In the unlikely event that there could have been collateral damage along the way, I've clearly stated that it was your responsibility to backup.
http://www.diyaudio.com/forums/soli...torch-preamplifier-part-ii-3.html#post1868453
/Hugo
http://www.diyaudio.com/forums/soli...torch-preamplifier-part-ii-3.html#post1868453
/Hugo
As has been said before, on this website, what we try to make, is power supplies that behave much like ideal batteries, as isolated as much as possible from the power line.
This requires high isolation transformers, good, fast regulators, and a number of passive parts, including caps and chokes.
On the other hand, many parts will contaminate the power supply, much like the power line will, if not addressed correctly. For example, slow, high feedback regulators will overshoot on current transients, sometimes generate excessive noise due to 'ringing', and almost always add extra noise from their 'zener equivalent' voltage reference, due to lack of cap decoupling. Typical power supply diodes will 'spit' pulses of current from the slowness of their turn-on and turn-off times, and many caps that are small, cheap and cute, will generate nonlinear distortion, have high dielectric absorption or linear distortion, or will not bypass very high frequencies properly due to self resonance and high DF.
Therefore, I tend to use 'soft recovery' power diodes in my best designs, polystyrene or polypropylene, and high quality smallish (33-330 uf) aluminum electrolytic caps (similar to Black Gate) in my power supplies. For gross filtering, I have found that Panasonic TS series or the rough equivalent large electrolytic caps work well, for preamps at least. Thats='s enough for now. Let's see if I can progress further.
This requires high isolation transformers, good, fast regulators, and a number of passive parts, including caps and chokes.
On the other hand, many parts will contaminate the power supply, much like the power line will, if not addressed correctly. For example, slow, high feedback regulators will overshoot on current transients, sometimes generate excessive noise due to 'ringing', and almost always add extra noise from their 'zener equivalent' voltage reference, due to lack of cap decoupling. Typical power supply diodes will 'spit' pulses of current from the slowness of their turn-on and turn-off times, and many caps that are small, cheap and cute, will generate nonlinear distortion, have high dielectric absorption or linear distortion, or will not bypass very high frequencies properly due to self resonance and high DF.
Therefore, I tend to use 'soft recovery' power diodes in my best designs, polystyrene or polypropylene, and high quality smallish (33-330 uf) aluminum electrolytic caps (similar to Black Gate) in my power supplies. For gross filtering, I have found that Panasonic TS series or the rough equivalent large electrolytic caps work well, for preamps at least. Thats='s enough for now. Let's see if I can progress further.
Last edited:
Well, that's where the compromises come in, isn't it. There's no way you can approach an 'ideal battery' without a feedback regulator.
And if you are afraid that a regulator overshoots etc, it's a badly designed regulator. John, you are I presume well-versed in feedback theory, and the issues of stability and settling time. You must know that it is quite simple to design an active regulator that behaves like an 'ideal battery' for audio loads, something you can only dream of with a zener stabilized emitter follower as is sometimes presented here.
It seems to me you are muddling the issues in your above post.
jd
Without a very high loop gain or a really big output C (both apparantly "bad") a shunt regulator as a rather high output impedance........
This is not true, unless by 'battery' you mean an ideal voltage source.
Batteries can be lousy power supplies. They may have high output impedance, and not even noise is guaranteed to be low. Actually, I wouldn't touch a battery with a pole, except for mandatory portable equipment and very low power consumption applications.
P.S. To the extend I am aware of, there are no batteries specified and guaranteed for ultra low noise. OTOH, designing a serial or shunt regulator for ultra low noise is not that difficult.
And here's another freebee for John: look at this evil $0.4 LM329. At 7uV wideband noise, you'll have a hard time finding a battery to beat this. Key word: buried zener. If you dig a little further, and you can afford a few bucks, you may find references with <1-2uV noise in 0.1-10Hz, like the AD44X series from Analog: http://www.analog.com/en/references/voltage-references/adr440/products/product.html
Last edited:
To quote Mr. Ben Duncan:" There should be some conductance between signal and safety grounds , otherwise the small but finite leakage capacitance and and insulation resistances across the mains transformers would cause the signal ground possibly to rise to the mains voltage potential .
Isolation guards against the effects of the induced AC power frequency and its harmonics.
4K7 resistor also limits inter-ground fault currents so PCB tracks on highly priced equipment are protected from costly vaporisation"
Except with a Class II chassis and a two conductor AC cord, there is no safety ground, so this point is irrelevant.
Yes, isolation from the safety ground. If there is no safety ground, there's nothing to isolate.
Ok, he's talking about fault currents burning up circuit board traces. That wasn't clear in your original post.
se
Please explain! Does the leakage current from class II equipment burns away PCB traces?
Regards
digging a little further:
Linear Technology - 775 Nanovolt Noise Measurement for a Low Noise Voltage Reference
make sure you check the AN124 application note.
Also, some good design ideas on battery power supplies (if you are so inclined):
SIM928 - Isolated Voltage Source
Regards
Linear Technology - 775 Nanovolt Noise Measurement for a Low Noise Voltage Reference
make sure you check the AN124 application note.
Also, some good design ideas on battery power supplies (if you are so inclined):
SIM928 - Isolated Voltage Source
Regards
Please explain! Does the leakage current from class II equipment burns away PCB traces?
I didn't say leakage current. I said fault current.
se
If you assume a 220 volt mains (seeing as it's Duncan who's in the UK), 4.7k would be only about 47mA. I don't see 47mA burning up any circuit board traces.
The resistor's another story.
se
Most overnight responses and questions about power supplies show a lack of attention to what I have said in the past. There is NO one power supply regulator that does 'everything'. If there was, we would use it. That is why I use 3 ACTIVE regulators for each channel just for the line amp of the CTC preamp. Each one does a different job, properly. For decades, I have used at least 2 ACTIVE regulators in series with EVERY gain stage.
For example, even the LM317-337 pair do hum reduction very well, but they are lousy with self noise, transient overshoot, and with high frequency rejection. Still they can be VERY useful to make front end AC-DC filters. I use them by the 1000's.
One of the advantages of active, open loop, local decoupling, besides removing high frequency garbage, and being low noise with a bit of care, is the isolation between stages, so that you truly isolate the left and right channels. This can be heard, and is therefore worth the effort. It is one of the main differences between the JC-80 and the Levinson JC-2.
As far a 'ideal battery' is concerned, that is the ideal, we can only try to approach it.
Ben Duncan makes a lot of sense to me in his approach to 3'rd wire grounding. We use much the same thing.
For example, even the LM317-337 pair do hum reduction very well, but they are lousy with self noise, transient overshoot, and with high frequency rejection. Still they can be VERY useful to make front end AC-DC filters. I use them by the 1000's.
One of the advantages of active, open loop, local decoupling, besides removing high frequency garbage, and being low noise with a bit of care, is the isolation between stages, so that you truly isolate the left and right channels. This can be heard, and is therefore worth the effort. It is one of the main differences between the JC-80 and the Levinson JC-2.
As far a 'ideal battery' is concerned, that is the ideal, we can only try to approach it.
Ben Duncan makes a lot of sense to me in his approach to 3'rd wire grounding. We use much the same thing.
Wouldn't the better approach be to design to Class II standards and then just take the safety ground out of the picture completely?
se
BZZZT...
- It's not cheaper; low noise JFETs are not free.
- You'll have some trouble in doing a negative reference.
- It's actually much worse. Do yourself a favor and calculate the equivalent noise of a Norton reference. About precision and thermal drift, in particular if you use the Vendetta very poor PSRR topology, another time...
- You can decouple the LM329 without problems.
- Comparing a reference noise with the Vendetta input stage noise is an irrelevant smoke screen.
As expected, you refuse to admit that anything good happened in the last 40 years technology.
- Status
- Not open for further replies.