And your amp boards have only minimal capacitance? I have an idea how this will sound and it is not a sound i strive for.
Ironically, some of the more successful commercial Ncore implementations use linear supplies and even batteries. And one would expect switchers to be perfect for class D...
Ironically, some of the more successful commercial Ncore implementations use linear supplies and even batteries. And one would expect switchers to be perfect for class D...
the Hypex is very quiet,
stabilized,
does not radiate,
is small,
does not vibrate and sounds better than a float.
NO, no cap downstream of the switching on the lines, but here you have to decide on what to choose:
- the performance of a power supply, see output impedances
or
- eliminate ALMOST all electrolytics, which are then on the signal, from the lines: do you know these differences?
stabilized,
does not radiate,
is small,
does not vibrate and sounds better than a float.
NO, no cap downstream of the switching on the lines, but here you have to decide on what to choose:
- the performance of a power supply, see output impedances
or
- eliminate ALMOST all electrolytics, which are then on the signal, from the lines: do you know these differences?
I had two power amplifiers with stabilized PSUs, using the transformers I'm using now in the WHA-217. 50+50Vac rms in, 50+50Vdc out. Those were simple and good stabilizer: a discrete Darlington and a Zener, no feedback; it only regulated the maximum voltage, that couldn't go lower because the transformer was oversized a lot. 4 PSUs in one amplifier. Very good sound. Very good! I know, people say that stabilized PSU gives a bad sound. Perhaps PSU with feedback do, I don't know, I never used one.
I would like to try an inductive PSU but it is heavy, bulky, expensive and you must find a good craftsman able to build the inductor that must not hum. You need at least 4 inductors. Or, perhaps, you can design the PSU trying to eliminate the saturation with the opposite fluxes flowing in the inductor. But still, you need two inductors, perhaps without any gap, perhaps a lot easier to build. I have not the skill to design such a PSU.
Switching PSU does have a feedback loop, cannot do less. But I think they are worth a try.
The bad of all these ideas is that my house is becoming again a HiFi shop, I'm spending a lot of money, and my wife is not that happy.
However, dear Domenico, would you like to explain to us your idea about eliminating almost all electrolytic capacitors from the (lines) signal path opponed to the performance/impedance? A short explanation, I don't want to force you, if you have any useful link, would you share it? I have to confess, I know nothing about switching PSUs.
I would like to try an inductive PSU but it is heavy, bulky, expensive and you must find a good craftsman able to build the inductor that must not hum. You need at least 4 inductors. Or, perhaps, you can design the PSU trying to eliminate the saturation with the opposite fluxes flowing in the inductor. But still, you need two inductors, perhaps without any gap, perhaps a lot easier to build. I have not the skill to design such a PSU.
Switching PSU does have a feedback loop, cannot do less. But I think they are worth a try.
The bad of all these ideas is that my house is becoming again a HiFi shop, I'm spending a lot of money, and my wife is not that happy.
However, dear Domenico, would you like to explain to us your idea about eliminating almost all electrolytic capacitors from the (lines) signal path opponed to the performance/impedance? A short explanation, I don't want to force you, if you have any useful link, would you share it? I have to confess, I know nothing about switching PSUs.
I'm talking about my experience, of course.
I have a power supply and I am trying to technically make it work at its best, and how?
I sow electrolytic caps bypassed by plastic along the lines and I can get, from a stabilized power supply, output impedances of 3 MILLIOHM ........ not bad! , almost a battery.
Then I look at the inside of a final whose design school teaches a lot and I know that, even with a float, it does not use any electrolytic on the lines.
I try to eliminate the electrolytics from the lines, the power supply remains stable and without problems, I only see that the impedance reaches 50 MILLIOHM.
oh well, let's hear.
and if you listen to the two identical stabilized ones, one with and the other without cap, ( the output cap of the stabilized power supply remains which manages to go down to 10 micro, 10 micro plastic ), there is no hesitation about which one to prefer
I have a power supply and I am trying to technically make it work at its best, and how?
I sow electrolytic caps bypassed by plastic along the lines and I can get, from a stabilized power supply, output impedances of 3 MILLIOHM ........ not bad! , almost a battery.
Then I look at the inside of a final whose design school teaches a lot and I know that, even with a float, it does not use any electrolytic on the lines.
I try to eliminate the electrolytics from the lines, the power supply remains stable and without problems, I only see that the impedance reaches 50 MILLIOHM.
oh well, let's hear.
and if you listen to the two identical stabilized ones, one with and the other without cap, ( the output cap of the stabilized power supply remains which manages to go down to 10 micro, 10 micro plastic ), there is no hesitation about which one to prefer
Downgraded LC Audio "The End"
Lars Clausen from New Class D tells me that Dartzeel looks very much like a downgraded version of his LC Audio "The End" made in different versions.
Lars Clausen from New Class D tells me that Dartzeel looks very much like a downgraded version of his LC Audio "The End" made in different versions.
Copenhagen Hificlub
Some members of Copen Hagen Hifi Club tells me they thought Dartzeels in some aspects sound better then our old Holton 500 with vertical Mosfets. 150 versus 400 W.
Some members of Copen Hagen Hifi Club tells me they thought Dartzeels in some aspects sound better then our old Holton 500 with vertical Mosfets. 150 versus 400 W.
Last edited:
These are the results of measuring hFE with a 20-year-old Metex digital multimeter
MJL 3281 transistors have values of 49-54 mostly 50
MJL 1302 transistors have values of 89-91 mostly 90
MJE 15032 transistors have values of 45-48 mostly 46
MJE 15033 transistors have values of 149-155 mostly 150
2N 5401 transistors have values of 127-136 mostly 133
2N 5551 transistors have values of 101-112 mostly 111
Looks like I have a bunch of useless parts and I won't be able to pair the transistors.
I'm expecting a transistor tester to arrive in 2 weeks
Does anyone have transistors with similar values?
MJL 3281 transistors have values of 49-54 mostly 50
MJL 1302 transistors have values of 89-91 mostly 90
MJE 15032 transistors have values of 45-48 mostly 46
MJE 15033 transistors have values of 149-155 mostly 150
2N 5401 transistors have values of 127-136 mostly 133
2N 5551 transistors have values of 101-112 mostly 111
Looks like I have a bunch of useless parts and I won't be able to pair the transistors.
I'm expecting a transistor tester to arrive in 2 weeks
Does anyone have transistors with similar values?
Last edited:
Since I will not work with any PSU having feedback (despite this, perhaps I will try a switching PSU), for my next project I will try to use a capacitance multiplier whose capacitor will be a polypropylene one. I have some ideas but all parts are in the lab in Rome.
respectfully, but this story of demonizing feedback in an SS power amp isn't easy to understand, much less in a PSU.
The SS work with feedback either local or ring, otherwise they do not stay together and among the best endings in SS there are high feedback ones.
Then it depends if it is a question of philosophy because, both technically and how sound the feedback it takes.
Do you want an open ring? Well, but the sound will be accordingly and don't hope in the usual joke that the 108 sounds like a 300B, just NO!
In my opinion, it's very bad to introduce an error then correct it once happened. Simply avoid the error! According to my oldtimer knowledge, you cannot design an amp with transistors using only the local feedback, because you cannot avoid to introduce the error and then fix it with a feedback loop.
Even if the LC Audio "The End Millennium" actually has no feedback loop, I watched at the schema but renounced to understand where the magic is hidden.
However, when designing a linear stabilized PSU, you can avoid the error so you do not need any feedback.
I abandoned the feedback loop in 1983 and never had to repent.
I opened and closed a short parenthesis playing with LJM L20s, but just because I had a hard time with fake transistors while working on the WWHA-217, and needed a fast solution in order to hear music in the living room. Now the L20 is used in my lab.
In 2004 I launched an idea about designing a high power amplifier with current dumping and feedforward correction using a 2A3 or 300B in single-ended. A 200W amp with the 300B sound. Nobody did follow.
You are right, it is philosophy, we all have our own one.
Even if the LC Audio "The End Millennium" actually has no feedback loop, I watched at the schema but renounced to understand where the magic is hidden.
However, when designing a linear stabilized PSU, you can avoid the error so you do not need any feedback.
I abandoned the feedback loop in 1983 and never had to repent.
I opened and closed a short parenthesis playing with LJM L20s, but just because I had a hard time with fake transistors while working on the WWHA-217, and needed a fast solution in order to hear music in the living room. Now the L20 is used in my lab.
In 2004 I launched an idea about designing a high power amplifier with current dumping and feedforward correction using a 2A3 or 300B in single-ended. A 200W amp with the 300B sound. Nobody did follow.
You are right, it is philosophy, we all have our own one.
Yes, they are a lot different NPN/PNP, being usually PNPs a lot higher than NPNs. Yours have values a lot lower than mine. My MJLs P/N are all almost the same.
However, even with that big differences, I used them the same. The midpoint setting (servo or not), the short feedback loop, and the heat will someway compensate the differences.
A recommendation? Try to buy some other MJE15032s but not on eBay!
I tend to buy these transistors in batches of 10-20. A batch bought a few months apart behaves differently and you get more chances for a good match between batches. You also have to match at realistic currents and not leave that choice to the tester.
Some fresh listening impressions on power amp regulation from another thread:
https://www.diyaudio.com/forums/pass-labs/271543-cubie2-6.html#post6332755
"So, just for a starting point, and to make things simpler, I decided to replace the MOSFET circuit with a power resistor of 2.2 Ohms, just on one channel, and make a direct A-B comparison. Well, I must say that the difference was huge! It was one of the rare occasions where I felt that a veil was removed from the scene. "
🙂
https://www.diyaudio.com/forums/pass-labs/271543-cubie2-6.html#post6332755
"So, just for a starting point, and to make things simpler, I decided to replace the MOSFET circuit with a power resistor of 2.2 Ohms, just on one channel, and make a direct A-B comparison. Well, I must say that the difference was huge! It was one of the rare occasions where I felt that a veil was removed from the scene. "
🙂
When I’m done with work I’ll review it in the evening, it could certainly help. Thanks, I'll post in the evening 🙂
these are statements that, without knowing what one is talking about, that is, without seeing the patterns, appear quite superficial