I just ordered a bunch of those "Audience" (Douk Audio, breeze audio, whatever other "Chinese" brand, they're all the same) SX52B op-amps.
Because I also bought a Breeze audio active crossover.
The crossover is a two way and uses a total of five NE5532 in sockets.
So my dumb question is: how can I be sure that the PSU will be able to supply enough current to the 5 op-amps, knowing one op-amp quiescent current is 30mA? Is it just a matter of adding it all? Like, in this case, total quiescent would be 150mA? Then I should just check with the regulators specs if I'm good to go? or is there something that escapes me?
And other question: has anyone here tried those op-amps? I'm usually working with OPA2132 and LM4562, this is my first foray into (cheap) discrete, if I like what cheap discrete do I might invest in something like Sparkos Labs. This is meant as an experiment)
Because I also bought a Breeze audio active crossover.
The crossover is a two way and uses a total of five NE5532 in sockets.
So my dumb question is: how can I be sure that the PSU will be able to supply enough current to the 5 op-amps, knowing one op-amp quiescent current is 30mA? Is it just a matter of adding it all? Like, in this case, total quiescent would be 150mA? Then I should just check with the regulators specs if I'm good to go? or is there something that escapes me?
And other question: has anyone here tried those op-amps? I'm usually working with OPA2132 and LM4562, this is my first foray into (cheap) discrete, if I like what cheap discrete do I might invest in something like Sparkos Labs. This is meant as an experiment)
Perhaps you should identify this active cross-over model?
I can't find any decent datasheet for these SX52B's. What's the current noise? Are they immune from phase-reversal or latch-up? What's the driving ability (actually that ought to be reasonable given the immense quiescent current).
By the way you have 10 opamps, 5 packages, is the quiescent current 30mA per opamp or 30mA per package?
The design of the power supply needs to be checked overall, not just the reglators.
I can't find any decent datasheet for these SX52B's. What's the current noise? Are they immune from phase-reversal or latch-up? What's the driving ability (actually that ought to be reasonable given the immense quiescent current).
By the way you have 10 opamps, 5 packages, is the quiescent current 30mA per opamp or 30mA per package?
The design of the power supply needs to be checked overall, not just the reglators.
Yeah I don't know... I read that 30mA SOMEWHERE but there are hundreds of pages selling that op-amp and no one gives a decent data sheet. There was ONE with a more complete technical description but I can't find it anymore :-/
Here under is the link to the active crossover; as you can see it is a super cheap thing also from china.
It's still in china at the moment so I cannot have a closer look on the PSU.
Please remember this is meant to be a fun experiment of "how far can you go with a Chinese 49$ crossover" and if the discrete op-amps suck I'll stuck some OPA2132 in there until I have the dough to get something from Sparkos Labs (if the PSU is strong enough and if the crossover is decent enough).
I do have an old TDM crossover that I upgraded myself: OPA2132, LM4562, added decoupling, better caps, signal caps replaced with film caps etc. It does sound great but being a pro unit I feel there's a LOT of "unnecessary" stages, caps and opamps (unnecessary in a home-audio context I mean) plus the amount of op-amps makes it impossible, both budget-wise and current-wise, to try discretes.
That's why I went for that cheap Chinese crossover: some guy on another forum saying "it's not as good as my Marchand XM44 but definitely much better than commercial units from Rane, Ashly, etc". For the price I'm not taking a big risk but I certainly don't want to destroy its PSU immediately - it takes too many weeks to order a new one ^^
here's the link:
WEILIANG AUDIO linkwitz riley diviseur de frequence electronique preamplificateur a deux divisions de frequence-in Amplificateur from Electronique on AliExpress - 11.11_Double 11_Singles' Day
Here under is the link to the active crossover; as you can see it is a super cheap thing also from china.
It's still in china at the moment so I cannot have a closer look on the PSU.
Please remember this is meant to be a fun experiment of "how far can you go with a Chinese 49$ crossover" and if the discrete op-amps suck I'll stuck some OPA2132 in there until I have the dough to get something from Sparkos Labs (if the PSU is strong enough and if the crossover is decent enough).
I do have an old TDM crossover that I upgraded myself: OPA2132, LM4562, added decoupling, better caps, signal caps replaced with film caps etc. It does sound great but being a pro unit I feel there's a LOT of "unnecessary" stages, caps and opamps (unnecessary in a home-audio context I mean) plus the amount of op-amps makes it impossible, both budget-wise and current-wise, to try discretes.
That's why I went for that cheap Chinese crossover: some guy on another forum saying "it's not as good as my Marchand XM44 but definitely much better than commercial units from Rane, Ashly, etc". For the price I'm not taking a big risk but I certainly don't want to destroy its PSU immediately - it takes too many weeks to order a new one ^^
here's the link:
WEILIANG AUDIO linkwitz riley diviseur de frequence electronique preamplificateur a deux divisions de frequence-in Amplificateur from Electronique on AliExpress - 11.11_Double 11_Singles' Day
NONE of the IC op-amps runs anywhere near 30mA. At 30V (+/-15V) that would be ONE WATT which is way too much for a precision 8-DIP.
The quiescent current is "always" on the datasheet. See attached.
http://www.ti.com/lit/ds/symlink/ne5532.pdf
Yes, this is no-load. If you run LARGE signal in LOW impedances, you must figure dynamic current. If you run up to clipping, multiply your load impedance by six and assume a resistor of that value rail-to-rail. 30V peak-to-peak in 600 Ohms: 6*600 is 3,600r, 30V/3600 is 0.008A or 8mA, two channels makes 16mA, plus the whole-chip 16mA Max is.....
Ah, I see. Yes, a very conservative design for a high-level low-Z console must assume 32mA per chip.
You are not driving 600 Ohms, possibly over 2k each amp.
You are not beating +20dBm (7Vrms) to flog miles of phone lines with steady tone, but under 2Vrms on peaks and average 0.2V. Your dynamic current is 0.1mA per amp, 0.2mA per chip, plus 8mA-16mA per chip for idle.
98% of '5532s will be far closer to 8mA than 16mA. Supply current is not a key parameter for a good audio amp but is tightly watched (if not Controlled) in production because current double what it should be implies a Problem in fabrication (crappy silicon, fuzzy masks, under-cooked...).
Attachments
Its a common misconception than fewer components is better - in fact the opposite is the case, more components allows better protection circuitry, reduced noise, reduced distortion, better headroom, better performance overall (except for power consumption perhaps!).
Engineers don't throw in extra stages without a reason, they are engineering for minimum cost for the wanted performance. An active crossover needs many filters, usually 4 pole or more, a few dozen opamps is to be expected. Of course a 2-way crossover is a lot simpler than 3-way, that's the main option for simplification. For a pro unit it will be 3 or 4-way, with time-delay all-pass filter(s), and individual level controls. If you only want simple 2-way crossover a pro unit is not the right choice!
Of course. but all I need is a fixed 2 way crossover at 400Hz, 24dB L-R. In a house system driving a 60cm single ended interconnect. No need for symmetrizer circuitry, low pass filters, etc. No matter how good that crossover is (the tDM is very good) all those "options" have a cost in transparency. A simple crossover with a short signal path and a limited number of op-amp will allow me, if circuit is good enough, to upgrade what needs to be upgraded, and try "expensive" op-amps that I couldn't try when there's 24 op-amps to replace...
A crossover certainly isn't "transparent", its modifing one signal into several in a very obvious fashion. I'm sure distortion isn't an issue if that's what you mean, what matters is getting the response to match the speakers and room acoustics well, not whether its 0.001% or 0.01% distortion. The drivers in the speakers are much worse than that.
You know what I'm talking about... each stage of passive or worse, active components impairs some fogging, hazing, bleaching of the sound. It may not be measurable (or barely), it may be a minute amount, but it keeps adding up, and it's very easy to hear it.
If its barely measurable, you certainly won't find it under double blind listening tests! Several stages with 0.001% distortion will end up being worse than one, but the total distortion will be well below the threshold of perception, either for harmonics or intermodulation products. The reason we aim for very low distortion is so that after many many stages its still indistinguishable from the original. As in an old analog mixing desk with dozens of stages involved, carefully designed for world-class performance overall (tough market to compete in).
Without double-blind protocol no listening tests are worth much I'm afraid. The limits of human perception are pretty well understood and measured. If you have exceptional hearing then its up to you to prove that, most people don't (yes hearing can be trained to some degree).
Just having slightly different volume levels between two test setups will favour the louder one under blind testing for "quality". So yes double-blind testing is hard to do, but its the only way to compare equipment fairly - otherwise you may just be comparing expectations.
Subjective impressions for sound are often dominated by expectation bias, not the actual audio quality (which is usually always adequate using modern equipment). You can rig up an ABX box and see if your beliefs about SQ are real or bogus.
Without double-blind protocol no listening tests are worth much I'm afraid. The limits of human perception are pretty well understood and measured. If you have exceptional hearing then its up to you to prove that, most people don't (yes hearing can be trained to some degree).
Just having slightly different volume levels between two test setups will favour the louder one under blind testing for "quality". So yes double-blind testing is hard to do, but its the only way to compare equipment fairly - otherwise you may just be comparing expectations.
Subjective impressions for sound are often dominated by expectation bias, not the actual audio quality (which is usually always adequate using modern equipment). You can rig up an ABX box and see if your beliefs about SQ are real or bogus.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.