And...
Since this is the Salas Shunt regulator thread we are talking about here: with a hefty current margin in the CCS, and powering a low current draw circuit (line level), grossly oversize transformers will not be a good thing. The CCS current and output impedence of the reg is what is going to matter, even the dynamic demands on the smoothing cap(s) is going to be low/nil with a generous margin in the CCS.
Now oversizing a transformer in a different design may have totally different consequences... with the Salas regs, ~2x current margin in the transformer has proven to be entirley adequate for me, and going bigger has not offered any sonic benefits to my ears. When I do hear benefits is when I make sure there is plenty of margin current in the CCS.
Since this is the Salas Shunt regulator thread we are talking about here: with a hefty current margin in the CCS, and powering a low current draw circuit (line level), grossly oversize transformers will not be a good thing. The CCS current and output impedence of the reg is what is going to matter, even the dynamic demands on the smoothing cap(s) is going to be low/nil with a generous margin in the CCS.
Now oversizing a transformer in a different design may have totally different consequences... with the Salas regs, ~2x current margin in the transformer has proven to be entirley adequate for me, and going bigger has not offered any sonic benefits to my ears. When I do hear benefits is when I make sure there is plenty of margin current in the CCS.
yeah I did say regulator and caps, unless the reg is right next door and you have removed all decoupling caps (sometimes the best thing to do IF the reg is local enough), the decoupling caps and potentially planar PCB capacitance will still be the first thing delivering any current for a transient. depends on the transient
Last edited:
Indeed but I found that using a double sized TX can make a BIG difference even behind a shunt reg.
Quality of the smoothing caps are also audible through the shunts
I do not know if there is a limit for tat but using one 50VA tx for the simplistic is totally different from using a 80VA or even a 120VA
My best builds use two 120VA tx in double mono config.
What'll make you raise your eyebrows a bit is if you change a 50VA "china" Torroid for an "old stock" 50va one, or even more noticeable is the difference of a 50va EI - (haven't tried an R-core yet)
The differences aren't major, but quite clear and if you combine variations of diodes and filters caps, the sum of these things adds up to a quite noticeable overall change, and this is before the Shunt Reg!
There are few of us that are interested in using snubbers (R-C networks) after the diode bridges and the ready availability of fast, soft Shottky diodes makes it a minor detail these days but the old idea of "tuning the transformer windings" has been all but forgotten but it's remarkable to find that this has quite dramatic effects on the overall sound after the Shunt reg - (and a real PIA to do!)
Fascinating things, power supplies!
The differences aren't major, but quite clear and if you combine variations of diodes and filters caps, the sum of these things adds up to a quite noticeable overall change, and this is before the Shunt Reg!
There are few of us that are interested in using snubbers (R-C networks) after the diode bridges and the ready availability of fast, soft Shottky diodes makes it a minor detail these days but the old idea of "tuning the transformer windings" has been all but forgotten but it's remarkable to find that this has quite dramatic effects on the overall sound after the Shunt reg - (and a real PIA to do!)
Fascinating things, power supplies!
Indeed but I found that using a double sized TX can make a BIG difference even behind a shunt reg.
Quality of the smoothing caps are also audible through the shunts
I do not know if there is a limit for tat but using one 50VA tx for the simplistic is totally different from using a 80VA or even a 120VA
My best builds use two 120VA tx in double mono config.
+1
Guys, lets keep potentially offensive opinions out of the mix.
People who post here come from many different backgrounds. Some need a lot of help, some don't. I try to be someone who always helps (even though it's not usually technical). A lot of people private message because they are afraid of being brow beaten on the list. People have a hard time getting a lot of concepts. We learn by analogies, not often well done in one comment. We learn by making some mistakes, and best to be caught on the list doing so then releasing the white smoke.
I hope that you come hear to either make good audio, or help others get there - either by good assistance or opinion.
People who post here come from many different backgrounds. Some need a lot of help, some don't. I try to be someone who always helps (even though it's not usually technical). A lot of people private message because they are afraid of being brow beaten on the list. People have a hard time getting a lot of concepts. We learn by analogies, not often well done in one comment. We learn by making some mistakes, and best to be caught on the list doing so then releasing the white smoke.
I hope that you come hear to either make good audio, or help others get there - either by good assistance or opinion.
Indeed but I found that using a double sized TX can make a BIG difference even behind a shunt reg.
Quality of the smoothing caps are also audible through the shunts
I do not know if there is a limit for tat but using one 50VA tx for the simplistic is totally different from using a 80VA or even a 120VA
My best builds use two 120VA tx in double mono config.
+2
Tea, that's great! I found some of these boards next to the DCB1. Although I can navigate the complex stuff, some of the simple stuff trips me up! Go figure.
I'm building the Marsh headphone amp and Richard Marsh said a shunt reg circuit should work great with his head amp. But each channel is powered separately by V+/V- rails and I'm not sure how to wire up a V+/V- SSLV1.1 to two separate boards for the head amp. I'm not there yet but when I'm ready, it will be great to know exactly how it should all be set up
I'm building the Marsh headphone amp and Richard Marsh said a shunt reg circuit should work great with his head amp. But each channel is powered separately by V+/V- rails and I'm not sure how to wire up a V+/V- SSLV1.1 to two separate boards for the head amp. I'm not there yet but when I'm ready, it will be great to know exactly how it should all be set up
As qusp said, I thought better of it afterwards but that post was already quoted. My intended meaning for that post has been misunderstood I fear, though I wanted to leave it and move on, there is little to be gained now.
I do not ever want to be seen to be discouraging people from asking a question. Recognising what you do not know is a very important skill and asking questions is a large part of what this forum is about!
I can imagine that the construction of two transformers of similar size (toroid vs r-core/EI or china toroid vs quality toroid) could make a difference. This could be for a number of reasons. The construction of the transformer will change it's interaction with noise from other circuits on the other secondary windings via mutual inductance of the windings. The change in tx type will also change the amount of radiated noise inside an enclosure.
I don't have a lot of experience with snubbers but I will say this, the snubber is not only filtering noise from the circuit that it is supplying. It also filters noise from the rectifier flowing back into the transformer. This, depending on other circuits on the same transformer (it will also add to the mains noise) may also change the performance of other devices in your system. You are damn right that they are a PITA to do well, I have not got the gear to measure properly the transformers that I have to be able to add them to my power supplies, yet.
Both of these, changes to high quality transformers and snubbers, (where values are actually chosen for specific measured tx properties) are far more worthwhile undertakings than brute force by simply adding larger VA toroidal transformers. The performance of these components contributes to the noise seen at the output and IMHO that is more likely to produce an observable change in performance.
I do not ever want to be seen to be discouraging people from asking a question. Recognising what you do not know is a very important skill and asking questions is a large part of what this forum is about!
I can imagine that the construction of two transformers of similar size (toroid vs r-core/EI or china toroid vs quality toroid) could make a difference. This could be for a number of reasons. The construction of the transformer will change it's interaction with noise from other circuits on the other secondary windings via mutual inductance of the windings. The change in tx type will also change the amount of radiated noise inside an enclosure.
I don't have a lot of experience with snubbers but I will say this, the snubber is not only filtering noise from the circuit that it is supplying. It also filters noise from the rectifier flowing back into the transformer. This, depending on other circuits on the same transformer (it will also add to the mains noise) may also change the performance of other devices in your system. You are damn right that they are a PITA to do well, I have not got the gear to measure properly the transformers that I have to be able to add them to my power supplies, yet.
Both of these, changes to high quality transformers and snubbers, (where values are actually chosen for specific measured tx properties) are far more worthwhile undertakings than brute force by simply adding larger VA toroidal transformers. The performance of these components contributes to the noise seen at the output and IMHO that is more likely to produce an observable change in performance.
When Jeremy mentioned using smaller Txrs and smaller caps for some situations, it brought to mind that there is a quite consistent attitude to power supplies in the French diy scene that advocates non electro caps in power supplies (adding chokes tho) and obtaining superior results
So, as Tea and others have mentioned before a few times (!) "there's many different ways to audio nirvana" (where did that come from?!)
What I found quite surprising is that changes/variations before the shunt reg would be so obvious - I expected the shunt filter to drown out any effects of diodes, caps, etc and could mostly ignore the first part (or "raw" supply) considerations, provided it was respectable.
Now I will add that the basic supply + shunt reg produces pretty good results, no doubt at all, and adding bigger trx, soft/fast diodes, etc is like 'icing on the cake', a bonus, and it just depends where to focus resources, unless you're a "upgrade nutter" like me!
It's great that we don't have any Trolls on these threads - very pleasant and thanks to "The Salas" (he's earned a title, I think!).
So, as Tea and others have mentioned before a few times (!) "there's many different ways to audio nirvana" (where did that come from?!)
What I found quite surprising is that changes/variations before the shunt reg would be so obvious - I expected the shunt filter to drown out any effects of diodes, caps, etc and could mostly ignore the first part (or "raw" supply) considerations, provided it was respectable.
Now I will add that the basic supply + shunt reg produces pretty good results, no doubt at all, and adding bigger trx, soft/fast diodes, etc is like 'icing on the cake', a bonus, and it just depends where to focus resources, unless you're a "upgrade nutter" like me!
It's great that we don't have any Trolls on these threads - very pleasant and thanks to "The Salas" (he's earned a title, I think!).
Here's a building thread for the SSLV1.1 BiB PCB reg. Its to discuss builds, ask questions, tell stories.
*''Lazybutt'' Fred's Excel calc to help you configure has been added on 1/14/12
**The pdf guide on GoogleDocs
Salas...I am trying to build a pair of +SSLV1.1 (6.5V - 7.0V; at 1.5A + safety) to power the filaments of a pair of VT25A or 10.
As suggested, I am trying to use the calculator....I entered as follows; and the VDC out = 2.....does it sound right?? and have I done anything not correct??
Pre-filter calculatorrsion 1a 01/15/12
DC Feed Parameters VAC In (Tx) 12 Vac RMS AC Peak 17.0 V
Target Current 17000 mA
Rectifier Drop 1 V DC Rectified 16.0 V
Tx DC Resistance 0.9 Ohm R Tx drop 15.3 V
L Filter DC Resistance 0 Ohm L Filter drop 0 V
R Filter 0 Ohm R filter drop 0 V
P(R filter) #DIV/0! W
V DC Out 2 Vdc
Pre-filter calculatorrsion 1a 01/15/12
DC Feed Parameters VAC In (Tx) 12 Vac RMS AC Peak 17.0 V
Target Current 17000 mA
Rectifier Drop 1 V DC Rectified 16.0 V
Tx DC Resistance 0.9 Ohm R Tx drop 15.3 V
L Filter DC Resistance 0 Ohm L Filter drop 0 V
R Filter 0 Ohm R filter drop 0 V
P(R filter) #DIV/0! W
V DC Out 2 Vdc
DC Feed Parameters VAC In (Tx) 12 Vac RMS AC Peak 17.0 V
Target Current 17000 mA
Rectifier Drop 1 V DC Rectified 16.0 V
Tx DC Resistance 0.9 Ohm R Tx drop 15.3 V
L Filter DC Resistance 0 Ohm L Filter drop 0 V
R Filter 0 Ohm R filter drop 0 V
P(R filter) #DIV/0! W
V DC Out 2 Vdc
Well said my friend....I am a reproductive biologist (of marine fish and prawn) by training who just love live music in my living room...when not go to concerts.....I was using commercial units (C.J. & Marantz 7...etc) ....but still have a feel that all is not there.....and then after searching in the net for a while...I started my first 10 type tube line amp (that must have been about 10 years ago now.....and after switching on my first 10 type line amp...I right away sold my Marantz 7.....ever since I have worked on my 10/VT-25A line amp...and never looked back.....this is DIY....and I learn a lots about E.E. (but never enough...and always something new everyday or everywhere you care to learn (about the facts but not the ads).....learn the most from Salas....and benefit from boards from Tea-Bag....
I dont think there were any, he probably just copied the values from the excel/salas V1 calculator. copying from Excel often creates unpredictable results because of all the tabs and internal formatting, creating large spaces in the text.