best sounding 78XX/79XX drop in Regulator Replacements?

[stefanobilliani]

Peter , I would be grateful if you can sell me some ( a lot ) of the panasonic regulators. Me too cant find a source for they.

 

[Peter Daniel]

Whoever needs those regulators, send me an e-mail with your reguest and I'll provide you with a pricing.

Here's technical info http://rocky.digikey.com/WebLib/Panasonic/Web data/AN80xx,AN80xxM Series.pdf

and here's Digi Key pricing on them http://www.digikey.com/scripts/DkSearch/dksus.dll?KeywordSearch

They come in different ratings and price is about $1/pc. Shipping shouldn't be more than $3

 

[Peter Daniel]

Re: Re: Re: An8005

That was always my impressions too.
Talking GCs, no matter if it's an IGC or a NIGC, it just likes a good pre.
It's better than the chip alone with a pot, even a good one.
Purists will kill me.

Well, it depends on the preamp. Before I modified ML preamp, I preffered passive switching attenuator at the amp's input over the preamp. The preamp sounded mechanical, without top end extention. It's only after modification, that it became better than passive attenuator. Purists might kill me too

 

[Pedja]

Anyone compared 78xx to 78Lxx (100mA) ones?

 

[ThorstenL]

Konnichiwa,

Or may be one is loosing 1 or more bits thru, noise.

Unlikely. Let's take an I/V conversion stage with OPA627 and a PCM1704.

The LM2990-12 negative reg has 500uV typhical noise, a standard 78/7912 has around 80 - 100uV.

Now our OPA627 has an input noise of 5nV|/Hz in the main audio band (more at low frequencies), or around 0.7uV Self Noise in unity gain, from around 1KHz up. Our noisegain will be higher than unity, actually around 10db, so we probably get closer to 2uV output Noise with "clean" supplies, that is -120db below 2V RMS.

Now we connect our "noisy" supply.

At 1KHz our PSRR is > 100db for the (worse performance) positive rail, at 10...20KHz the PSRR is > 80db. So our 500uV Rail noise will increase our output noise of 2uV by around 0.05uV, hardly "Loosing 1 Bit", more like "nothing that I can measure reliably". So, no material worsening of noise levels

Okay, lets supply the analogue side of a PCM1704, after all that is the DAC. The worst case PSRR is 90db on the negative supply, the noise from the LM2990-5 is 150uV. There is no unweighted noise level specified, but 120db S/N ratio are claimed (A weighted) for the PCM1704, so the PCM1705 also contributes at least 2uV output noise after the I/V.

So, our 150uV are attenuated by 90db after the I/V conversion, or to 0.005uV noise contribution. Again, I fail to observe "loosing 1 Bit".

On paper the regulators may look a little noisier than classic 78XX/79XX, however, unless you seriously want to go past 20 Bit equivalent resolution I would not worry too much. When playing back CD's I'd not worry at all....

Sayonara

 

[fmak]

On paper the regulators may look a little noisier than classic 78XX/79XX, however, unless you seriously want to go past 20 Bit equivalent resolution I would not worry too much. When playing back CD's I'd not worry at all....

Sayonara [/B][/QUOTE]
------------------------------------------------------------------------------------

I am aiming for 24/96/176; not interested in 16/44.1 with or without brickwall filter.

 

[tiroth]

But it is still almost silly to ask for even 120dB SN, that is 1uV ref 1V FS.

 

[miguel2]

What happens if we put a big cap in front of the regulator? In this case the regulator is only feeding the cap and any noise generated will be smoothed. The bigger the cap, the bigger the smoothing effect, and that is what the circuit sees. Just like a big reservoir tank fed by a small current of water - small variations in the flow of the feeding current will not affect the tank's level.
High frequency can be treated in the usual small caps way. Or am I missing something?

Miguel

 

[Peter Daniel]

Are you talking about big cap before or after regulator, to be more specific?

 

[miguel2]

After.

But I allways put a big cap before too. And one should consider bypass caps near the active devices to get a correct analysis.

Miguel

 

[Peter Daniel]

Well, it seems like a common practce to put big caps after regulators and in unregulated PS of amps. But I did few experiments and comparisons, and I tell you, the smaller the cap you can get away with the better the sound. Much less mechanical and more "alive"

For instance, Kusunoki is using 33u on receiver chip and 1000u on the 1543 DAC.

I found out than 4.7u on receiver and 100u on a DAC sounds much better. I also tried 4.7u on a DAC, and although it sounds better than 1000u, a 100u cap provides still more liquidity, but I wouldn't go anything more than that. Of course this applies to BG N type caps only and I don't know how would other type of caps perform with such small values.

 

[fmak]

But it is still almost silly to ask for even 120dB SN, that is 1uV ref 1V FS.

----------------------------------------------------------------------------
Not silly. You get the nos right; then listen to the result. May be that doesn't work but you have tried.

There is a big difference between 16 and 24 bit sonics. Those who claim that degarded resolution sounds better regardless are silly.

 

[tiroth]

There is a big difference between 16 and 24 bit sonics. Those who claim that degarded resolution sounds better regardless are silly.

There is no debate there, but is there a difference between 20 and 24 bit sonics? There are not many DACs that can achieve 120dB S/N in any topology, even harder to maintain this through the entire signal chain.

I am not saying this is not a laudable goal, but we cannot "lose" bits that do not exist with any significance even at the DAC.

Well, it seems like a common practce to put big caps after regulators and in unregulated PS of amps.

A lot depends on the noise spectrum. If we are talking 100's of microvolts of low frequency noise, even 1000uF is probably not contributing much in terms of noise reduction at the lowest frequencies, but probably has a lot more parasitic inductance than 100uF.

 

[carlosfm]

I have some LM2931 that my brother gave me (they are made mainly for the automotive industry).
They seam quite good to me, I've already used them on a Dac.
There are fixed versions: 3.3v and 5v, I have the 5v ones, and also a variable voltage version.

This is on the datasheet:
"Ideal for automotive application the
LM2931 is protected from reverse battery
installations or 2 battery jumps. During the
transient, such as a a load dump (60V) when the
input voltage can exceed the specified maximum
operating input voltage (26V), the regulator will
automatically shut down to protect both internal
circuit and the load."

 

[Fred Dieckmann]

Its a "Swamp" Thing

"Often "good" regulators have such a low output impedance that they swamp any bypass capacitor sonics and determine the sound, usually in a negative sense."

The rising impedance caused by the falling open loop gain of the regulator makes the output impedance of the regulator look like an inductor. Sticking a cap on the output causes a resonance between the regulator and the cap at some frequency, often in the audio band. Put a 0.5 ohm (or there about) resistor in series with the cap to provide damping of this resonance. This has be discussed at length on the forum, and mostly in the AWL regulator threads, if you want to go back and get more info. Low drop out regulators have better stability with output caps since they are collector out instead of emitter follower output stages. Their stability is at the expense of higher output impedances. Low output impedance has been correlated with better sounding regulators for at least 30 years, providing the regulator has good transient response and PSRR.

 

[carlosfm]

The girl makes this thread much more alive.

 

[ThorstenL]

Konnichiwa,

Those who claim that degarded resolution sounds better regardless are silly.

Please note that my point was the supposedly "noisy" and "poor performance" regulator used in the cointext of a sensible "standard" circuit contributed very little noise compared to that generated by DAC and I/V Op-Amp. In fact, if you get DAC and Op-Amp noiseless and retain the Op-Amp's PSRR as original your noiselevel of 0.05uV absolute would limit you to a noisefloor in the -155db below 2V region, a cool pretty close to 26Bit equivalent.

So, in a classic POOGE situation swapping the LM2937/2940 for 78Xx and LM2990 for 79XX is likely to make minimal differences to noise, yet the improved stability and transient performance may very well pay off well. Definitly worth trying.

As always with Datasheets, reading them is one thing, one must understand how the numbers relate to the application circuit to judge the worth or unworth of trying them.

So, thank you Peter for bringing them to my attention, excellent.

Maybe you know an excellent choice of 8-Pin DIL Case Op-Amp's that improve on the OPA2604 as well (I know the LM6172 but it may be too twitchy for the job), lowish cost preferred?

And what about some 15mm PCB Lead Pattern Film Cap's with really good sonics, around 4.7uF Value, 35V minimum DC Voltage - again cost is an object?

Sayonara

 

[Peter Daniel]

As to the small, 4.7 cap, for me there is only one choice: BG N type I know you might not like it, but I'm using those in my DAC and they sound beautiful. It is important to use proper orientation regarding start of the foil (one way they sound better than the other). Otherwise, maybe Panasonic ECQ-E(B) (or F series) could be used. However I never tried them, although P series is reported to sound very good (but not available in large values). Digi Key # is EF1475 for those , but lead spacing is slightly more than 15mm.

Levinson preamp is also using OPA2604, and the sound I'm getting now is pretty good. I don't think I would like to change anything. For output buffers they use PMI SSM 2210/2220 dual transistors.

 

[jewilson]

It's been my experience that low drop out regulators or mainly for battery powered equipment. From my pass experience they just don't perform as that well in audio. Most of them have higher noise, higher output impedance to go along with their lower bandwidth.

Many years ago I found out in the lab that the Unitrode 78/79 regulators had the best performance. I don't believe there make them anymore. As far as adjustable regulators go the parts from Linear Tech or about the best.

So if your requirement is low power design a low drop out is the way to go.

 

[Fred Dieckmann]

The numbers racket

"In fact, if you get DAC and Op-Amp noiseless and retain the Op-Amp's PSRR as original your noiselevel of 0.05uV absolute would limit you to a noisefloor in the -155db below 2V region, a cool pretty close to 26Bit equivalent."

First of all I don't know what "get DAC and Op-Amp noiseless and retain the Op-Amp's PSRR as original" means.

Op Amps and DACs have finite PSRR which usually become lower as frequency increases and is lower than you might think. Noise modulation on digital supplies results in jitter. The desirability for low noise on the supply rails seems pretty straight forward.

I am trying to figure out where the 50 uV number come from.... broadband noise for a typical three terminal regulator?

I really don't know where the -155dB number comes from...........
-160dB in reference to 2 volts is 10 x 10^-9 or 10 nV . Doubling this
for -154 dB is 20nV. This is way below the noise floor of any audio equipment know to me. Since the noise floor doesn't allow most "24
bit" digital to get much below 20 bit resolution in the real world,
I can only wonder what "26 bit equivalent" as any real measurable quantity of figure of merit.

I think numbers have their place although the correlation between measured numbers (distortion being the usual subject of discussion) and sonics is controversial and not always reliable.

I find it amusing that you push all this numbers at us without any context of what they are in reference to, and then turn around and ask for recommendations for parts based on others subjective experience with them (also without much context of the rest of the circuit in which they were used ). For instance, the quality of the power supply can easily mask difference between op amps.