Purifi class-D amplifiers

[ergo]

Not sure if we'll be so lucky and this comes available as DIY module also, but it sure is interesting and from people with great track record.

https://www.audioxpress.com/news/purifi-audio-promises-to-reduce-distortion-in-speakers-and-amplifiers

https://6moons.com/audioreview_articles/purifi/

https://6moons.com/wp-content/uploads/audioreviews/purifi/1.pdf

https://www.monoandstereo.com/2019/02/new-lyngdorf-class-d-amplifier.html

[Evil Twin]

Impressive... most impressive.

Unquestionably these tiny technological terrors could have an outsized impact in some sectors...

The Force is strong with Bruno and others on this team... I think we will see for the first time that Bruno is truly collaborating with peers now...

[JonMarsh]

Very interesting.

Two observations-

• Notice the gain level, and the implications for the input necessary to achive rated output. This only makes sense in order to achieve THD+N at this level, N must be very low- and level must be higher at the input. This is a pro audio input/output gain structure setup. My Benchmark preamp wouldn't have any problem with this.
• THD is now falling into a level I have only seen approached with amplifiers like my Halcro DM68's. "Triple oh two" is getting down there, and you need a unit like the APx555 to measure reliably at that point.

[ergo]

The AudioXpress news also gives a hint
"At the company's first presentation they will be unveiling the PTT19-W02-03 ultra-low distortion 6.5" long-stroke woofer (Xmax +/- 14.7 mm), rated for 4 Ω nominal impedance. Purifi says this new driver is optimized for excellent transient behavior and is able to reach high SPL without distortion, thanks in great part to moving parts designed with low internal damping."

That'll be interesting too. I'm wondering if the long goal would be some amp/speaker driver combo. Lars Risbo has bee working with TI SmartAmp technologies with real time 'IV Sense' that can keep small woofers from overdrive etc. It does not say anywhere thought that such interaction would be in between these two new products, but as Evil Twin says 'Force is strong with these guys'

[ergo]

As for gain - indeed. At least I would need to either mod by Bruno Putzey's Balanced Preamp or then go for some other design that could pull that much pre-gain off.

[5th element]

They have comparable gain to my TI post filter feedback tpa3250s. One of the design benefits of that is lower noise.

These seem like amazing amplifiers it does bring up implications of cost mind you.

[kimmosto]

Many speakers need different gain for tweeter, mid and woofer. This kind of modules should have switch which affects to both gain and noise voltage at the output. It would be useful with sensitive (115 dB/2.83V/1m) drivers without requirement of passive components for noise reduction. For example 14 dB and 26 dB would be quite adequate gain options. Hypex FA for 3-way has different nCore module for tweeter. That is usually okay. At least more okay than e.g. fixed 26 dB gain such as with Pascal Audio S-PRO (which I really like but output noise is way too high for compression drivers and PA mids).

[Evil Twin]

They are using this design concept in a new Lyngdorf 8 channel amplifier, but the physical realization is not in modules as shown in the PuriFi data sheet.

Looking at the module render, I think the physical construction is not very different from nCore. Note that the output level is somewhat lower, even compared with the nCore NC500 module; I think they are most likely using pre-clipping/limiting like the class D modules from Merus audio, who was bought last year by the company I work for in my day job. Unusual compared with Hypex modules, the power spec is essentially at clipping, like many other Class D modules, approaching 1% distortion.

The performance should give them an opportunity to charge more, but on the surface it doesn't look like they'll be significantly more expensive to manufacture than nCore modules. Same RM core inductor concept, same 6x WIMA caps in parallel for the output filter. The secret sauces is in the gain structure and feedback loop, I think.

I did advise our own teams about this development, as we have been developing GaN HEMTs focused on the Class D market.

[ergo]

https://www.purifi-audio.com/press/

"May 8th, 2019

NAD Electronics and PURIFI Announce Cooperation Agreement

NAD will license PURIFI’s patent-pending amplification technology to create next generation amplifier platform..."

[ergo]

And more good news

[JonMarsh]

And more good news

[ATTACH=CONFIG]29267[/ATTACH]

Hmm, as a moderator, am I allowed to say, "Hot Damn!!"

:T

I saw the NAD announcement yesterday and forwarded it to a few friends...

[ergo]

And some more news from Puriti

A new speaker driver data here
Whitepaper here

Very unique looking surround on that driver. The 2nd harmonic THD has a rise at 2k, but I guess there need to be some tradeoffs as the Xmax in turn is very big and some other parameters look very good.

[Zvu]

Regardless of what they say, that smells to me like very good pure woofer driver, not midwoofer. Really looks a lot like Denovo Anarchy 704 (costs 66$) but Denovo being a little more potent with +/-12.5mm Xmax.

https://www.diysoundgroup.com/anarchy7-4ohm.html

Measurements available at John Krutke's blog.

http://www.zaphaudio.com/temp/EX-Anarchy-HD.gif

http://www.zaphaudio.com/temp/EX-Anarchy-FR.gif

http://www.zaphaudio.com/temp/EX-Anarchy-CSD.gif

https://i.postimg.cc/7hPgkQrj/Ex-Anarchy3.jpg

https://i.postimg.cc/xC2vs1xf/Ex-Anarchy2.jpg

[JonMarsh]

I'm thinking an MTM with a large wave-guide loaded tweeter designed for a low crossover point.,..

These specs are somewhat different from what earlier blurbs implied. But it's an interesting looking driver- I like the impedance curve, too, pretty well controlled inductive rise.

Might be great for a mini- Ardent. I'm in agreement that three way applications seem more reasonable.

Bottom line is the price and availability.

[AJINFLA]

I'm not seeing an Anarchy beater. At all.

[5th element]

I'm most impressed with the drivers very low levels of 3rd harmonic distortion. We're talking Satori levels here.

The rise in 2nd harmonic coincides with cone breakup at 2kHz. It is described as being a mid/bass and it would seem it's definitely targeted at 2 way use.

Given the no nonsense design methodology of Bruno and taking cues from his Kii loudspeaker he would clearly not use this beyond where the driver started to beam and would probably match it to a tweeter with some control over its directivity. So just as Jon said, with a waveguide. As a mid/bass it's fine for that.

The surround idea basically looks as if they've attempted to make half of it a forward roll surround and half of it a rear roll. No doubt so the surround non linearities in the forward vs reward going motion can make a stab at cancelling out.

The higher order midband distortion looks to be an order of magnitude better than the anarchy but the price is where things will really be determined. Not to mention there's the reckhorn sub driver if you want just bass and that's dirt cheap.

It does make for a decent proposition on a small active speaker with lots of watts, a low xover point and suitable tweeter. If the mid band distortion is to be believed it would sound fantastic and have the ability to churn out lots of bass. It does have a pretty low qts though, has anyone simmed it ported? I get the impression that it might not optimise into giving too much extension without an ebs alignment. Maybe the idea is to use it sealed with eq and lots of those class D watts.

[tktran]

Looks pretty good to me-

A quick look with my trusty online box modeller-

https://speakerboxlite.com/

This 6.5” midwoofer can be squeezed into a 1/4 to 1/3 cu.ft cabinet for an F3 of 50Hz to 44Hz in a traditional vented box. But with a couple of passive radiators and DSP/Class D power on tap...

At 94dB@1m, the 3rd harmonic is -60dB (0.1%) from 100Hz to 2000Hz.
If the 3rd harmonic is at 35dB, that likely means that the 3rd and higher order harmonics would like be buried beneath the noise floor.

The most interesting graph is this one:

[wolf_teeth]

Also appears to be a Lyeco-made product, according to the frame used.

Later,
Wolf

[Zvu]

It models reasonably well in volume of 15 liters and tuned to 35Hz.

https://i.postimg.cc/KzWH0hMs/image.gif

Port resonance problem can be seen and while it looks severe, it can be absorbed or mitigated to some extent by placing the vent on the back side. Other problem is port noise and at 18W when port diameter is 50mm and length is 26.7cm with two flares it looks like this.

https://i.postimg.cc/k4wSxW2n/image.gif

For loudspeakers that i want clean bass from i keep port air speed bellow 14ms. Lets say that at 17m/s it becomes audible, and that you really don't want to go above 26m/s then maximum power you can let through these woofers is 39Watts. All this happens at 60% of Xmax, over that and chuffing becomes serious distortion source so the point of low distortion woofer with 10mm Xmax is lost.

If we make the port larger, let's say 68mm, to minimize port air speed, port resonance goes down to around 320Hz and good luck mitigating that.

https://i.postimg.cc/8zRLp7hm/68.gif

Again, the port induced distortion is too large to be neglected in my opinion.

I'd say that passive radiators should be used with these woofers or closed cabinets with appropriate low frequency boost. Then there are some advantages of low distortion that can be put to good use. But it would have to be quite a boost to get these results in closed cabinets (6-8dB boost at 40Hz actually). Distortion levels rise as well.

Thing that really puts me off for two way is 30 grams Mms, more than double from Revelator, Satori, Seas Nextel, Excel and Prestige woofers. Although Cms is 1.3 i am doubtful how would it sound. I guess that we'll just have to wait and see.

[wolf_teeth]

Upon closer inspection, did anyone notice one HD plot is -35dB at the 2k resonance, and the other HD plot is -45dB?
Wolf

[JonMarsh]

It models reasonably well in volume of 15 liters and tuned to 35Hz.

https://i.postimg.cc/KzWH0hMs/image.gif

Port resonance problem can be seen and while it looks severe, it can be absorbed or mitigated to some extent by placing the vent on the back side. Other problem is port noise and at 18W when port diameter is 50mm and length is 26.7cm with two flares it looks like this.

https://i.postimg.cc/k4wSxW2n/image.gif

For loudspeakers that i want clean bass from i keep port air speed bellow 14ms. Lets say that at 17m/s it becomes audible, and that you really don't want to go above 26m/s then maximum power you can let through these woofers is 39Watts. All this happens at 60% of Xmax, over that and chuffing becomes serious distortion source so the point of low distortion woofer with 10mm Xmax is lost.

If we make the port larger, let's say 68mm, to minimize port air speed, port resonance goes down to around 320Hz and good luck mitigating that.

https://i.postimg.cc/8zRLp7hm/68.gif

Again, the port induced distortion is too large to be neglected in my opinion.

I'd say that passive radiators should be used with these woofers or closed cabinets with appropriate low frequency boost. Then there are some advantages of low distortion that can be put to good use. But it would have to be quite a boost to get these results in closed cabinets (6-8dB boost at 40Hz actually). Distortion levels rise as well.

Thing that really puts me off for two way is 30 grams Mms, more than double from Revelator, Satori, Seas Nextel, Excel and Prestige woofers. Although Cms is 1.3 i am doubtful how would it sound. I guess that we'll just have to wait and see.

Certainly agree that's a problem- either a passive radiator or a more radical port design than a tube is required.

[tktran]

Upon closer inspection, did anyone notice one HD plot is -35dB at the 2k resonance, and the other HD plot is -45dB?
Wolf

One is 2.83V/1m

The other is 94dB/1m...

A higher drive level to reach 94dB will raised distortion products?

[wolf_teeth]

This still will not typically vary by 10dB. No matter the level unless swamped by the noise floor, it should remain at the same level down from nominal. The gap should be presumed equal, right? If not, then it has compression issues too.

Please correct me if I'm wrong,
Wolf

[5th element]

This could be a different driver that's being measured too, but the driver is nominally 87dB @ 2.83Vrms. So 94dB is 7dB higher in drive level. Harmonic distortion at breakup like this does degrade with drive level so that's not a surprise.

[ergo]

As it appears there is an interview coming to Computer Audiophile about Purifi, so opportunity to chime in with questions.

Questions For Interview With Bruno Putzeys & Lars Risbo About Purifi Audio

[bvbellomo]

I don't see any pricing or 1ET400A's for sale. Is this not yet available? Any idea what it costs?

[tktran]

I've just realised that amp has a gain of around 13dB. There's no way to achieve 400W/4ohms with a standard 2V output of a disc player, or let alone the more ubiquitous ~1-1.5V of portable players or DACs.

If my mental calculator is correct, that works out to be an input sensitivity of some 8-9V...

[ergo]

I don't see any pricing or 1ET400A's for sale. Is this not yet available? Any idea what it costs?

The other forum has info
"The "Eigentakt" should be available for "DIY audio segment" approximately end of the year, prices may be announced end of Q3/2019. A "standard" I/O board may be supplied as well. And they are working on therir own SMPS solution."
So a bit of time to wait and collect the funds.

[ergo]

I've just realised that amp has a gain of around 13dB. There's no way to achieve 400W/4ohms with a standard 2V output of a disc player, or let alone the more ubiquitous ~1-1.5V of portable players or DACs.

If my mental calculator is correct, that works out to be an input sensitivity of some 8-9V...

Yeah, will need a hefty pre-amp which has me pondering options that perhaps I should start research into suitable DIY options for achieving high gains and be able to provide such output voltages with comfort.

[Evil Twin]

Not having seen the architecture, I can only speculatively remind one that this sounds quite similar to the gain architecture of the core of an nCore module, before putting the high performance op amp front end on to make a complete amplifier.

It should be most interesting to see complete details when available...

[tktran]

Purifi audio 6.5” midwoofer datasheet updated.

Wowsers...Looks nicely improved...and ready for prime time.

[tktran]

Exclusive measurements of This new model vs ncore 400

As usual, looks like Bruno has further refined his designs.

Summary:
eigentakt > nCore
Available to DIYers
Stereo Pre-amp board available
Compatible with Hypex SMPS1200A400

For the OEM this low gain eigentakt allows for custom input buffers for market differentiation (eg. Customised valve preamp, discrete preamp). whilst the eigentakt does the (transparent) heavy lifting

Full details here:

https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-purifi-1et400a-amplifier.7984/

[dwk]

Yeah, not what I'd call a massive improvement over the NC400, but still a clear step forward. A basic NC400 setup (one spms1200 + 2xNC400) is pretty darn affordable (relatively speaking) with the current exchange rate, so it will definitely be interesting to see where these come in price-wise.

[JonMarsh]

Thanks very much for the link. Actually, not intending to be disagreeable, and truly it's all relative, I suppose, but in a technological sense achieving an order of magnitude improvement in the fundamental SINAD performance is a pretty massive improvement, IMO. Now the front end stage and approach will be even more critical. It will be interesting to see what people come up with, and this goes on my list of things I MUST do next year if I can find the time. My APx555 is drooling at the thought of having a switching amp to compete with the Halcro's.

[5th element]

As far as I can see it isn't an order of magnitude better. Noise levels are comparable and distortion is only of minor improvement when approaching higher power levels.

Of course this is the main area of issue with the ncore so it's nice to see that Bruno has solved it.

The other issue with the ncore, that seems to pop up, is the performance variability. Sometimes you see distortion down to 0.000x% and other times it only reaches 0.00x% The same is true for the distortion rise approaching 10kHz. Sometimes measurements show the distortion rise as quite severe Vs non existent. Measurement sampling rate aside a cover up of higher frequency harmonics isn't cool.

I do know that the various stock hypex boards have quite variable performance, which appears to be from variations from the input circuitry. At least according to the datasheets.

It's interesting that one bare/basic ncore goes down to 0.000x but others only 0.00x. The one with the highest performance is the OEM nc500 which doesn't come with an input gain stage. The Purifi under test is very much like the nc500 so this is probably an example of as good as the Purifi gets. This is no bad thing as it's superb. What will be interesting is to see is if this performance is replicated in other modules and when it's also tested by others.

[Zvu]

Price of midwoofer will be 450 euros for one piece.

I'm guessing that they've looked at Seas graphene and Scan Speak Ellipticor and decided this is a good price.

[tktran]

Have you seen it on sale? Or did you hear it from a little birdie?

With the majority of speakers sold these days (looking at consumer devices in stores both online and brick and mortar) my wild guess is 2-4” drivers accounts for the majority of the market; in 2019-2020 a 6.5” midwoofer has become a pretty niche market.

I guess the strategy is to go after the “high end” to establish supremacy. Bruno did say they they would be price competitive, and with Ellipticor 18WE going at €850, I guess one could argue it is. As long as they measure and sound better, then I guess they’ve achieved their goal. As for the consumer; the more competition, the better?

As a DIYer or small volume manufacturer I think the best standard values are the SB Acoustics standard line, except for <100Hz. But finding gems in the rough is one of the fun parts to being a DIYer, and my last build was Jim Holtz’s/Curt Campbell’s Statement II, which uses drivers all less than €100. Phenomenal value, but very few have space or spouses that could accommodate 200L worth of cabinets.

Looking forward to a SOTA 2 way in 2020.
Purifi/BlieSma?

[Zvu]

Guy from Netherland on diy forum wrote it.

I'd always buy two Satori MW16P-8, instead of one Purifi woofer (if it turns up that is the real cost), and made an MTM with elliptic waveguide. 260€/pair vs 450€/piece - i think value is on my side and it just might be that performance would be too - less power compression, no port distortion at higher volume, lower combined Mms of both Satoris than one Purifi, better distribution of low frequencies, larger cone surface (something that i can't measure but definitely has a subjective perceivable difference), less doppler effect at same volume... In 50 litre cabinet, true.

I'm not seeing ingenuity either - small square box with Mundorf AMT (remarkable) tweeter.

https://audiophilestyle.com/uploads/monthly_2019_05/purifi-bruno-lars-etc.jpg.43bc6291f702d4c5ff185e1a5ef91ad6.jpg

Or maybe there is something i don't see and only listening will tell. For now, i'm not seeing the value there.

[5th element]

That price is ridiculous. Then again the ncore modules were ridiculously priced too as DIY module things go. As far as parts costs go the ncore modules cost about the same as the UCD modules but they charged many times the price for one. It's not a surprise to see that they've decided to try charging a massive amount for the driver too.

It looks like it should cost $100.

As to the speaker they've built with it, that's no surprise really given that it's a prototype thing. If you're constantly making changing to the driver you don't want to go with complex boxes that you may be rebuilding regularly.

[Evil Twin]

Time will tell if your intuition is correct, but the reasoning seems quite sound...

While it is a national theme in this country, that, “In God We Trust”, the wise also note that, “All others we verify”. Bruno is good, but he is not God.

[Zvu]

That price is ridiculous. Then again the ncore modules were ridiculously priced too as DIY module things go. As far as parts costs go the ncore modules cost about the same as the UCD modules but they charged many times the price for one. It's not a surprise to see that they've decided to try charging a massive amount for the driver too...

Price for a product is always more than value of bundled parts. I don't have a problem with that. Ncore NC400 was good example when it came out since you had to work really hard to achieve those parameters when making an amplifier and it would almost certainly cost you more than what 2xNC400+SMPS cost in Hypex shop.

This time perfomance offered is not on par with asking price since we can get same/better for less money, if we disregard the size of loudspeaker - but i really don't consider 50l volume as large.

.... “In God We Trust”, the wise also note that, “All others we verify”...

I paraphrased you couple of times. Some people really needed to hear that.

[Jonasz]

While it is a national theme in this country, that, “In God We Trust”, the wise also note that, “All others we verify”. Bruno is good, but he is not God.

I would never ever trust any so called god.

[ergo]

I guess being a new startup company they have weighed their options regarding pricing and set it to a high side first. One can always come down from that if numbers in first months do not meet the expectation. Going the other way is very much harder. So business wise I do understand the logic I think.
Would be interesting to try these out, but yeah with 450 a pop makes one think that perhaps going after them right at start seems a bit expensive.

[5th element]

It's a 6.5" driver. The world is your oyster when it comes to such devices.

Yes this driver does seem like a nice combination of specs. Very low distortion. No apparent cone edge surround resonance, although there is definitely something happening at ~400-700Hz. Needs more detailed measurements to see what, though the frequency response definitely wiggles about a bit. Huge xmax. Useable up high enough to mate with a tweeter.

My gut feeling is that if you want completely pistonic/resonance free operation go with the SB17NAC/BNAC/CAC. If you're okay with some soft cone properties pick a Satori.

The textreme Satori, if the marketing can be believed, should offer resonance free operation, within the typical passband, and without aggressive breakup later on. The cone materials are also very cheap. Carbon and epoxy?

If you absolutely need the xmax for a potential 'one driver does it all' speaker then it seems like a good choice. Otherwise choose a different mid/bass and use a small sub.

It makes me wonder what Bruno would use it for? A remake of his kii but with 4 of them? Front rear and sides? That would seem like a nice upgrade over the real basic peerless drivers he currently uses but that would completely screw the current speakers economics. I mean the peerless bass drivers are €30 each? Yeah replace those with €450 drivers.

[Zvu]

Those are about 14$/piece if you are buying over 500 pieces. And if you're using 8 woofers for pair of speakers, you will need a whole lot.

https://www.components-shop.com/product/Peerless-by-Tymphany/SBS-160F35AL01-04.html

In that 14$ there are all of the production expenses in making an alu cone 6.5" woofer with profit. Puts things in perspective how large the profit is for Kii Three. With Kii Three the idea is what matters to me. It is one of the reasons i've given up on hyped expensive drivers that are better than the rest by few tenths of percent but cost up to 10 times more. Bruno just did engineering job properly using what we consider (at very best) mediocre drivers and the result is phenomenal. I, and probably rest of the world, can't wait to see the midrange in it. Purifi driver is produced in Indonesia and costs out of the factory probably whopping 30-40$ a piece - and it is that expensive only because it is not made in tenths of thousands. So it might be profitable for him to start using Purifi woofers in updated Kii but i'm guessing that the end price for updated Kii would also be substantially higher. Not because it is justified, but because upgrade must not cost less.