That would be a grey zone as class A amps and their PSU's + mains connections are discussed openly. "All animals are equal but some animals are more equal than others"
Just looked it up and it says:
While most projects on this site deal with electricity and construction which inherently involve some risk, particularly dangerous topics and procedures should include a warning in the thread that adequately explains these risks. Certain inherently dangerous topics are not allowed. At this time they include but are not limited to: discussing power supplies directly fed by mains current without a transformer, and mucking about in CRT video monitors. Posts and projects are those of individual members of diyAudio. The forum itself is not in the business of vetting projects or posts for safety, accuracy, performance, reliability, function, or fitness for use. If you attempt to make something and it blows up, or turns expensive parts into charcoal, or just doesn't work the way you were hoping, that's between you and the person posting the project or idea. The forum is merely a bulletin board which allows anyone to post ideas, criticisms, or discussions. It is up to the individual to make the final determination of how appropriate a project is for them to attempt, based on their own experience.
Mmm, this would indeed rule out DIY mains fed SMPS. Problem solved !
I see, wer lesen kann ist klar im Vorteil!
That one has some similarities with the Silentswitcher....Nice schematic and ultra low noise/EMI. In fact a very nice chip with good specifications. LT3042 would be a good companion.
But no 115V/230 mains input...I suppose adding a 6V transformer, rectifier and filter cap for 6V DC is no real problem but apparently the classic transformer is heretic in audio Since mains fed without transformer is forbidden at diyaudio.com we might as well make this a classic/SMPS combo so with the dreaded transformer as there is no other possibility for DIY. I sure see some possibilities here. Adding a third output voltage seems no problem with an LT3042 set at 5V connected to 6V directly.
http://www.mouser.com/catalog/specsheets/213664es.pdf
5V @ 0.2A plus +/- 15V @ 100 mA max. that's a whopping 4W. Low power rules I was planning to stay away from this thread but I will put some time in this.
But no 115V/230 mains input...I suppose adding a 6V transformer, rectifier and filter cap for 6V DC is no real problem but apparently the classic transformer is heretic in audio
Since mains fed without transformer is forbidden at diyaudio.com we might as well make this a classic/SMPS combo so with the dreaded transformer as there is no other possibility for DIY. I sure see some possibilities here. Adding a third output voltage seems no problem with an LT3042 set at 5V connected to 6V directly. http://www.mouser.com/catalog/specsheets/213664es.pdf
5V @ 0.2A plus +/- 15V @ 100 mA max. that's a whopping 4W. Low power rules
I was planning to stay away from this thread but I will put some time in this.
Last edited:
Please explain that. It is possible to tune this in a classical transformer-rectifier-capacitor way. The current low noise LDO regs need less than 0.5V drop so with low power supplies things become trivial quite fast
The LDO needs only 0.5 V if the input was stabilized. But a transformer+diodes+cap doesn't produce stable voltage. If you need 5.5V at full load and minimal line voltage then at light load and nominal line voltage it will give you about 8V. And a typical graetz drops (dissipates) further 1.8V.
I dont know what do you mean tuning.
SMPS is replacing transformer, not LDO in highend PSU. It was told already.
Why me? You were searching for. I can't tell it is neccessary. I can only tell it is beneficial. I told why.
Instead of what?
Well, not neccessarily. There are many other ways, this is just the simplest.
This proves at least some point
If you think low weight, high efficiency and low leakage current are not benefits...
Forget that datasheets! They aim for a different goal! There are completely different SMPSs even with the same topology. You can't make a correct general decision based on some items not intended for your purposes.
I will make a low power, regulated LLC ZVS SMPS. For me 12V, 5V, 3V3 and maybe -5V is needed. Probably I won't use any LDO, unless the DAC (not choosen yet) requires it, but adding some LDOs can be done any time. Tell if you want to try when ready!
An R core transformer has relatively high primary-to secondary capacitance. Wether this is acceptable or not depends on the whole system.
I see very important possibilities missed in this schematic. With center tapped push-pull topology at switching off a very high and fast impulse is created. This wastes power and generates EMI. Instead of this a simple series LC resonant coupling could provide low loss, low EMI, low primary to secondary capacitance and low droop at the same time.
I dealt with only the transformer section, since I think the main purpose of the circuit is isolation. Why not to do it well then?
How about a SMPS working from 350V/DC ?
You can use a depletion mosfet to get the driver voltage down to a tolerable level for the chip. Supertex shows how to use one for the auxiliary power/bias.
In a world where one is force fed english on a daily basis one needs a guilty pleasure now and then. Sorry. Already trying to maintain 5 languages.
Last edited:
Wow, long post. I thought I was good at that.
Tuning of LDO circuits means choosing the right transformer, the right value filter caps, ultra low drop Schottky diodes, ultra low drop LDO etc. Losses can be made less than usual with some careful thinking. Will never be as efficient as SMPS, that is true. You know what I said about low weight, high efficiency and low leakage current as benefits. Do they really count with 10VA power supplies that much ? A trade off with some specifications for better sound quality is what I have in mind.
There are various nice techno babble words involved with SMPS design but despite the beautiful language that is used I see very little audiophile SMPS either from manufacturers or from DIYers.
All the other items take too time to respond to but let's state that if it was that simple and standard there are surprisingly little audiophile low power SMPS. I wasn't adressing to you personally and the normal power source for the Silentswitcher is a power bank. Never experienced problems with Rcores because of their relatively high primary-to secondary capacitance. I think that problems sometimes are problems if one wishes them to be. You say I should not read data sheets and I can't make a correct general decision based on some items not intended for my purposes. So I need to build many SMPS without reading datasheets ?! Should I build many and see what is best ?
I will try your LLC ZVS resonant SMPS out when it is ready. Promised. Any EMI and it will join the potato peels
How about a SMPS working from 350V/DC ?
I see what you did there
Last edited:
Data sheet of LT1533 specifically mentions it was aimed for greatly reduced conducted and radiated EMI... To my surprise the chip is designed in 1998.
http://cds.linear.com/docs/en/datasheet/1533f.pdf
How to do isolation well ?
Last edited:
I repeat: ZVS, LLC. If you prefer perfection over simplicity: full bridge. This can already be better than the LT solution (both in terms of EMI and efficiency), but you can further decrease EMI by CM choke, shielding, symmetrical transformer, RF grounding.
The 100 uVpp as they measured (in differential mode voltage, after additional filter) doesn't describe EMI compatibility. Common mode current is what counts.
They say only minor increase in loss. But what does this mean? I found one overall efficiency spec in that datasheet:
With LLC achieving 90% is easy.
Efficiency is not real parameter to care about with 10VA PSU's IMHO. The difference between 68 or 90%....It would not be the key parameter when compared to the efficiency of linear PSU's. In other words, anything with better efficiency than linear PSU's would be good enough as long as performance is stunning.
Anyway I put time in thinking this over and see no reason to continue with a 115V/230V input DIY linear/switcher combo. It is way more complex, has some nasty stuff to care about and I am still satisfied with linear PSU's even though they are plagued by magnetic waves (which pose no real threat to me but it costs some metal).
Anyway I put time in thinking this over and see no reason to continue with a 115V/230V input DIY linear/switcher combo. It is way more complex, has some nasty stuff to care about and I am still satisfied with linear PSU's even though they are plagued by magnetic waves (which pose no real threat to me but it costs some metal).
Last edited:
How much less? Can you go significantly above 60% overall efficiency with a 20 VA transformer at 10W output?
No I don't. I don't memorize all of your posts, and takes a hell of a lot time to search on my insanely limited mobile bandwith. I only remember you listed good efficiency as a goal. Low EMI also. And low primary to secondary capacitance is part of low EMI. You also mentioned compactness. I thought you were serious.
How much? You decide. For me small size and lack of 50 Hz disturbance are more important than simplicity and 4 USD cost reduction. I don't know what drawback do you see.
I can't tell I aim for a definitely better sound quality, because in an unknown sound system the effect of a PSU change for sound quality is unpredictable for me. I'm not a mistic high-end guru, only an engineer. I can only try to make a PSU that produces much less disturbance through the already known mechanisms then the existing ones. And I found common mode current the most important disturbance (obviously differential mode noise is more noticeable, but it can be filtered extremely effectively in any kind of PSU).
A babble word for outsiders, very meaningful for people know the area.
So much errors in one sentence, but I try to correct them and respond to the argument hidden inside.
"Little" Do you mean few? "New" ideas arrive very slowly to audio. Most things that are "new" for DIY audio are studied 30-40 years before in the areas of motor drives, control theory, etc...
And audiophile market is negligible compared to PC/multimedia market. Most developers knowing technology well probably simply don't want to waste his time, the rest can't get support in marketing, or afraid of the very fast chinese copycats, etc...
Your fallacy is questionable reason.
I didn't say it was simple, nor standard. I only said every elements to a good (much better then available now) solution are known.
You didn't (or didn't realised), others did. I said it depends on the system. I explained the effect chain with simulation in a case when somebody experienced such problem here on diyaudio. Maybe I will find the topic if you ask.
I didn't wrote not reading any datasheet, but that datasheets (optimized for low power, low price). And you didn't wrote you are reading those datasheets because you are designing with them.
But if you really wanted to design an SMPS, then first study this area, then read as many datasheets as you want. But datasheets don't teach you all the neccessary basics. This is explicitly stated also in the linked LT datasheet.
This is generally unavoidable. Unless you are a genious or very-very lucky. Or unless you can convince yourself that your first try succeed perfectly (which is the nature of human and most people behaves like this).
I will try your LLC ZVS resonant SMPS out when it is ready. Promised. Any EMI and it will join the potato peels
I connected an antenne of an FM receiver to the output of my ClassD amp under development. The only change I could notice is that some weak, noisy stations become clear (because the amp as any conductive material increased the effective area of the antenne). No interference. This is the level of Electromagnetic Compatibility I aim for.
You seem to mix up some things too in trying to make your point but never mind. Again repeating makes me feel old so I won't. The tone is becoming unpleasant and somewhat fanatical. Many engineers display underdeveloped social skills. I was serious in case you have your doubts but like said, I don't continue with this bag of hurt as I already have my linear solution for low power. I do not see the need to put many hours in this. It might occur to you that I rather use laymen's terms instead of techno babble. I think this is more easy to read for non insiders i.e. readers here as you mustn't expect that everyone is of your kind here. In my profession I do the same, it is not forbidden I hope.
Yes, I meant "few". Sorry for the language mistake. Off to Ebay to find myself a superior mains input low power SMPS for audio.
Yes, I meant "few". Sorry for the language mistake. Off to Ebay to find myself a superior mains input low power SMPS for audio.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.