Purifi Class D

[jaytor]

I'm interested to hear what you think about the Weiss op amps. I recently completed a DIY build of a pair of monoblocks using the Purifi modues with Neurochrome buffers. I decided to go with a large linear power supply because I was disappointed with the performance of my last class D amps (Nord NC500 with Hypex SMPS), and also since I could build it myself. So far, I'm liking the Purifi amps much more than the Nords. Much better dynamics and sound staging, and a smoother, airier midrange and high-end (although my old ears don't hear high frequencies like they used to). The amps currently have a few hundred hours on them, so they may still have a bit more break-in to go.

 

Each monoblock has a 1500VA custom Toroidy transformer with approximately 200,000uF of filter caps (CRCRC for high voltage rails, CRC followed by low noise regulator for buffer stages). 

 

The Neurochrome buffers are very neutral sounding with very low noise and distortion, but I'm wondering whether an input buffer that adds a little more character would be preferable (I like to experiment). So I'm considering building (or buying) an input buffer with the Weiss op amps, or maybe VTV's tube buffer. 

[jaytor]

Thank you. I have read from other designers, and found from my own experience, that using a power supply that is oversized for the amplifier (or preamp) provides an improvement in sound quality, particularly in dynamics and less congestion when the music has a lot of mid bass. It's probably not a significant difference, but since the Purifi module can deliver 25 amps before the current limiter kicks in, I don't think that it is terribly overdesigned. I had originally planned on using a 1200VA transformer, but the 1500VA Audio Supreme transformer from Toroidy is the same size, so I decided to go with the larger capacity. 

 

The transfomers have five secondaries - 2x45V and 3x15V. The latter for the input buffers and the gate drive, so not all the current is allocated for the high voltage rails. 

 

I am currently using the Neurochrome input buffers. This is a well designed op-amp based design, and I am also using the plug in linear regulators using AMB Omega 78/79 regulators for the buffer stages. 

 

I have not tried the Micro Audio SMPS. My only experience with using an SMPS with class D is with the Nord NC500 which uses the Hypex supply. I tried both the Sparkos and Sonic Imagery Op Amps. I really liked the black background, good bass punch and excellent detail, but they just sounded kind of sterile to me. Dynamics and sound stage were not as good as my Krell FPB300 or Parasound JC5. 

[jaytor]

@fas42 - yes, I agree. But commercial designers have to optimize performance for the cost and, while I believe that a high capacity power supply has definite SQ advantages, it’s not as clear that the improvements would justify the cost increase if building a commercial product - at least as far as I have taken it.
 

But since I only have to please myself, my general philosophy is to cram as much power supply into the box as I can make fit. 
 

@RickyV - I’m using a preamp I built myself. It uses a Khozmo stepped attenuator with an opamp output stage. I’m very pleased with the clarity and detail (it was a substantial improvement over my Parasound JC-2), but it is also fairly neutral sounding. I’m in the process of building another preamp which will use a fully discrete output stage with a high bias single ended class A output. 
 

I’ve had tube preamps in the past and the slow deterioration in SQ drove me nuts. I think I’ll be able to dial in the sound of the Purifi amps to my liking with a few more tweaks (or perhaps with just a bit more burnin). 
 

The power supplies for the high voltage rails are built using two Daisy chained universal power supply boards from the diyaudiostore. The low voltage supplies are built with two boards I bought from Jims Audio, again Daisy chained, but with resistors between the two boards since these boards aren’t designed with a CRC topology.  

 

[jaytor]

9 hours ago, hopkins said:

The audiophonic Purifi model has 3 gain settings on its interface card: https://www.audiophonics.fr/en/blog-diy-audio/36-set-gain-bypass-on-purifi-amplifier-module-ampli.html

If you can live with only 12db gain from the amplifier, would it not make sense to use the "bypass" mode which avoids the use of opamps entirely ?

The Purifi modules have an input impedance of 4.4K across the differential inputs (2.2K to ground). The general rule of thumb is that you want to have a source output impedance that is at least 10 times lower than the input impedance. So if your source has a differential output impedance under about 450 ohms and plenty of drive, and you are using short interconnects, you might be happy with the results. 

 

I tried bypassing the input buffer, and even though I have a low 66ohm differential output impedance from my preamp, I thought the sound was more dynamic with more weight with the input buffer active. However, I do have long balanced interconnects (~8M). So your results may be different. You won't hurt anything by trying it. 

[jaytor]

I listen to my system about eight hours a day since I've been working at home for the past year and my desk/computer are at the back end of my listening room. I've been quite pleased with my Purifi based amps. I also have a number of class A and A/B amps that I've compared them to including a couple older amps (Krell KAV-250a, Bryston 4B-ST), a Parasound JC5,  DIY FirstWatt F5 turbo monoblocks, and Neurochrome Mod-286 monoblocks. The Purifi amps compare favorably to all these. 

 

For the last couple weeks, I've been listening to the Purifi amps with VTV buffers that utilize a 6922 dual-triode feeding a pair of Weiss discrete op-amps. At first, I was disappointed with the sound compared to the Neurochrome buffers, but I'm starting to enjoy the sound. I'm not sure if I'm just getting used to them, or if they are still breaking in. I have around 100 hours on them now. Unfortunately, it's a pain to switch back and forth since they use different connectors. After I put another 100 or so hours on them, I may switch back to give the Neurochrome buffers another try. 

 

Unless the VTV buffers continue to improve a fair amount more, I'd have to say that the Neurochrome buffers are a much better value. But I'm also considering trying a different set of tubes to see how much difference that makes. 

[jaytor]

16 minutes ago, John Hughes said:

Well tubes can make a huge difference, but I do trust VTV to have selected some that match well (he's been in the tube business for decades).

Yes, I suspect you are right. It comes with Genalex Gold Lion tubes which seem to be fairly well regarded for a new production tube. 

[jaytor]

I built a pair of monoblocks with a (way overkill) linear power supplies incorporating a custom Toroidy 1500VA transformer, 134K uF caps for the high voltage rails, 54KuF for the low voltage rails. This is for each monoblock. I have experimented with a couple of different input buffers. I currently have the VTV tube input buffer installed which uses a single 6DJ8 and a pair of Weiss discrete op amps. 

 

The amp sounds very nice, but doesn't quite match either my Pass XA60.8s or my DIY PSET 300B monoblocks which use the Western Electric reissue 300B tubes. I've had the 300B amp in my system most of the time over the past year - definitely my favorite.

 

I have another Purifi stereo amp built with an SMPS and Neurochrome buffers. The monoblocks have a deeper sound stage, more natural decay, and a bit more meat on the bones in the upper bass (most obvious with male vocals, saxophone, etc.). 

[jaytor]

On 12/26/2020 at 10:20 AM, RickyV said:

I wondering about transformer based class-d amps. So not only for the input stages but also for power. Is somebody doing that? 600va transformer, rectifier and big capacitance. 

And the question would be if that would be better then the smps?

 

I built a pair of DIY monoblocks using the Purifi modules and a large linear power supply. The power supply is WAY overkill, but I've found in my other amplifiers (both DIY and commercial), that oversized power supplies generally sound better. My DIY amps have a 1500KVa custom toroidal transformer, and employ 134,000uF in the high voltage supply in a CRCRC configuration. The low voltage supplies have another 54,000uF. 

 

I am currently using a tube-based input buffer from VTV with Weiss discrete op-amps. Prior to using the tube buffer, I tried a simple discrete op-amp buffer using the Sparkos opamps, and the Neurochrome buffer. I slightly preferred the Neurochrome buffer over the Sparkos buffer - a bit cleaner in the midrange. In other respects, the these two buffers sounded slightly different in tonal balance, but I didn't really have a preference. Imaging, soundstage, dynamics, timing were pretty similar. 

 

The tube based buffer adds a bit of richness to the presentation with better soundstage depth and a bit more forward midrange (which I like). One of the things I look for is how smoothly notes decay. On some amps, the notes seem to abruptly stop instead of smoothly decreasing. This amp with the tube buffer does a really nice job. 

 

I also built a stereo purifi amp using the standard Hypex SMPS and the Neurochrome buffers (with low-noise regulators). This is a nice little amp that I use in my bedroom system. Compared to the monoblocks with tube buffer and linear power supply, it's pretty easy to hear the differences. It is nice and clean sounding, but lacks the mid-bass energy, huge soundstage, and more natural decay of the monoblocks. 

 

Until recently, I owned a pair of Pass XA60.8 monoblocks. IMO, the Pass amps were a bit better sounding, but not by much. It's hard to put my finger on it, but they just had a richness to the music that I preferred, with a bit more meat-on-the-bones in the mid-bass. 

 

My current favorite amps are a pair of DIY 300B P-SET monoblocks. My current speakers are quite efficient (at 98db/w) and I use separate powered woofers for the lower few octaves (crossing over at about 170Hz), so the 300B amps are a great fit. They just have a naturalness of tone and incredible detail that the Purifi amps just can't match (neither could the Pass amps). 

 

EDIT: I just realized that I had already posted about these amps last fall. Sorry for the duplicate content. 

 

[jaytor]

It is nice to have the option to be able to tailor the sound with different input buffers, particularly for a product designed for the DIYer. Although the all-in-one approach is certainly simpler for the less-experienced builder and is more likely to deliver a consistent result. 

[jaytor]

I've been following this recent discussion with some fascination. I've been an audiophile for close to 50 years, and was an electronics engineer in my early career (I've run a modest e-commerce business and developed web software for the past 20+ years).

 

In the last several years, I've become an avid DIYer - building a dozen or so amps, preamps and speakers. I do a lot of SPICE simulations and use test equipment (scope, distortion analyzer) during prototyping, but my final arbiter has always been listening.  

 

When I first started out as an audiophile, it was all about the measurements - these were the days of Julian Hirsch and his compatriots. But I soon discovered that I preferred the sound of equipment that was not always the best measuring. 

 

I am not a believer in the "Absolute Sound" as espoused by Harry Pearson, etc. I don't really care if the music sounds exactly the same as it did in the original recording venue. I want the music to move me. It's not an analytical response - it's an emotional one. I've often found that equipment that has the best "measurements" sounds, at best, boring, and, at worst, fatiguing. Some "distortion", whether it be frequency response variations or 2nd and 3rd harmonics, can be very pleasing to my ear, and make the music much more enjoyable. A lot of global feedback can often result in better measurements (particularly if you're just measuring THD and noise), but far worse sound.

 

In my DIY designs, I've found that different parts can make a big difference in my enjoyment in the music, even though they measure virtually the same. Since I'm not building my equipment to make a profit, I don't really care if the cost is justified on a pure price-performance curve. If I like the way it sounds, that's all that counts. For example, some of the copper foil in oil film caps sound a lot better to me than cheap film caps, let along electrolytics, yet they have very similar parameters (although electrolytic caps tend to have higher ESR and dieletric absorption). 

 

So, while I have a background in electrical engineering, I generally fall much more in the subjective camp than the objective one. I don't discount measurements - they are an invaluable tool for design and production, but, in my opinion, they are not the final arbiter in sound quality.

 

[jaytor]

I really wish that it was easy to tell how something was going to sound based on the published measurements, but I have not found that to be the case.

 

For example, I had the Orchard Audio GaN amp on home trial for a week and compared it to my other amps. I had a couple audiophiles over where we compared the amp to my other amps. In my system, the Orchard amp sounded like I had heavy drapes in front of my speakers. The soundstage was shallow and recessed, and overall sounded somewhat muted. Based on its excellent specs and many positive reviews, I'm sure that this amp can sound great, but it sure didn't in my system. 

 

By far my favorite amp in my system is a 300B SET amp. This amp has far worse objective measurements than my Purifi monoblocks or even my Pass XA60.8 class A monoblocks (now sold), but has more detail, better tone, and a deeper, wider soundstage. 

 

Of course, these are subjective opinions, but shared by many of my audiophile friends that are familiar with my system.

[jaytor]

21 minutes ago, fas42 said:

 

Because what one hears is the system, not a component - there is not a "good car tyre!", but there is a tyre which best complements the engineering of the vehicle's suspension and drive train, at that moment ... that's the way I think about getting the best from a system.

Yes, I agree, which is why I have found published specs next to useless in determining the components I prefer. 

 

That said, I build a lot of my own equipment and use measurements a great deal to help refine my designs and verify my builds. But this is after using sound engineering in the design. 

[jaytor]

This hobby is extremely subjective - everyone hears things somewhat differently and has different preferences for what they like. Telling people that what they hear must be confirmation bias is extremely disrespectful and rude. Just because you'd prefer to rely on test measurements instead of your own ears doesn't mean everyone else feels that way. 

 

I would love it if I could figure out how something was going to sound just from simple measurements, but in my 50 years as an audiophile, I have not found that to be the case. Many of the products that I have tried that appear to have stellar measurements sound, at best, boring and uninvolving, and at worst, dull or fatiguing. And this includes equipment I have built myself.

 

Measurements are a very useful tool for identifying design or manufacturing problems, and I use test equipment a great deal when I am designing and building my own equipment, but I have not found them to be a substitute for listening. 

 

This thread was (is) supposed to be about the Purifi class D modules and amps built using them. If you want to start a thread to discuss your beliefs that most audiophiles are delusional and wasting their money by using their ears, by all means go ahead, but please let's get this thread back to its original topic.