Jump to content

Art Vandelay

Member
  • Posts

    5,314
  • Joined

  • Last visited

  • Days Won

    18

Everything posted by Art Vandelay

  1. You're comparing oranges with apples. Bruno achieves very low distortion with a very high feedback factor, and multi-pole compensation to extend open loop gain to the top of the audio band. All things being equal he should be able to achieve lower distortion with GaN with less feedback factor. Not only are they a faster device but the much lower gate capacitance also reduces IMD Vs MOS. Of course there are other factors to consider and you could mount an argument that the performance is now at the point where further improvements may be barely audible, and probably difficult to achieve, regardless of topology or output device types. GaN also allows for a higher switching frequency. This comes at the expense of dissipation, efficiency, and hence output power, but I'm sure we'll be seeing wider bandwidth class D in the future, for the sake of marketing, if not engineering.
  2. I had it in use driving Soavo's in the AV system but since replaced with a Purifi 1ET400 stereo amp. It sounded fine and quite sweet for class D, though it is a bit power limited with normal efficiency 2-way speakers. Fwiw, the specs aren't quite as impressive as with Hypex, Purifi amps, but that's because they're a low feedback design, so distortion is close to linear with frequency, and nominally 0.01%. The GAN transistors will produce less distortion at high frequencies due to the reduced "dead time". Given that DT is a fixed quantity, it increases as a percentage of wavelength with rising frequency, so for low distortion it's switching speed rather than switching frequency that matters. IMO, probably the best low cost option for a compact low power class D amp is the Topping PA-5, if you can live with the 6.5mm connectivity.
  3. There's good and not so good examples, but that applies to all class types. Class A bias gets around all crossover distortion mechanisms - in typical complementary push pull class B / AB amplifiers. It achieves this by ensuring that both positive and negative halves of the push pull output stage are always dc biased (and conducting ). This (in theory) removes related high order distortion products, and also removes some slew-induced distortion mechanisms. From a subjective standpoint (actually mine), a well designed class A amplifier will sound more transparent and 'grain-free', and typically a better soundstage with more delineation of instruments in complex music passages. The downside of class A is power dissipation and heat, which led Krell to develop 'plateau bias'. Plateau bias simply moves the bias point up (and down) to meet a minimum level of bias current in order to maintain class A operation. This removes lots of heat dissipation if you're only listening at low to moderate levels.
  4. I have 4x NCX500 modules just arrived. The intention is to make a couple of mono's. I'm currently designing a discrete class A preamp that I'll be using in preference to the typical op-amp based input buffer. Spice sims look good thus far. It will eventually be housed in a separate enclosure with linear psu.
  5. Class D amps have improved hugely in the past decade but there's still a signature sound that you may or may not be able to live with long term. The Krell FPB series is still one of the finest, so definitely "relevant" imo. However, it's worth taking into account age and condition, and 20+ years is getting up there. I know of one FPB300 owner who had to spend a small fortune in repairs. One of the downsides of high bias class A, AB etc is that the heat does age components more quickly, and capacitors in particular, and failure or degradation of supply or rail bypass caps can sometimes cause serious and potentially destructive, costly failures. Krell amplifiers are equipped with last resort 'crow bar' type protection, which guarantees that speakers are protected at all costs, but that includes at the expense of the amplifier too. So I recommend Krell, but..... check the history, including the service history if possible. I would also recommend that you audition a Hypex or Purify based class D amp, preferably in your own system.
  6. Sure, not all preamplifiers are fitted with output relays, including many expensive DC coupled designs, so when powered up or down there can be momentary DC at the output, and if connected to a power amplifier with (typically) 25-30dB of gain it only takes 0.5 volts of DC to result in 10 Volts or more at the output. It's why it's always good practice to switch the power amp on last when powering up the system, and conversely to switch it off first when powering down. In the case of integrated amps it's usually not a problem because the relays at the speaker terminals will have the appropriate delays to accommodate the full stabilisation period.
  7. We can probably all expect to see more supply outages and regulation problems as intermittent renewables are phased in, and firming from coal and gas supply is phased out. There's definitely a risk to components and also speakers if the supply is regularly being switched on and off, and it goes without saying that it's a real nuisance too. My recommendation would be to install a UPS for your sources and pre-amplifiers as a priority, because power cycling can result in momentary DC being extended to the power amplifier. Power amplifiers are generally well protected, and in the case of larger class A/AB you might not even be aware if there's been a momentary outage or brief interval of low voltage, so it's less of a priority in most cases. Class D amplifiers don't have the same supply reservoir but they usually have excellent measures for speaker and self-protection. Fwiw, I recently added a Tesla battery and I've ridden through several outages without even being aware until notification via the app.
  8. Yes, definitely, although there are still limits due to thermal resistance with the output devices. The new Purifi certainly looks the goods though. 40A Vs 25A output current will definitely drive a 2 ohm load, and I'm sure that addresses the main concern with the 1ET400A. I have been using Hypex NC1200's to drive my 800 Diamond's woofers, but do still have a preference for class AB driving mids and tweeters. That said I will look forward to hearing 1ET7040SA based modules in the near future.
  9. Stereophile uses a metric they call 'EPDR' to express complex impedance. Yes, it looks torturous, but isn't actually a simple Polar-Rectangular conversion as many probably think. It's main purpose is to equate the resistive impedance for equivalent dissipation in the connected power amplifier. If you look at the latest iteration of Alexia (V), the impedance dips to 2.6 ohms at 86Hz, but with a phase angle of 15 degrees, the power amplifier effectively sees 2.5 ohms, which is still acceptable. Where Stereophile quotes EPDR of 1.15 ohms @ 66 Hz, the impedance appears to be 3 ohms / -45 °, which in 'R' form is 2.1 ohms. So you should be able to safely drive the latest Alexia's with the latest Purifi without invoking the current limiter, but of course that doesn't explicitly mean that the Purifi will drive them without shutting down due to thermal protection etc.
  10. Hi George, From the figures specified in the 1ET7040 datasheet, the DC rail voltage works out to be +/- 63 VDC. Noting that 'RMS' power is a marketing term, but we'll assume equal to 'maximum average power', based on the RMS voltage or current. This is different to 'peak power' which is an instantaneous value. Purifi does state a maximum current of 40A, based on the action of the current limiter. Assuming a resistive 2 ohm load, the maximum current through the load will be determined by the DC rail voltage, and @ +/-63VDC that equates to 22A rms, with a peak (instantaneous) value of 31.5A. Probably safe to say that the current limited output of 40A is unlikely to be reached or even approached in the real world, even with 'difficult' speakers.
  11. As always, buyer beware. Success in the audio industry is often more to do with marketing than performance. 'Negative feedback' was given a bad name back in the 70's and 80's on the back of amplifiers that measured well on the test bench but sounded harsh and sterile. Although much has been learned since, and the evils of NFB have been mostly eradicated, there remains a lingering curse, and hence the challenge remains for engineers to produce an amplifier that employs little or zero global NFB without compromising test bench performance. (Yes, Stereophile charts matter) But there are others reason too why zero GNFB is attractive for designers, particularly at the high end of the price spectrum. Few, if any, high feedback amplifiers are completely insensitive to capacitive loads, and most exhibit instability at or near clipping, and / or recovery from clipping. In short, it's much more difficult to guarantee freedom from a catastrophic failure or a misbehaviour that destroys very expensive speakers with high feedback designs, particularly with respect to very high power / high current amplifiers. The big Krell amps are actually designed to destroy themselves if DC or HF is detected, rather than risk destroying speakers - that may cost 4x as much, because it ends up cheaper for the manufacturer, and bad press isn't easy to recover from either.
  12. It's a form of 'feed-forward', and was experimented with a bit back in the 80's, though less popular today. The Aussie made ME580 also employs a feed-forward approach, whereby an inverse distortion signal is created and reinjected to cancel the unwanted distortion. But the amp in question also contains a fair amount of local 'degeneration' negative feedback, and indeed the typical EF output stage is an example of 100 percent local negative feedback. Bear in mind that it's not actually possible to design an amplifier with 'zero' negative feedback, and absolutely the case if local degeneration is deemed to be NFB. To that end 99.999 percent of bold claims will amount to semantics for marketing purposes. That said there's nothing inherently wrong with global negative feedback, and that's supported by the many great amplifiers that sensibly utilise GNFB to great effect. It really boils down to the design topology and layout. There are fine examples of both 'zero' and 'non-zero' global NFB designs.
  13. That's true, but more than that, he will be (and should be..) remembered for being one of the first, if not the first, to employ lots of capacitors in parallel, which has the primary effect of lowering the power supply impedance. This is actually quite significant in a class AB design, because the power supply can determine the ultimate slew rate of a class B or AB amplifier. Peter will indeed leave a great legacy but it will also remain something of a tragedy that he was never quite able to achieve the complete realisation of his ideas in the form of new product upgrades, as was his vision. Unfortunately, none of us get to chose the timeliness of our passing, and nor do we always get to chose the course of our lives. When the time for grieving has passed it will be appropriate to start some Peter Stein tribute discussions, to fully reflect on his achievements and to hopefully learn more about what he was working on and working towards, from his closest friends and colleagues.
  14. Very sad to hear. I'd heard from Zaph that he wasn't travelling too well, so although the news isn't completely unexpected it still hits hard. Well done Zaph for heading north and spending some precious time with him, and adding some happiness to his final days. RIP Peter.
  15. +1 Jaycar sells a range of very silent 12V fans - with hydro-dynamic bearings, and under $30, so not expensive. As an aside, I purchased a property in the mountains west of Sydney - with a view to retirement in a cooler place, and have the exact opposite problem. The hottest day this summer failed to reach 30C, and rarely above about 22C, and of course this recent weather has been terrible, with persistent fog and rain showers and temps in the mid teens at best. Last winter was really cold too, with morning frosts, several days of snow (6-8 cm), and failing to reach much above 4 or 5 degrees, even on many of the fine sunny days. The intention is to renovate the existing cottage (inside and out) and to extend to the rear with - a dedicated listening room / HT, so I'll be WTB amps and preamps that generate as much heat as possible.
  16. This post cannot be displayed because it is in a forum which requires at least 10 posts to view.
  17. This post cannot be displayed because it is in a forum which requires at least 10 posts to view.
  18. Very nice. We kicked off with a Bolly and followed up with the Tattinger, and it was in definitely in another league. The Bollinger lacked the elegance from start to finish and was less intense and more one-dimensional by comparison.
  19. Big day on Wednesday with a few drinking buddies for a mostly Penfolds perfect day. It was my birthday recently too so there was a cake and a celebratory glass of Penfolds brandy. I'm the old guy on the right. Gotta say that I'm not a Chard man normally but the 16A was stunning. The Grange was superb too but my pic of the reds was the bin 707. In a word, incredible.
  20. This post cannot be displayed because it is in a forum which requires at least 10 posts to view.
  21. This post cannot be displayed because it is in a forum which requires at least 10 posts to view.
  22. This post cannot be displayed because it is in a forum which requires at least 10 posts to view.
  23. Although most class D amplifiers use either ICE power or Hypex modules and identical SMPS's, so at their core they're all very similar or near identical. The primary differentiator is the input stage (buffer) which adds almost insignificantly to the measured performance in most cases. This is very unlike class AB amplifiers which are mostly discrete designs with a wide range of topologies and ditto measured performance.
  24. For me, a few months is long enough. I come to class D without any preconceptions or bias, but my opinion ultimately only matters for me. If you're happy with class D then you should use class D because it's more efficient and occupies less shelf or floor space.
  25. ABX testing is meaningless in my opinion. The only valid test is the long term listening experience. I have NC500 and 1200 monoblocks and they're both great sounding amps, but as much as I want them to be as good as my class A,AB references they're not. However, my NC1200 mono's give me the best bass I've heard in my system so I use them to power the woofers.
×
×
  • Create New...
To Top