Hi Tryphon,
In a perfect world, yes. You can get some good rejection.
Tell me, when you have small capacitance, what passes through more easily. A high frequency or a low frequency? Why don't linear regulators (used after the switcher) deliver a perfect, clean voltage? Should these switching regulators be encased in a metal enclosure?
-Chris
In a perfect world, yes. You can get some good rejection.
Tell me, when you have small capacitance, what passes through more easily. A high frequency or a low frequency? Why don't linear regulators (used after the switcher) deliver a perfect, clean voltage? Should these switching regulators be encased in a metal enclosure?
-Chris
I can imagine a SMPS with near total isolation. The first stage as usual. A square HF oscillator transmit the DC to a second capacitance (during the positive half period), and a third is filled from the previous the other half of the time. The rest as usual.
There is no moment where a low impedance will exist between output and input.
There is no moment where a low impedance will exist between output and input.
Hi Tryphon,
Your imagination is a perfect world. Now for a taste of reality.
Have you ever used an oscilloscope to look at the output from these regulators?
Some time, connect a coupling capacitor to the DC rail and into the input of an amplifier connected to a speaker. Short the RCA side (if you use a plug) until the power supply has come up and is running. Then connect it to your amplifier. Have a listen, but low so you don't damage your tweeters. Best done with a full range speaker.
-Chris
Edit: Of course you use the power supply ground for the RCA common connection.
Your imagination is a perfect world. Now for a taste of reality.
Have you ever used an oscilloscope to look at the output from these regulators?
Some time, connect a coupling capacitor to the DC rail and into the input of an amplifier connected to a speaker. Short the RCA side (if you use a plug) until the power supply has come up and is running. Then connect it to your amplifier. Have a listen, but low so you don't damage your tweeters. Best done with a full range speaker.
-Chris
Edit: Of course you use the power supply ground for the RCA common connection.
In the shop at the moment we have 30, 50, and 115 kw PFC corrected switchmode power supplies under construction. These cost in USD $30 to 100k region so not in the realm of affordable for audio front ends. Arriving today were the dual electrostatically shielded transfomers we run them from so the power plants they work in don’t destroy them when they lock up their 15,000 hp motors on the coal grinders and put 4 kv transients on the power plant’s AC distribution bus. The dual ESL transformers have been pretty effective at preventing front end damage on the switchmode supplies during these significant building AC transients
Last edited:
Of course not, I was born during the last rain.
I have a whole world of wonderful things to discover.
Like an oscilloscope, then how to design and build low noise regulator supplies.
That was NOT the question. The question was to null as much as possible the currents in the ground connections between two devices connected together and powered from the same AC.
Last edited:
In a double shielded transformer, there are two electrostatic shields between the primary and secondary windings insulated from each other. This allows you to connect the primary shield to AC mains ground and the secondary shield to your reference ground point in the circuit. It greatly reduces capacitive coupling of AC mains EMI to the secondary. Also, the secondary shield now is pulled to the potential of the circuit reference ground, reducing potential differential between it and the secondary winding.
In a triple shielded transformer like the Topaz, there is an additional shield between the two which is brought out separately and should be tied to the quietest earth ground available, often back to the power entrance grounding electrode (near which these transformers are often located).
(another old story approacheth) In the audio plant we needed a local ground reference. We sawed through the concrete pad around the perimeter of the audio area to break the galvanic connection with the rest of the factory's rebar, and caulked the joint. Then we drilled and chipped a hole through the 6" thick concrete pad, cadwelded to the rebar to make the pad under the audio machines an Ufer ground and drove grounding electrodes in the hole. We then doped the soil through the hole with electrolytic salts to enhance it's conductivity. This became our reference technical ground to which the (8) 5KVA Topaz's center shields were tied. It was a technicians job to water the hole once a week. We discovered the noise floor in the system would rise otherwise. Keep in mind this was unbalanced audio transport measured over a 400Hz to 1.6MHz bandwidth.
At home I use a 5KVA Topaz to power my ham setup, the center shield is tied to my partial ring ground outside with 6 electrodes bonded to the power entrance ground.
All of this verbiage has no direct application to a Blowtorch, but the principle of isolating power, especially when using unbalanced signal transport between chassis certainly has an application to home and pro audio. It specially applies to a home theater system which may have additional connections to sources of unknown potential.
Cheers!
Howie
Hi Howie,
Yes, but the "technical ground" can cause a ton of problems for people not used to the nuances. It's pretty easy to create the monster of all ground loops with a particularly nasty, harmonic rich 60 cycle hum!
At home you could sneak your grounding mesh and rods under the garden. It gets watered regularly that way and no one is the wiser.
-Chris
Yes, but the "technical ground" can cause a ton of problems for people not used to the nuances. It's pretty easy to create the monster of all ground loops with a particularly nasty, harmonic rich 60 cycle hum!
At home you could sneak your grounding mesh and rods under the garden. It gets watered regularly that way and no one is the wiser.
-Chris
Why do people talk about switching supplies as if there is only one kind?
Also, anybody besides me use more than a scope to track down the problems switchers sometimes create? A small battery-powered shortwave radio with an external antenna that can act as a probe can be a real eye opener. You can follow RF produced in a wall wart right along a 5v power cable shield, though a Silent Switcher, and eventually radiate off into space (I guess). That may be were some of the energy goes after doing mischief along the way with conductors under the shield, before and after the SS, after all we don't use triple shielded wire much either. Can't affect audio? Sometimes it can in my experience, not always, or necessarily even usually, but it can.
The other thing maybe good to keep in mind is that with 24-bit audio and 21-bit or even 22-bit converters, we are talking about extremely tiny signal levels that can affect performance. Sure, enough stages of power conditioning, or a built-in power supply that does the same thing in the end is possible, manufacturers have to decide what customers are willing to pay for and how many are likely to notice anything or figure out the cause if they do notice something. It can be hard enough for engineers sometimes.
Also, anybody besides me use more than a scope to track down the problems switchers sometimes create? A small battery-powered shortwave radio with an external antenna that can act as a probe can be a real eye opener. You can follow RF produced in a wall wart right along a 5v power cable shield, though a Silent Switcher, and eventually radiate off into space (I guess). That may be were some of the energy goes after doing mischief along the way with conductors under the shield, before and after the SS, after all we don't use triple shielded wire much either. Can't affect audio? Sometimes it can in my experience, not always, or necessarily even usually, but it can.
The other thing maybe good to keep in mind is that with 24-bit audio and 21-bit or even 22-bit converters, we are talking about extremely tiny signal levels that can affect performance. Sure, enough stages of power conditioning, or a built-in power supply that does the same thing in the end is possible, manufacturers have to decide what customers are willing to pay for and how many are likely to notice anything or figure out the cause if they do notice something. It can be hard enough for engineers sometimes.
Last edited:
Hi Tryphon,
I'm sure we'll return to ground differentials again. Be patient and the thread will get there again.
-Chris
Thought so!
No. That just became your question. We were talking about noise coming through the power supply and AC line conditioning. However, leakage currents can be related to the unit power supply if you want to talk about that.
I'm sure we'll return to ground differentials again. Be patient and the thread will get there again.
-Chris
I have to agree with Tryphon here. You can build an excellent isolated SMPS and you can filter it, post regulate it, etc. I don't think I have even come across any test and measurement gear that uses a linear supply that was produced in the past decade.
Linear supplies with overkill amounts of cap hung off them and high-speed rectifier diodes (as is common in audio/DIY) aren't the most EMI friendly thing either with the narrow conduction angle.
Hi Mark,
Yes. AM radios are really useful, and not just the broadcast band ones. I used to have a hearing aid amplifier, a silver one that was reasonably large. I plugged various coils into it and followed noise that way. And recently I bought an HP 547A current tracer probe and a pulser (HP 546A). I had wanted one for years and finally have one!
-Chris
Yes. AM radios are really useful, and not just the broadcast band ones. I used to have a hearing aid amplifier, a silver one that was reasonably large. I plugged various coils into it and followed noise that way. And recently I bought an HP 547A current tracer probe and a pulser (HP 546A). I had wanted one for years and finally have one!
-Chris
Hi Chris,
Yes, and I agreed with Tryphon about that. But outside of test and medical equipment I seldom see "nice" switchers.
I don't use large amounts of capacitance for that very reason. I've often had to draw people's attention to things like conduction angle. You're preaching to the choir here.
-Chris
Yes, and I agreed with Tryphon about that. But outside of test and medical equipment I seldom see "nice" switchers.
I don't use large amounts of capacitance for that very reason. I've often had to draw people's attention to things like conduction angle. You're preaching to the choir here.
-Chris
It's all about the design of the entire product. I am sure the switcher Benchmark is using is nothing special but it is carefully chosen and the rest of the system is designed for it.
Most of the medical SMPS I come across are pretty much identical to the commercial grade but without the Y-caps (or smaller value) to ensure leakage current passes IEC 60601-1-2.
People here are very obsessed with individual parts when it's the sum of the parts that counts. From the measurements I have seen, the outputs of the Benchmark are clean and you wouldn't know it uses a SMPS and internal switching regulators.
I know they are frowned upon but I'll give an automotive analogy:
Porsche uses MacPherson struts in front on the 911. People often consider that an inferior setup, yet the 911 still beats its more powerful rivals on track, even though they are using "better" front suspensions.
Most of the medical SMPS I come across are pretty much identical to the commercial grade but without the Y-caps (or smaller value) to ensure leakage current passes IEC 60601-1-2.
People here are very obsessed with individual parts when it's the sum of the parts that counts. From the measurements I have seen, the outputs of the Benchmark are clean and you wouldn't know it uses a SMPS and internal switching regulators.
I know they are frowned upon but I'll give an automotive analogy
: Porsche uses MacPherson struts in front on the 911. People often consider that an inferior setup, yet the 911 still beats its more powerful rivals on track, even though they are using "better" front suspensions.
That I would agree with. It is possible to make better switchers both in terms of conducted and radiated RFI/EMI they produce. And, I'll bet the switchers in the Benchmark gear are of that type. They say the same things you guys are saying, after all. And, it's all true, no question. However, try to find one at a competitive cost at Digikey or Mouser for your quick one-off project. For one consideration, a linear supply doesn't care if it is unloaded. A switcher might be able to regulate over a 10:1 output current range, and it might not be equally noisy at every output current. The more one can characterize the exact load the supply will have to regulate with the more it can be optimized to be low noise under that condition. Of course, SMPS controller chips are getting more sophisticated so it easier to make a supply with some compensation for realistic operating condition variations.
True, because safety is the universal factor for medical power supplies. And that includes non-interference with life safety devices. But, most medical devices don't need performance down to 22-bits or thereabouts either. Those that do get what they need and we all pay for what it costs one way or another, and I can assure you first hand it costs a lot for some of that stuff.
They connect the RCA shell right to the chassis too by the looks of it. How ever will they let the RF in? They need to contact Lampizator to figure out how to do it right.
That is true! Thus the need for an isolation transformer to break the galvanic and capacitive coupling between chassis as well as to incoming power. No connection, no loop. The code mandated safety ground on a separately derived system must be connected to the secondary side ground.
Cheers & GN,
Howie
- Status
- Not open for further replies.