun]

And by way of a value comparison: $90 3eAudio boards compared with what is considered stellar and desirable by some audiophiles, I recently repaired a guy's Spectral Audio DMA-100S stereo power amp - nasty circuit boards with tracks that lifted as soon as you bring a soldering iron near them; a pair of (obsolete) custom toroid transformers designed to fit the chassis height rather than the rated current required, hence one had burnt out (had to get seriously creative to fix that). After hours of work and frustration I got it working and measured it at 80dB sinad 5watts into 4ohms: back within Spectral's spec, but 20dB worse than if I could have just dropped a replacement TPA3255 board and a cheap smps into the pretty Spectral chassis instead! Price on eBay for a 30 year old Spectral Audio DMA-100S? Typically 3000 to 3500 Euros. Go figure!

This is a good example that shows that we often pay a high price in the audiophile world to finally have poor equipment... This is obviously applicable for cables at completely prohibitive prices...

A good example here....

Nordost Blue Heaven AC Cord Review

This is a review and detailed measurements of the Nordost Blue Heaven AC Cord. It is on kind loan from a member and costs US $417 at this length (2m). This is "third-tier" product in Nordost line so naturally, it is not going to scream expensive. That is a good thing as the cable itself is kept...

www.audiosciencereview.com

 

[antcollinet]

thanks for sharing
Hot air welding would help a lot in this case)

My tips for hand soldering SM devices.

• Put a small amount of solder on ONE of the pads on the PCB. Try to remove solder off the other pad if there is any there.
• Hold the device in place with tweezers.
• Melt the solder on the pad (That side of the device sinks into the liquid solder)
• Remove iron and let the solder solidify.
• The device is now held in the correct position by the solder, so you can solder the OTHER side of the device, then when that is solid add some additional solder to the first side if necessary.

You can do the same with multi pin devices, starting by tacking down one corner pin using the solder on the pad technique, then soldering the opposite corner to fix the position. The rest are easy after that.

 

[antcollinet]

Bear in mind that 190W is already with 1%THD - I've no idea how flat the top is of a 1% clipped sine wave. It might be enough to create the extra power.

Having said that - I just found this in the datasheet. This suggests the table above is masively over optimistic. I make it about 130 to 135W at 36V 4ohm.

 

[S=klogW]

This is a good example that shows that we often pay a high price in the audiophile world to finally have poor equipment... This is obviously applicable for cables at completely prohibitive prices...

A good example here....

Nordost Blue Heaven AC Cord Review

This is a review and detailed measurements of the Nordost Blue Heaven AC Cord. It is on kind loan from a member and costs US $417 at this length (2m). This is "third-tier" product in Nordost line so naturally, it is not going to scream expensive. That is a good thing as the cable itself is kept...

www.audiosciencereview.com

$417 for a "third-tier" wire without gain. Wouldn't you have to be a sociopath to ask those prices and still sleep at night? and the end-user: how do you slide into the dinner party conversation that you paid that much for a piece of wire and consequently everyone else in the room had better be crushed and humiliated by your inate superiority. Emperor's New Clothes, n'est pas? Sante! (where are the acute accents on ASR/ UK keyboards when you need one?)

 

[antcollinet]

I don't know why we still use a 1970's-applicable 1% THD+N (*

It is not totally unreasonable. If run within that limit with real music, it will only be the most transient of musical peaks that are clipped. And at 1% that won't be audible.

 

[antcollinet]

BTW : What about the Mini Flex DSP Balanced via Bluetooth input (LDAC) ? Is Dirac Live really a substantial option?
I'd like to know if we stream via BT from the Mini Flex would there be a major interest compared to a very good DAC like the X26 Pro?

I'm currently bluetoothing my turntable to the Flex (just to piss off the vinyl/analogue purists ). Yes it is LDAC. Yes dirac can equalise the bluetooth signal.

I'm also in the camp that any DAC with sufficiently good measurements is audibly perfect. The flex qualifies here, so no need for any "good" or "better' DAC such as the X26 pro.

 

[MCH]

My tips for hand soldering SM devices.

• Put a small amount of solder on ONE of the pads on the PCB. Try to remove solder off the other pad if there is any there.
• Hold the device in place with tweezers.
• Melt the solder on the pad (That side of the device sinks into the liquid solder)
• Remove iron and let the solder solidify.
• The device is now held in the correct position by the solder, so you can solder the OTHER side of the device, then when that is solid add some additional solder to the first side if necessary.

You can do the same with multi pin devices, starting by tacking down one corner pin using the solder on the pad technique, then soldering the opposite corner to fix the position. The rest are easy after that.

Agree, the orientation of the op amps is perfect, what makes it an easy job.
(If they were 90° the white connectors would make it a bit more difficult)

 

[daniboun]

I'm currently bluetoothing my turntable to the Flex (just to piss off the vinyl/analogue purists ). Yes it is LDAC. Yes dirac can equalise the bluetooth signal.

I'm also in the camp that any DAC with sufficiently good measurements is audibly perfect. The flex qualifies here, so no need for any "good" or "better' DAC such as the X26 pro.

Makes sense, I just read the review from Amir. Really good unit !

 

[daniboun]

Having said that - I just found this in the datasheet. This suggests the table above is masively over optimistic. I make it about 130 to 135W at 36V 4ohm.

I don't know to what extent we will make a comparison by analogy with the Jlester Module with a Connex PSU @36V / 300W
But I guess this can give us an idea anyway...

Checkout :

JL ELectronics Sylph-D200 Amplifier Module Review

This is a review and detailed measurements of the JL Electronic's (JLE) Sylph-D200 stereo amplifier module built on top of TI's TPA3255 amplifier IC. Depending on the configuration you pick, the module costs US $130 to $170. You can read more about it here...

www.audiosciencereview.com

 

[S=klogW]

Having said that - I just found this in the datasheet. This suggests the table above is masively over optimistic. I make it about 130 to 135W at 36V 4ohm.

View attachment 310093

Yes, that's more realistic. Here TI are allowing a 2volt bite into the 5.5V headroom and so are setting the bar at 1% THD+N:

(((36V -3.5V headroom)/sqrt2)all squared)/4ohms =135watts rms Q.E.D.

I was suspicious of the previous table as it showed ever increasing amplifier power output when keeping the same psu voltage by merely changing to a psu with a higher current output capability. Ohm's Law doesn't work like that: e.g. my local electricity substation is capable of delivering thousands of amps, but at 230Vac the 26ohm resistance of my 2000watt electric kettle means it can only ever use 8.7Ampère of those thousands on tap - Thankfully! Otherwise:

ça veut dire:

​

At 36volts, a 6amp rated psu is more than enough to take a BTL TPA3255 channel to power levels where THD+N goes greater than 1%. Any more than 6amps will only take it into much more distorted power territory. If a person needs this much power and doesn't mind paying the price in increased THD+N then: a) at ASR, they are probably in the wrong place; b) they could consider instead the Behringer iNuke nu6000 2 x 3,100 Watts @ 4 Ohms ; THD+N: 12.13 pounds (Yes, I know that's its weight but it's the only other spec available on Behringer's detailed spec sheet!

 

[S=klogW]

I don't know to what extent we will make a comparison by analogy with the Jlester Module with a Connex PSU @36V / 300W
But I guess this can give us an idea anyway...

Checkout :

JL ELectronics Sylph-D200 Amplifier Module Review

This is a review and detailed measurements of the JL Electronic's (JLE) Sylph-D200 stereo amplifier module built on top of TI's TPA3255 amplifier IC. Depending on the configuration you pick, the module costs US $130 to $170. You can read more about it here...

www.audiosciencereview.com

I'd forgotten/dismissed the Jlester module as it's clearly an imperfect implementation of Post-Filter Feedback (PFFB) - only modest gains in thd+n over non-PFFB, even with total gain at a low 14dB so as to minimize the output noise power. Consequently this unit produces what is, even for me and my more modest requirements, a pretty anaemic 61watts into 4 ohms. In contrast, 3eAudio does it all correctly; daring to stray from the modest performance gains afforded by TI's cautiously suggested PFFB implementation and using so much feedback (Sorry, Nelson, feedback is a "Good Thing". It's why human's have touch sensors to feedback when their arms would otherwise be crushing the baby. Feedback is central to Control Theory: Control is a "Good Thing".) that the TPA3255 itself only has a gain of 2 times, the rest comes from its input buffer. Well done, 3eAudio! you did a veritable Bruno Putzeys on your high-order PFFB filter optimisation. (No I am not Mr/Mrs3eAudio, just a fanboy/girl/fanGrandad).