most of the major manufacturers in this world have understood this, too much reliability kills profitability
Not only small pF values, but relatively low voltages as well - especially in the SMDs. And leaded are getting more difficult to find unless old surplus. Hence the 500 volt silverED mica in the Miller position of most of my amps, or anywhere I need a compensation cap with tubes. Preferable to 1kV Z5U’s.
The problem with silver mica the high ESR due to a bit of dielectric absorption and to the very thin deposited metal. In a high impedance position a few ohms (or tens of ohms) of ESR isn’t the end of the world. Not like trying to push 200 watts of RF power through an 0402 cased SMD cap. Even a COG has issues with that (And mica would be a ball of fire).
“Hence the 500 volt silverED mica in the Miller position of most of my amps, or anywhere I need a compensation cap with tubes. Preferable to 1kV Z5U’s.”
I guess you can’t compare mica with polystyrene due to the HV. If you could I predict you would find the latter is better.
I guess you can’t compare mica with polystyrene due to the HV. If you could I predict you would find the latter is better.
Off-topic again but public telephone systems, at least those using Ericsson technology, once consumed literally tons of polystyrene film caps and some massive ones were rated at 400, 600V or 1000V up to and above 0.1 µF. These were apparently for the high-Q filter sections once necessary to squeeze 3 lines into a single 10khz bandwidth line and split them to 3 x 3.4kHz lines again, at a local distribution box or exchange facility. They can still be hand-made (small-order maintenance quantities actually were) with simple workbench jigs and machinery. As far as I know, many suitable thicknesses of PS film are still available at commercial rates.
Capacitors are so important. Especially in high feedback designs. They don't teach this stuff at uni. because it's too specific to audio. We are taught about capacitance, voltage, polarity and temperature range (if lucky). Problems with dielectrics and the physical structure of capacitors are unlikely to be covered. These effects are not made-up.
So high-fidelity audio is a great subject because it's not simple at all and some of the engineering challenges would cause conventional academics to wibble in disbelief.
So high-fidelity audio is a great subject because it's not simple at all and some of the engineering challenges would cause conventional academics to wibble in disbelief.
Good day
Dear gents you try clone of NAP250DR?
Maybe anybody have of DR PSU schematic? (Not for public and commercial use and if you have it please mail me)
Interesting for me testing old PSU regulator and DR for NAP amp board
I think old power reg better than for DR and 250.2 reg
I testing NAP boards with old power reg and without and sound very same but with old reg better strong bass
I like this
Dear gents you try clone of NAP250DR?
Maybe anybody have of DR PSU schematic? (Not for public and commercial use and if you have it please mail me)
Interesting for me testing old PSU regulator and DR for NAP amp board
I think old power reg better than for DR and 250.2 reg
I testing NAP boards with old power reg and without and sound very same but with old reg better strong bass
I like this
I search many times on the net, i never find the DR schematic. I just find the PCB of the 250. with schematics but as you saw there is mistakes on the PCB ... So it should be enterely verified on the original PCB by reverse engineering
There are two in-depth series of articles by the late Cyril Bateman, which are about deriving, measuring and proving the facts on caps, principally for audio applications. Both series were published in Electronics and Wireless World magazine and the original series is available online at Jan Didden's Linear Audio website or on USB stick from the DiyAudio webstore and Amazon.
An updated version of the original series of E&WW articles including project details project for building a complex, super-low distortion measurement instrument, is here: www.dadaelectronics.eu/uploads/downloads/13_Theoretical-and-practical-Electronics-papers/Capacitor-Sound-C-Bateman.pdf
I don't understand why there should be an educational rift in electronics, where practical component qualities can be ignored. What a "first day in the job" disaster that could lead to. Surely some idiot isn't going to try somewhat lossy dielectrics in high voltage AC systems, for example. From the 1970s to 2000 at least, telecommunication technician level courses were covering capacitor types in detail by the second year here. Perhaps that's due to the masses of analog audio systems still in use.
I would not expect to find the designs of current high-end products on the net. There is a reason why you only find a few genuine old Naim product designs (likely stolen) from 50 years in business.
Luckily, new graduates aren't normally allowed near anything important. It takes years of experience before that.
Your observation is not entirely surprising. Why did Naim introduce the DR after decades without it? It's not like they invented separate regulation of the low power stages of an amplifier. I have to question the motivation between sound benefit and marketing benefit.
Doing this sort of regulation is not a guaranteed upgrade to the sound. Regulators bring technical difficulties. I spent a lot of time on this myself and I can confirm that they don't always make things better, sonically. I found it very difficult to design a regulator that doesn't do harm. Consider that a conventional regulator is often a low-power, NFB amplifier, with all the warts that can entail like linearity and stability. My experience is that the downside of having the noisy power rails shared with the input stage is often better than the downsides of a regulated voltage.
Management who has both a fat salary and supposedly experience will often try to enforce the use of cheap capacitors. Specific case which has happened was to “redesign” an internally-matched RF power FET from using $3 MOS caps and 2 mil gold bond wires to 3-cent SMDs on an FR4 PCB. And wondering what happened to 10% of the power and why the goddam thing nearly catches on fire during qualification test. All the models say it works. I can’t understand what is happening here. Must be a designer error. At least audio frequency won’t usually cause caps to be a safety hazard - unless you’re trying to push 500 watts average thermal power through an NPE inside a speaker cab (which makes a loud pop and confetti, then the mids/highs go out). At least it’s contained in the box so the DJ doesn’t lose an eye.
Hello, For the output transistor what do you think about trying 2N6341 ? it's expansive but it seems to be good part better than BUV22.
That is a mighty high power switch (25A) and fast too but I can't tell whether it's sufficiently linear and I doubt it with those specs. On Semi still produce MJ15015, a souped-up 2N3055which is a compromise at 6MHz Ft but recommended as a substitute for BDY56 based on switch speeds which could mean it switches fast enough and it's otherwise acceptable too. Not cheap at US $6-10 but bona-fide parts are never sold by distributors at give-away prices.
Ok ! Why not, i have some 15024 at 4Mhz too, not very fast. But how can we know if a transistor is linear ? In some case it's indicated in the datasheet as a linear transistor for audio application but when it's not indicated ?
The transistors in the above pic, #337, will be cheap enough as they're only available from platform sellers but who made them and to what spec? I've bought a number of Chinese 2SD1047/B817 copies and most are way off Sanyo's spec. though they are clearly intended as copies of original Sanyo products - once used in almost every Rotel amplifier. FWIW, some samples branded K and sold as "KTD1047/KTB817" were actually quite good quality - I used them as replacements in pairs or full sets on a number of old designs, with mostly good results.
The 6341 probably isn’t any more NON linear than the 15024. All big *** power transistors have hFE fall off at very high current. Look at the 6341 spec sheet. Of course gain goes in the toilet at 25 amps. But 30 min at 10A is pretty good as far as old TO-3’s go. And 50 min at 0.5A. That’s less than a 2:1 change over the useful range. Not as good as a modern Toshiba or Sanken, but really not bad considering how nonlinear most of those old epi-base audio types were (maybe 150 to 200 at tens of mA, and single digits at 10A).
What they mean by characterized for linear operation on the 15024 data sheet means it was designed as a linear amp with a high DC SOA. These switching types CANNOT take the pounding that hard PA use will put them through. The SOA graphs show just a couple of amps at 30 V, dropping under and amp at 60V. For the hundreds-of-microseconds to a couple milliseconds time frame typical of “music” waveforms this can be exceeded (see the pulse thermal resistance graphs). But don’t go shorting the speaker wires or playing 30 Hz ghetto bass at high levels and expect them to survive. For that you need MJ15024’s.