Why 7805 is better than LM1085?

[Christer]

Originally posted by peranders


Can this really be good?

Leakage from the first approx. 50 nA x 300 x 300 x 300 x 300 = 405 A

Hfe 50 => 300 mA

Leakage 10 nA, Hfe 100 => 1 A

Isn't this a VERY sensitive circuit? Very small leakage generates VERY much current which you don't can control.

Am I wrong?

I think you mean that we cannot ignore recombination and surface leakage for the first one or two transistors, but
your calculations are wrong, since you do not take all effects of
this into consideration. As you might remember from your
semiconductor-physics books, and as can also be seen in the
datasheets for most BJTs, we have a beta-droop not only for
high IC values, but also for low values. Hence, you cannot use
the same hfe figures for all transistors. The first ones will have
a very low Vbe and also a very low hfe.

[peranders]

Originally posted by Elso Kwak
Hi peranders,
I don't know where you got these leakage numbers from and it may escaped your attention but my circuit is a <B>working</B> one!
Audionote uses 680 Ohm resistors from bases to ground in the last four transistors as also Guido Tent does.

I look at BC546 datasheet: Collector cutoff current: 15 nA, worst case at 150 deg C 5 µA.

Do you mean four emitter followers connected in chain, with load down to ground from each one of them? If, so that's better.

A "real" quadrupleton has a huge current gain if not measures are taken.

[peranders]

Originally posted by Christer
we have a beta-droop not only for
high IC values, but also for low values. Hence, you cannot use
the same hfe figures for all transistors. The first ones will have
a very low Vbe and also a very low hfe.

Still, some transistors have substantial Hfe at low currents.

This was interesting though, I might test what happens with a quadrupleton.

[ALW]

I don't doubt your subjective sonic results for one second, after all, if it sounds good, it is good.

But any theory as to why the Jung type reg's sound 'hopeless' in this application.

I mean, DC is DC, closer to DC is a better PSU, does the chip somehow thrive on imperfect supply rails?

Line rejection of a bootstrapped Jung reg. is almost immeasurable without great care in measurement technique, so I don't feel that's the issue. Output impedance and noise will all be lower with the Jung reg too.

Any thoughts?

Andy.

[Elso Kwak]

Originally posted by ALW
I don't doubt your subjective sonic results for one second, after all, if it sounds good, it is good.

But any theory as to why the Jung type reg's sound 'hopeless' in this application.


Any thoughts?

Andy.

Hi Andy,
I am not sure but I suspect the TDA1543 is spitting HF-garbage back into the powersupply. And second with all these clocks running in the same enclosure, Latch, Bitclock, etc. these might couple into the powersuply too, upsetting the regulator.

[Elso Kwak]

Originally posted by peranders



A "real" quadrupleton has a huge current gain if not measures are taken.

Hi peranders,
May I quote from the Low-Noise Electronic Design book for the one transistor capacity multiplier:
<B><I>"The capacity multiplier circuit has the potential for increasing the filtering time constant by a value equal to the transistor Beta. This conclusion follows because the filtering takes place in the base current, whereas the load current is the transistor emittor current and Ie= BetaxIb. The presence of ripple at the collector of Q1 has little effect since Ie is essentially independent of that voltage " </B></I>
and<B><I>
"If lower ripple is needed, a second transistor can be added as in fig.11-5. This can reduce the ripple by another factor of Beta. It is possible to use the extra transistor as a trade-off for a smaller capacitor. It may be less expensive to buy the second transistor than a large electrolytic capacitor. This two-transistor circuit is also useful when the supply must deliver a large load current. Transistor Q1 handles the current and Q2 provides the gain. When Q1 is a low-Beta powertransistor, Q2 can be selected to increase the current gain".</B></I>


You see peranders gain is just what we want.

[Elso Kwak]

Originally posted by ALW
I don't doubt your subjective sonic results for one second, after all, if it sounds good, it is good.

But any theory as to why the Jung type reg's sound 'hopeless' in this application.

I mean, DC is DC, closer to DC is a better PSU, does the chip somehow thrive on imperfect supply rails?

Line rejection of a bootstrapped Jung reg. is almost immeasurable without great care in measurement technique, so I don't feel that's the issue. Output impedance and noise will all be lower with the Jung reg too.

Any thoughts?

Andy.

Hi Andy see also;
http://db.audioasylum.com/cgi/m.pl?f...41&r=&session=

[peranders]

Originally posted by Elso Kwak

Hi peranders,
May I quote from the Low-Noise Electronic Design book for the one transistor capacity multiplier:
<B><I>"The capacity multiplier circuit ...

You see peranders gain is just what we want.

But are your circuit really a capacity (capacitance) multiplier???

[ALW]

I am not sure but I suspect the TDA1543 is spitting HF-garbage back into the powersupply. And second with all these clocks running in the same enclosure, Latch, Bitclock, etc. these might couple into the powersuply too, upsetting the regulator.

The only way I could see this being a problem (assuming a stable regulator) would be non-ideal power or earthing connections, the current flows in which could upset other items / the DAC. An unstable reg. obviously needs rectifying rather than blaming.

That layout issues can certainly be a problem when modifying items, but when designing from scratch you have control, one should never overlook (I'm sure you don't) the importance of layout, the PCB is part of the circuit. it's characteristics rarely shown on the schematic

Interesting stuff anyway,

Andy.

[Elso Kwak]

Originally posted by peranders


But are your circuit really a capacity (capacitance) multiplier???

Peranders,<B>
WIIILMAAAAAAAAAAAAAAAAAAAAAAAAAA!!!!!!!!!!!!!!!!!! </B>
Wilma (in her well controlled and friendly voice): "What is the matter Fred?"
Fred:<B> "*&#$%^@!&%#" </B>
In other words, Peranders Give It up or Let Me Go.
Bonnie Raitt, my sweetheart, thanks for the line.


Sorry Fred Dieckmann this is not meant to be you but Fred from the comic TV cartoon "The Flintstones"