Of course it was designed to work that way.
And of course I pass the wall socket grounding through, but it goes out just as it came in, no better, no worse. Remember, in Europe, we mostly use German type of DIN sockets with two central prongs and ground on the sides.
BTW, I note that those Dutch devices use the same sockets as I do, with spring loaded cover for when they are not being used. They are rated at 25 Amps continuous, 40 Amps peak (1 in 4 cycles, T= 1 sec).
Yes ...? 
I've done many experiments over the years, and it's an evolving approach, philosophy, still. Everything works to some degree, in some areas - the trick is to get the 'killer' solution; that is, a total "mechanism" that does the job 100% in all situations - and that's something I can't say I have, at the moment. Possibly, if I combine everything I've considered then it will be bulletproof - but I haven't got that at the moment.
A simple solution I use for the low end things at the moment is the borin' ol' surge protectors, with MOV's - they help a bit. Quite often I just turn off and disconnect the interfering devices - the simplest solution of all!
What I'm after in any situation is sound that's "good enough", for my mood at the time. If I had an extremely expensive setup, I would expect, demand, nigh 'perfect' results from it; with something much less pretentious I will excuse the less irritating 'failings' ...
I've done many experiments over the years, and it's an evolving approach, philosophy, still. Everything works to some degree, in some areas - the trick is to get the 'killer' solution; that is, a total "mechanism" that does the job 100% in all situations - and that's something I can't say I have, at the moment. Possibly, if I combine everything I've considered then it will be bulletproof - but I haven't got that at the moment.
A simple solution I use for the low end things at the moment is the borin' ol' surge protectors, with MOV's - they help a bit. Quite often I just turn off and disconnect the interfering devices - the simplest solution of all!
What I'm after in any situation is sound that's "good enough", for my mood at the time. If I had an extremely expensive setup, I would expect, demand, nigh 'perfect' results from it; with something much less pretentious I will excuse the less irritating 'failings' ...
What the man said. 'Nice clean earth' is a popular audio myth - the reality is grounds are just sewers for everyone's (including one's own) electrical rubbish. Common-mode filtering, its the only way to go.
When I see a supposedly 'high end' audio component with one of those filter 'blocks' with an IEC inlet and some internal caps and chokes, I'm chuckling to myself. It tells me they probably just stuck it in there without much thought. Those pre-packaged filters get their datasheet figures for mains noise suppression from testing labs where's there is a decent groundplane to make the measurement. Given a noisy earth (most people's setups) those figures go out the window pronto.
I was intrigued with this beast, EARTH HENRY - AC power earth-line noise reducer - By Ben Duncan, at the time. No-one else seems to do something like it, I don't know how effective it is, and I haven't tried doing a version like it for myself ... any thoughts ...?
I was kinda joking there - sure the very high frequency performance of the coiled up extension cable is going to suck because the dielectric (PVC usually) is very poor. So the coil gets lossy. But then there's the chance to sell audiophile grade spacers for the cable isn't there?
I dunno what problem that coil is going to fix. Certainly I use a ground lifting coil in my system, but its a segmented choke which gives a much higher impedance well below 100kHz. But it has been known to give rise to problems - particularly when I connected my laptop to my DAC when the laptop was plugged in. The leakage current from the SMPSU gave rise to an appreciable common-mode voltage on my amp's 0V.
You really have to know what problem you have in order to know how to fix it. Which is pretty much dvv's attitude with his filters - try them and see. A ground lift coil it seems to me could be a hazard or a boon, depending on the system and whether the primary sources of noise are within the system (like an LCD TV connected via HDMI to a receiver) or beyond.
I dunno what problem that coil is going to fix. Certainly I use a ground lifting coil in my system, but its a segmented choke which gives a much higher impedance well below 100kHz. But it has been known to give rise to problems - particularly when I connected my laptop to my DAC when the laptop was plugged in. The leakage current from the SMPSU gave rise to an appreciable common-mode voltage on my amp's 0V.
You really have to know what problem you have in order to know how to fix it. Which is pretty much dvv's attitude with his filters - try them and see. A ground lift coil it seems to me could be a hazard or a boon, depending on the system and whether the primary sources of noise are within the system (like an LCD TV connected via HDMI to a receiver) or beyond.
Yes, it's all very suck-it-'n'-see stuff - there's no universal solution ... as yet. Obviously it, the Earth Henry, didn't work brilliantly enough for sufficient people - otherwise it wouldn't have virtually disappeared, as it has.
The simple answer is that every part of an audio system, every aspect that has a physical nature, has to be totally neutered,, with regard to be any sort of electrical activity having a poke at it, that's apart from the audio signal manipulation. Obvious, simple to say, but damned difficult to get 'just right' ...
The simple answer is that every part of an audio system, every aspect that has a physical nature, has to be totally neutered,
, with regard to be any sort of electrical activity having a poke at it, that's apart from the audio signal manipulation. Obvious, simple to say, but damned difficult to get 'just right' ...
This is a truly grey area, Rick, those IEC sockets.
Technically, you can't bitch about them because they do work, but the problem is that they kick in after 99.9% of the garbage has already gone by. Typically, they kick in at 600...700 kHz, but according to my experience, some 95-96% of the unwanted junk is done by around 80 kHz, which includes harmonics.
The usual excuse for those IEC sockets is that we are surrounded by ultrasonic noise from wireless table 'phones, cell 'phones, radio, TV, etc. While all that is true, no-one seems to know what power levels of those disturbances are we talking about. Not the same thing, I hope you'll agree, whether it's a cell phone with a very low power capability, or a ham radio which is not a joke any more.
Agreed on common mode as being the best all around solution. That's what I use.
Last edited:
All done in 1999-2001. Several versions made, from similar variations to different versions. In the end, overall the best was the version I now use, even if that evolved a bit as well.
I like to periodically set aside the time to muse over them, see if I can come up with something useful. If not, no hard feelings, I still think of it as time well spent.
I do exactly the same with my prototype version of my headphones amp. Yes, it's really good, but I have yet to encounter anything technical which could not be made better still. That's why I always keep the prototype, it isn't pretty and all dressed up, but it alllows me easy access to everything, so I can play with it.
And there's no race like the race to outrun yourself.
I would like to answer, but I honestly cannot because I don't remember. Measuring much of what I wanted measured required specialized lab equpment which I do not have, so I let a local institute working on such matters do the whole measuring gig.
They did and came back with an impressive set of measurements. And a warning: power quality widely varies across the city of Belgrade alone (app. 2 million pop.), and varies wider still in the 60 km radius.
I was primarily interested in the filter's behavior on the 50 Hz to 100 kHz range. They said that most of the line noise did what it did and was done by around 80 kHz, including harmonics. By then, the filter was down at over -60 dB, while at 20 kHz, the 6 supplied samples were down 29-31 dB, advertised value being "-27 dB or better".
Their findings were very interesting in other aspects. For example, they claimed that PCs were the worst line polluters, followed by some (mind - not all) tube gear, due to thermal noise problems in some designs, and air cons and vacuum clenars were not far behind. With often junk grade motors built into them, I am not surprised.
Subsequent 12 years of practice have validated almost all of their findings.
I think I should talk to them again, in the meanwhile I have gathered more questions, and have changed the filter to a fair degree, perhaps evolved is a more precise description. Time to check up on myself.
Last edited:
Finally, I managed to get the Current Factory Mk. II under way.
There are several changes, so I attach both versions for review.
The input stage CCS has been completely redone, as this new version provides for a still better CMRR figure, admittedly only 2 dB in comparison with the old one, but I am now down to fighting for every dB I can get.
The second BIG change is that Q1 and Q4 of old, previously 2N5551 bipolars, have been swapped for 2SK170 FETs. Thorsten pinted out to me that there was in fact a very low voltage across these transistors, and it is indeed so, just 3.2V. This is child’s play for the FETs, as they can handle 40V, but the entire circuit gains by lower THD and IM figures, and shows sigificant improvement in its DC stabiility – just as you might expect from FETs. Why I did not see this myself is a mystery to me, but either way, thank you, Thorsten.
Some may ask why the old 2SK170? Well, if any FET has proved its worth in audio time and time again, 2SK170 would be the one. It is still available, unlike many others. And it's cheap enough to buy in some quantity to be able to hand match them to rather close tolerances.
The third change is that the simple zener based CCS for the input stage cascode now has a decoupling (zener to cascode) transistor, which should help remove the last traces of zener noise, theoretically at least.
The fourth change is that I have simply added to the schematic what I said was missing from the first version, and that is the 8th pair of output trannies. They are clearly shown now.
The fifth change is that somewhere along the line I realized that the protection output relay, rated at 16 Amps, will become a bottleneck for any serious listening at any higher volume. Since I would rather die than give it up, there was nothing for it but to add a second relay, doubling the nominal output current to 32 Amps continuous, and 44 Amps in peaks. If anyone is going to need that, it will be Wayne. And, last but not least, this paralleling halves the output impedance of the realys, usually not a problem, but might well be a problem with 1 Ohm loads.
A note here. The schematic shows a total of 3 protection fuses. All 3 are probably not really needed, but there are those why say that just the PSU lines need to be protected, and there are those who feel that a series fuse with the output is the only real deal. I don’t want to be the judge of who’s right, so I leave that to the builder – you decide. You can use only the series fuse, only the PSU line fuses, or all three, your choice. Personally, I’d settle on the line fuses only, but that’s just me.
Overall THD and IM figures are a little down over the first version, but that’s to be expected. Overall, it still uses just less than 26 dB of global NFB, and the open loop full power response is out to 50 kHz. It is still limited to around 200 kHz by the input filter, and I would strongly advise this to remain so. If Wayne wished to lower the cut-off frequency, he can simply increase the value of C12 to wherever he wants.
If this is deemed satisfactory, the next step would be to get in touch with Alex in Romania. The man is WAY more than a competent technician, he is in my view a true artist. Immaculate! I have seen some great PCB artwork in my time, and I believe Alex to be up there with the best. He can supply all required GERBER files to enable local manufacturing almost anwhere, which should be good for Wayne, but we can discuss that along the way. I would say that for somethig like this, 70 microns of copper would be the very least to even think about, but if I could, I’d go for 100 or even 130 microns.
Then we come to most difficult aspect – hands-on development. Given the distance between Wayne and me, and between Wayne and most of us in general, it appears Wayne will have to the hands doing the work, with out wise assistence to back him up.
Obviously, I will help with any and all parts in case there's a supply problem.
If anyone has an idea, this would be a GREAT time to let us know.
I’m off to finish my own Centurion project, which I had to set aside for this one.
There are several changes, so I attach both versions for review.
The input stage CCS has been completely redone, as this new version provides for a still better CMRR figure, admittedly only 2 dB in comparison with the old one, but I am now down to fighting for every dB I can get.
The second BIG change is that Q1 and Q4 of old, previously 2N5551 bipolars, have been swapped for 2SK170 FETs. Thorsten pinted out to me that there was in fact a very low voltage across these transistors, and it is indeed so, just 3.2V. This is child’s play for the FETs, as they can handle 40V, but the entire circuit gains by lower THD and IM figures, and shows sigificant improvement in its DC stabiility – just as you might expect from FETs. Why I did not see this myself is a mystery to me, but either way, thank you, Thorsten.
Some may ask why the old 2SK170? Well, if any FET has proved its worth in audio time and time again, 2SK170 would be the one. It is still available, unlike many others. And it's cheap enough to buy in some quantity to be able to hand match them to rather close tolerances.
The third change is that the simple zener based CCS for the input stage cascode now has a decoupling (zener to cascode) transistor, which should help remove the last traces of zener noise, theoretically at least.
The fourth change is that I have simply added to the schematic what I said was missing from the first version, and that is the 8th pair of output trannies. They are clearly shown now.
The fifth change is that somewhere along the line I realized that the protection output relay, rated at 16 Amps, will become a bottleneck for any serious listening at any higher volume. Since I would rather die than give it up, there was nothing for it but to add a second relay, doubling the nominal output current to 32 Amps continuous, and 44 Amps in peaks. If anyone is going to need that, it will be Wayne. And, last but not least, this paralleling halves the output impedance of the realys, usually not a problem, but might well be a problem with 1 Ohm loads.
A note here. The schematic shows a total of 3 protection fuses. All 3 are probably not really needed, but there are those why say that just the PSU lines need to be protected, and there are those who feel that a series fuse with the output is the only real deal. I don’t want to be the judge of who’s right, so I leave that to the builder – you decide. You can use only the series fuse, only the PSU line fuses, or all three, your choice. Personally, I’d settle on the line fuses only, but that’s just me.
Overall THD and IM figures are a little down over the first version, but that’s to be expected. Overall, it still uses just less than 26 dB of global NFB, and the open loop full power response is out to 50 kHz. It is still limited to around 200 kHz by the input filter, and I would strongly advise this to remain so. If Wayne wished to lower the cut-off frequency, he can simply increase the value of C12 to wherever he wants.
If this is deemed satisfactory, the next step would be to get in touch with Alex in Romania. The man is WAY more than a competent technician, he is in my view a true artist. Immaculate! I have seen some great PCB artwork in my time, and I believe Alex to be up there with the best. He can supply all required GERBER files to enable local manufacturing almost anwhere, which should be good for Wayne, but we can discuss that along the way. I would say that for somethig like this, 70 microns of copper would be the very least to even think about, but if I could, I’d go for 100 or even 130 microns.
Then we come to most difficult aspect – hands-on development. Given the distance between Wayne and me, and between Wayne and most of us in general, it appears Wayne will have to the hands doing the work, with out wise assistence to back him up.
Obviously, I will help with any and all parts in case there's a supply problem.
If anyone has an idea, this would be a GREAT time to let us know.
I’m off to finish my own Centurion project, which I had to set aside for this one.
Last edited:
Big centurion or little centurion .... ?
O tempora, o mores!
That "little" version uses 3 series/parallel 200W devices, and is "little"? Believe it or not, Wayne, but for 99% of us, that is more than enough. In the industry, very, very few manufacturers go that far, the usually best you can hope for are 3 pairs of 130-150W devices.
And if you're called Marantz, and the model is their upmarket 11S, you would use just one pair of 200W devices for nominally 100/200W into 8/4 Ohms. Just how one squeezes 200 Watts of effective power from a single pair is completely beyond me, they must have installed freon cooling in the trannies.
How would such a product survive the IHF power test, I have no idea, remember, 30 minutes of running at 1/3 nominal power prior to measuring maximum power.
- Status
- Not open for further replies.