Hi,
I guess it depends on what you want.
The C-02 is a really minimalist design (Tone controls non-withstanding), folded cascode circuitry and with an MC Stepup Transformer and overall uses lower levels of global NFB.
The C-03 by comparison uses generic (and fairly low grade) Op-Amp's for the Phono Stage and Tone controls (which arguably can be bypassed), has by far more switches in the signal path and all.
But, if you only use the Input switching, volume control and line stage there is likely not a lot between the two.
Currently playing with folded cascode VAS (or precisely "VAS-Less") circuitry in the Sim. A bit twitchy for DC stability, but I like the harmonic distribution better. Also have been looking at clipping recovery in the Sim.
Ciao T
I guess it depends on what you want.
The C-02 is a really minimalist design (Tone controls non-withstanding), folded cascode circuitry and with an MC Stepup Transformer and overall uses lower levels of global NFB.
The C-03 by comparison uses generic (and fairly low grade) Op-Amp's for the Phono Stage and Tone controls (which arguably can be bypassed), has by far more switches in the signal path and all.
But, if you only use the Input switching, volume control and line stage there is likely not a lot between the two.
Currently playing with folded cascode VAS (or precisely "VAS-Less") circuitry in the Sim. A bit twitchy for DC stability, but I like the harmonic distribution better. Also have been looking at clipping recovery in the Sim.
Ciao T
The key point being that it's moving along - that's what I wanted to hear.
After much simulation, I ended up with classic cascodes at the input, the voltage being stabilizied not by the usual voltage divider, but by a double filtered (22uF/16V tantalum in parallel with 0.1 uF) zener diode. This model returned slightly better results, although the classic method with the voltage divider is said to have a better PSRR. But I reckon I make up for it by using shunt stabilized power supplies for the voltage gain stages.
I am still undecided regarding full electronic stabilization of the current gain stages. It seems there are exactly as many pros as there are cons. It will require splitting a stereo amp into two mono blocs simply because it uses up quite a bit of space. The alternative would be using either one massive (1,000+ VA) toroid, with separate secondaries for each channel, or two separate smaller toroids (500+ VA) for a complete dual mono. Decisions, decisions, ...
As for switches and knobs, don't let it fool you - most of the switches are related to electronically switched inputs, selection of 1 or 2, etc. Of the rest, it has the more or less usual four switches for stereo/L+R (which I consider as essential for properly adjusting balance), loudness, tone in/out and subsonic filter. The two on the extreme right deal with phono type and load and the Rec Out selection, meaning that as far as I am concerned, they may as well not even be there, but thankfully, both have an OFF position - which is where they will stay.
Re: Luxman preamp. What's done is done, there's no turning back. I will have to refresh the preamp, and to do that I'll need its service schematic. I'm hunting for one right now. I can always use that opportunity to beef it up a bit, both using better components and larger capacitors, perhaps exchanging a few critical resistors, things like that. I can't really know until I see its schematic.
All I can say for sure at this point is that its two 3,300/50V uF caps in the power supply will grow in capacity. I have a reasonable stock of both Fischer & Tausche and Panasonic power supply caps, but if I can alway eschew them for something better still. I'm not shy in that way.
After much simulation, I ended up with classic cascodes at the input, the voltage being stabilizied not by the usual voltage divider, but by a double filtered (22uF/16V tantalum in parallel with 0.1 uF) zener diode. This model returned slightly better results, although the classic method with the voltage divider is said to have a better PSRR. But I reckon I make up for it by using shunt stabilized power supplies for the voltage gain stages.
I am still undecided regarding full electronic stabilization of the current gain stages. It seems there are exactly as many pros as there are cons. It will require splitting a stereo amp into two mono blocs simply because it uses up quite a bit of space. The alternative would be using either one massive (1,000+ VA) toroid, with separate secondaries for each channel, or two separate smaller toroids (500+ VA) for a complete dual mono. Decisions, decisions, ...
As for switches and knobs, don't let it fool you - most of the switches are related to electronically switched inputs, selection of 1 or 2, etc. Of the rest, it has the more or less usual four switches for stereo/L+R (which I consider as essential for properly adjusting balance), loudness, tone in/out and subsonic filter. The two on the extreme right deal with phono type and load and the Rec Out selection, meaning that as far as I am concerned, they may as well not even be there, but thankfully, both have an OFF position - which is where they will stay.
Re: Luxman preamp. What's done is done, there's no turning back. I will have to refresh the preamp, and to do that I'll need its service schematic. I'm hunting for one right now. I can always use that opportunity to beef it up a bit, both using better components and larger capacitors, perhaps exchanging a few critical resistors, things like that. I can't really know until I see its schematic.
All I can say for sure at this point is that its two 3,300/50V uF caps in the power supply will grow in capacity. I have a reasonable stock of both Fischer & Tausche and Panasonic power supply caps, but if I can alway eschew them for something better still. I'm not shy in that way.
Last edited:
Hi,
Not necessarily along, but moving, possibly more sideways or circular, who knows, maybe even funicular...
Try my way, CLC...
I'm not fooled. I looked at the actual switches in the direct signal circuit, that is how many contacts must be traversed... The C-03 is much worse than the C-02 which I already consider to have "way too many"...
Ciao T
Not necessarily along, but moving, possibly more sideways or circular, who knows, maybe even funicular...
Try my way, CLC...
I'm not fooled. I looked at the actual switches in the direct signal circuit, that is how many contacts must be traversed... The C-03 is much worse than the C-02 which I already consider to have "way too many"...
Ciao T
Amp: Eppur si muove.
CLC: Don't mind if I do. Any ideas? I have a LOT of Fischer&Tausche 10,000 uF/63V and 4,7oo uF/63V caps ....
Switches: Well, thank God I am not the purist you are.
I accept my fate stoically. 
And I'm still a small timer, Dan d'Agostino uses 118 of them altogether in the voltage gain stages.
I played a lot with the design, trying out different transistors and circuits, and I discovered that using 3 MPSA92/42 trannies reduces high frequency distortion to an extent, nothing spectacular, but it's there. And, being the Motorola/ON Semi affictionado, I just happen (by sheer chance, of course) to have about a suitcase full of them, so matching them is reduced to coffee sipping, cigar puffing, music listening nicely spent time of manual labor.
Is this three trannies in parralel ?? Maybe to replace for higher Ic part ??? Is this in real measured performance or simmed ?? Krell has used this before for the purpose too, but in their case they were looking at near medium levels of Ic I havent counted but in one of their amps they probably used over a 100 of them (six or seven in parallel).
Yes, three MPSA 92 and three MPSA 42 trannies in parallel, since it's a fully complementary design.
No, it does not replace anything, and especially not an IC. The only two ICs it uses is an LF411 for DC Servo and a an LM 393 comparator for overheat and DC protection, meaning they are out of the actual signal path.
I honestly don't know which ICs you refer to. The only Krells I know what they are made of are the FPB series from the late 90-ies and early 2000s. In them, the entire basic audio signal path is made 100% of dsicrete transistors, while ICs are used for FPB (Floating Point Bias), protection and other strictly secondary functions. The voltage amp consists exclusively of small signal Motorola/ON Semi devices, the MPSA an 80xx families. There are 118 of them in the voltage amp alone.
Perhaps you refer to another series?
By IC i meant max allowable current. I honestly cant see those 3 paralleled to92s say outperform a transitor like 2sa1381/2sc3503 at high or low frequency, it just doesnt make sense. Three to92s will have higher Cob capacitance and ultimately worse performance.
Thats the reason I ask if this is in real life or in sim. In sim models the value of cjc is in many cases grossly mistaken.
Its exactly the krells I was thinking of.
Thats the reason I ask if this is in real life or in sim. In sim models the value of cjc is in many cases grossly mistaken.
Its exactly the krells I was thinking of.
The reason why I use so few Japanese transistors is quite simply that they are unavailable to me where I live. And frankly, I'm sick and tired of having to import every little ninny each and every time.
As a point of interest, I believe that transistors do have a sound of their own, i.e. that they are NOT absolutely neutral in all respects. As most people (I think?), I tend to use devices I am well acquainted with, but more thjan that, devices which I can easily get hold of.
The local market has been taken over many decades ago by European manufacturers, and consequently, I am well acquainted with many (but certainly not all, not even most) European transistors. The only non-European company which rules with a strong hand locally is Motorola/ON Semi, they've been present here since the early 70ies.
I have yet to see and hear a transistor with better HF performance than the BF 720/721. They are grossly misrepresented in catalogs, in real life they do a LOT better than stated, assuming only you buy them from Siemens of Philips (no biggie, Farnell has them). Their only inconvenience is that they are intended for SMD, but, with a little patience nd a fine soldering iron tip, that's not really a big problem.
Of course, if I had a decent selection of Japanese transistors readily available, I would certainly give them a much more serious look.
Oddly enough, Dan d'Agostino didn't mind having many of them set in parallel, and nobody can call him an amateur. Not at his prices.
Dejan,
How true. It does feel like the holy inquisition here, at times...
Just get some local maker of high power inductors to make you something on those rare earth based torroids, which only take a few ten turns for a millihenry or so at 10 Ampere and a few 10th of milliohm DCR.
Use them so you have one bridge made from fast/soft switching diodes directly feeding one capacitor rated at enough ripple current. Then have one choke in each line (+/-) and follow with a second capacitor. The second Cap should be both correctly bypassed and snubbered. Create the local start ground by connecting the two separate supplies together.
For example my PSU will have per 56V line (that is four of them +/- Left and +/- Right) a bridge made from TO-247 encases 25A RUR-XXX Diodes with an additional 40A schotty diode connected in series. I call this kind of rectifier "schottky enhanced", the Schottky diode tends to block the reverse recovery spike.
Then 18,000uF/71V Elna "For Audio", then 2pcs 1mH Inductor, then another 18,000uF/71V Elna "For Audio". I did test these cap's on an LCR Analyser, they are very good and I also know the ESL values now.
The bypass will use some very compact Japanese made MKP 4uF Capacitors, they have phantastically low ESL/ESR and to snub the invariable impedance spike between the two cap's I will add a few 15uF/100V Electrocube Mylars, which have low ESL but high ESR, probably 4pcs per rail, precise number will depend on measured results, too many can add resonances back.
I will have to experiment with the bypass combo using our RF generator and 150MHz 'scope, it woirks quite horrorshow in the Sim.
I am quite worries about mechanical switches, especially ones that are open to the air. ideally you minimise any current flowing through them.
If I could I would avoid having any switches in series with signal lines (ask Ed Simmons [Simon7k] one day why, if you feel like) and probably use J-Fet input buffers where I simply switch off the supply for the buffer and have a shunt relay on the output of the buffer after maybe a 1K resistor, then a suitable series resistor before the volume control maing a classic summing circuit.
Luckily I have a small supply of mercury wetted reed relays remaining, I will be using these for my next pre. Sadly ROHS has comply nixed these from availability, even before ROHS prices where getting quite bezoomny.
But there are reasons why, where long term accuracy of measuring a few dozend uV per centigrade from thermocouples with better than point one degree accuracy only such relays where acceptable. Given the very real dynamic ranges in audio these should be "Audio required use" items.
Ciao T
Amp: Eppur si muove.
How true. It does feel like the holy inquisition here, at times...
Just get some local maker of high power inductors to make you something on those rare earth based torroids, which only take a few ten turns for a millihenry or so at 10 Ampere and a few 10th of milliohm DCR.
Use them so you have one bridge made from fast/soft switching diodes directly feeding one capacitor rated at enough ripple current. Then have one choke in each line (+/-) and follow with a second capacitor. The second Cap should be both correctly bypassed and snubbered. Create the local start ground by connecting the two separate supplies together.
For example my PSU will have per 56V line (that is four of them +/- Left and +/- Right) a bridge made from TO-247 encases 25A RUR-XXX Diodes with an additional 40A schotty diode connected in series. I call this kind of rectifier "schottky enhanced", the Schottky diode tends to block the reverse recovery spike.
Then 18,000uF/71V Elna "For Audio", then 2pcs 1mH Inductor, then another 18,000uF/71V Elna "For Audio". I did test these cap's on an LCR Analyser, they are very good and I also know the ESL values now.
The bypass will use some very compact Japanese made MKP 4uF Capacitors, they have phantastically low ESL/ESR and to snub the invariable impedance spike between the two cap's I will add a few 15uF/100V Electrocube Mylars, which have low ESL but high ESR, probably 4pcs per rail, precise number will depend on measured results, too many can add resonances back.
I will have to experiment with the bypass combo using our RF generator and 150MHz 'scope, it woirks quite horrorshow in the Sim.
Switches: Well, thank God I am not the purist you are.
I am quite worries about mechanical switches, especially ones that are open to the air. ideally you minimise any current flowing through them.
If I could I would avoid having any switches in series with signal lines (ask Ed Simmons [Simon7k] one day why, if you feel like) and probably use J-Fet input buffers where I simply switch off the supply for the buffer and have a shunt relay on the output of the buffer after maybe a 1K resistor, then a suitable series resistor before the volume control maing a classic summing circuit.
Luckily I have a small supply of mercury wetted reed relays remaining, I will be using these for my next pre. Sadly ROHS has comply nixed these from availability, even before ROHS prices where getting quite bezoomny.
But there are reasons why, where long term accuracy of measuring a few dozend uV per centigrade from thermocouples with better than point one degree accuracy only such relays where acceptable. Given the very real dynamic ranges in audio these should be "Audio required use" items.
Ciao T
Thorsten, don't forget thany before any grid power even reaches the electronics, it has to pass through one of my line filters.
In effect, these are made much along the lines you wrote above. The effects of the filter can be seen, if so desired, on a 'scope rather well, and they can be impressive. However, this impressive fact does not improve the sound of many devices as much as the 'scope view simply because they are poorly made from the outset.
It's like dressing a hobo in an Armani suit - you get a hobo in an Armani suit, no more.
But they do their job admirably, I guarantee it. This means that my power supply arrives at the device's transformer already filtered from whatever happens to travel along the grid lines at the moment. With a rating of 20A/230V EACH, let's say I have no immediate problems with overloads, capacities, etc, I am well taken care of.
Since you, to the best of my knowledge, do not use an outside filtering device, obviously your problems are the same as mine, but you have to solve them inside the power amp, preamp, or whatever. I have no doubt at all that your system work very well, but in my case, that would be putting a second system in series with the first.
Very often, this can cause unpredictable and never pleasant effects. In brief, it can kill the sound dead as a doornail. I know - I tried putting both mine and other similar devices in series, and never once got anything but a catastrophe back.
Which is why I will pass on this matter, with the knowledge that I already have the problem well under control. Also using LC filtering.
In effect, these are made much along the lines you wrote above. The effects of the filter can be seen, if so desired, on a 'scope rather well, and they can be impressive. However, this impressive fact does not improve the sound of many devices as much as the 'scope view simply because they are poorly made from the outset.
It's like dressing a hobo in an Armani suit - you get a hobo in an Armani suit, no more.
But they do their job admirably, I guarantee it. This means that my power supply arrives at the device's transformer already filtered from whatever happens to travel along the grid lines at the moment. With a rating of 20A/230V EACH, let's say I have no immediate problems with overloads, capacities, etc, I am well taken care of.
Since you, to the best of my knowledge, do not use an outside filtering device, obviously your problems are the same as mine, but you have to solve them inside the power amp, preamp, or whatever. I have no doubt at all that your system work very well, but in my case, that would be putting a second system in series with the first.
Very often, this can cause unpredictable and never pleasant effects. In brief, it can kill the sound dead as a doornail. I know - I tried putting both mine and other similar devices in series, and never once got anything but a catastrophe back.
Which is why I will pass on this matter, with the knowledge that I already have the problem well under control. Also using LC filtering.
Dejan,
And?
This does not filter the main sources of noise in the Amplifiers PSU. Do you know what the turnover (-3dB) for a 2mH + 18,000uF filter is?
To give you an idea, at 5A one supply of my style compares to the same sans inductors as follows in Duncanamps PSU-Designer:
Parallel @ 0.5A 120mV Peak-Peak, pure sawtooth << 1mS rise time
Parallel @ 5A 1V Peak-Peak, pure sawtooth << 2mS rise time
Parallel effective DCR = 1.65 ohm and effective 70mS time constant
CLC @ 0.5A - 14mV peak-Peak, slightly distorted sinewave 5mS rise time
CLC @ 5A - 130mV Peak-Peak, slightly distorted sinewave 5mS rise time
CLC effective DCR = 1.62 ohm and 60mS time constant
So, around 18dB lower absolute noise and a much less spikey spectrum and only a little worse for time-constant (how long it takes the PSU voltage to fall to within 30% of the final 5A load following a 0.5A to 5A step)
Your line filters cannot do anything like that, because they are BEFORE the rectifier (where they are of limited use) not AFTER the rectifier (where they belong).
BTW, my frontend supply uses 3,300uF/2 * 6.8mH/3,300uF per rail for around 100mA prior to the shuntreg, from the original 2 * 1,000uF in parallel. This reduces the ripple prior to the regulators from over 0.5V P-P to 8mV P-P...
Sure. I bought a NEC-Tokin line filter for my Amp that is about the size of the head of German stick hand grenade (as opposed to the US style pineapple ones). It has around 30KHz differential cutoff, common mode a little lower.
It is true, I design all my gear not to need external band-aids. Most of the times I like to design the filtering directly into the mains transformer...
However the CLC in the PSU has noting to do with the RF stuff, or external issues, but all with happens inside the Amp.
Your call, but in reality you are merely deceiving yourself into a false security that will create a lower performance device in the end.
Ciao T
And?
This does not filter the main sources of noise in the Amplifiers PSU. Do you know what the turnover (-3dB) for a 2mH + 18,000uF filter is?
To give you an idea, at 5A one supply of my style compares to the same sans inductors as follows in Duncanamps PSU-Designer:
Parallel @ 0.5A 120mV Peak-Peak, pure sawtooth << 1mS rise time
Parallel @ 5A 1V Peak-Peak, pure sawtooth << 2mS rise time
Parallel effective DCR = 1.65 ohm and effective 70mS time constant
CLC @ 0.5A - 14mV peak-Peak, slightly distorted sinewave 5mS rise time
CLC @ 5A - 130mV Peak-Peak, slightly distorted sinewave 5mS rise time
CLC effective DCR = 1.62 ohm and 60mS time constant
So, around 18dB lower absolute noise and a much less spikey spectrum and only a little worse for time-constant (how long it takes the PSU voltage to fall to within 30% of the final 5A load following a 0.5A to 5A step)
Your line filters cannot do anything like that, because they are BEFORE the rectifier (where they are of limited use) not AFTER the rectifier (where they belong).
BTW, my frontend supply uses 3,300uF/2 * 6.8mH/3,300uF per rail for around 100mA prior to the shuntreg, from the original 2 * 1,000uF in parallel. This reduces the ripple prior to the regulators from over 0.5V P-P to 8mV P-P...
Sure. I bought a NEC-Tokin line filter for my Amp that is about the size of the head of German stick hand grenade (as opposed to the US style pineapple ones). It has around 30KHz differential cutoff, common mode a little lower.
It is true, I design all my gear not to need external band-aids. Most of the times I like to design the filtering directly into the mains transformer...
However the CLC in the PSU has noting to do with the RF stuff, or external issues, but all with happens inside the Amp.
Your call, but in reality you are merely deceiving yourself into a false security that will create a lower performance device in the end.
Ciao T
Well, that's an easy one to crack. I can simply make two power supply boards, one with and one without your suggestions built in. Since I already have most of the materials, that shouldn't be too much of a problem.
Then I give it a spin, see what I come up with. I will not dismiss the idea out of hand, I allow for the possibility that I might be wrong.
I just inspected my stock, and I do have Radiohm (France) 42H42 sintered toroidal ring inductors, rated at 1.8 mH, 10A/250V. They are doubles, meaning two in parallel on the same ring, so one will cover both "+" and "-".
It's not 1 mH, but then my caps are not 18,000 uF but 10,000 uF. Kind of works out rather nicely, doesn't it?
So, I stick in one such inductor after the first 10,000 uF cap, then follow it with another 10,000 uF cap, followed by one 4,700 uF cap - though this last one I should perhaps put on the board as near to the power trannies as I can. Or just add one more 4,700 uF cap, as I said, I have a LOT of them.
So, that should work out at 10,000+10,000+4,700+4,700, for a total of 29,400 uF per supply line, rated at around 53V offloaded, or around 51V at full load. Locally, I order my toroids according to the minimum voltage I want from it when it is 100% loaded, and besides, I tend to go out to town when I buy toroidal transformers.
I believe two things should be massive - toroidal transformers and heat sinks. Put in as much as you can possibly fit in.
It's not 1 mH, but then my caps are not 18,000 uF but 10,000 uF. Kind of works out rather nicely, doesn't it?
So, I stick in one such inductor after the first 10,000 uF cap, then follow it with another 10,000 uF cap, followed by one 4,700 uF cap - though this last one I should perhaps put on the board as near to the power trannies as I can. Or just add one more 4,700 uF cap, as I said, I have a LOT of them.
So, that should work out at 10,000+10,000+4,700+4,700, for a total of 29,400 uF per supply line, rated at around 53V offloaded, or around 51V at full load. Locally, I order my toroids according to the minimum voltage I want from it when it is 100% loaded, and besides, I tend to go out to town when I buy toroidal transformers.
I believe two things should be massive - toroidal transformers and heat sinks. Put in as much as you can possibly fit in.
Dejan,
These are utterly useless. The rating is not DC.
You need 1mH @ 10A DC, single. You will not find anything like them at Farnell or Radiospares. Mine where custom made by a German company that specialises in high power inductances.
Nope. Utterly useless.
Not as such disagreeing, except Torroids. They cost more to fix the problems in many cases than they save money.
Try double C-Core or LL Core instead of Torroids if you can. Torroids, especially off the shelf models are the worst possible choice for audio. R-Cores are second worst.
Ciao T
These are utterly useless. The rating is not DC.
You need 1mH @ 10A DC, single. You will not find anything like them at Farnell or Radiospares. Mine where custom made by a German company that specialises in high power inductances.
Nope. Utterly useless.
Not as such disagreeing, except Torroids. They cost more to fix the problems in many cases than they save money.
Try double C-Core or LL Core instead of Torroids if you can. Torroids, especially off the shelf models are the worst possible choice for audio. R-Cores are second worst.
Ciao T
- Status
- Not open for further replies.