John Curl's Blowtorch preamplifier part III

Status
Not open for further replies.
Jump to Latest
#17,544
Patrick, picture (e) is the best, Zener is filtered, 22V Vbe multiplier is not simple.
There are ones using different principles, e.g. LT3045.
Click to expand...
For $4 I'll go for conventional Didden/Jung without 1/f

Chis, ML No5xx and No5xxx have LM317/337 with BJT filters, only No523/6 has discrete regulator
 
Last edited:
#17,548
Bonsai said:
And of course, if there are ferrites anywhere near audio circuits, the usual anti-ferrite sniping will start.

Re regulators, the standard 78xx series is surprisingly quiet. If you use a split secondary, you can rectify, smooth, regulate and then combine to make a single split supply using the same regulator type. It’s also of course much easier to manage charging currents and layout. The cost is a single rectifier, but you will get that back because the negative regs are usually more expensive.
Click to expand...
I have always liked the dual secondary/dual supply idea. I advocated for it in some longish power supply thread. You get to choose the either the "least expensive," "best" or some compromise regulator type to use in both supplies. The schematics and performance are identical (within parts tolerances) for positive and negative.
chris719 said:
I thought this was high end, why use LM317 still?

I have nothing against it and it works fine. Audibly transparent in this case I'm sure, but doesn't it go against the typical attitude here?
Click to expand...
Worst case, it can be a pre-regulator for something better, maybe a high-end LDO regulator.
 
#17,549
I stuck with 'B' because it was easy to make and reliable enough.
What many here do not appreciate is that the 317-337 combination does a pretty good job of REDUCING RIPPLE, and it is fairly stable with DC voltage as well. Not precision, but precision DC was not necessary, because of the servos that kept the offset near zero.
Let's concentrate then on the POWER SUPPLY BUFFER. It further regulates the DC, as well as removes the AC noise. How can it be improved? What does it do right? And what does it do wrong?
 
Last edited:
#17,550

Attachments

  • 337.jpg
    337.jpg
    29.1 KB · Views: 238
#17,552
I'm clearly a heathen. Balanced in, balanced out and a lithium pack powering a silent switcher for the rails. Local additional regulation if required (after all if one regulator per amplifier is good enough for AP must be good enough for us).



Single ended and mains powered in a phono stage? ugh 😛
 
#17,555
john curl said:
And why don't you like single ended? What is wrong with balanced? (Hint:noisy)
Click to expand...


Hint: Not as noisy as vinyl, and can be less noisy than SE if approached the right way. For me signal integrity comes first. For marketing I understand noise figure drag racing can matter but in the real world other things matter just as much.



If I ever find a nagatron ribbon cartridge I might go transformer, but H&H even have a solution for that with their bucket of bipolars preamp.
 
#17,559
john curl said:
Let's concentrate then on the POWER SUPPLY BUFFER. It further regulates the DC, as well as removes the AC noise. How can it be improved? What does it do right? And what does it do wrong?
Click to expand...

OK, you have a FET configured as a CCS pushing a "constant" current through a resistor to make a "constant" voltage reference on the gates of the parallel pass elements. However, Idss varies with Vds, so is not really constant if the raw DC supply varies somewhat. At least the R is bypassed with a cap. An LED has a pretty constant forward voltage drop as If varies, so makes a better voltage reference, but for >19V you would need a lot of them. You could still put a string of them in series with a resistor and gain some benefit.

Is there any concern if the pass elements do not match? Would you be better off with a single higher power device, or even a MOSFET?

I don't know much about this stuff but that's my first-pass analysis.
 
Status
Not open for further replies.
Top Bottom