Power Supply
Schematic for power supply attached. Note that MOSFET regulation was used for voltage driver and phase spliter stages, not output tubes.
Also my design ensured that the current drain from the two driver stages and phase spliier had a constant current flow hence low valued non electrolic capacitors
Schematic for power supply attached. Note that MOSFET regulation was used for voltage driver and phase spliter stages, not output tubes.
Also my design ensured that the current drain from the two driver stages and phase spliier had a constant current flow hence low valued non electrolic capacitors
Interesting. This sounds kind of like the "hot spot" breakdown of the older epitaxial bipolars. Where the conductance of the hot spots increase with temperature, increasing localized current and runaway occuring in a local spot on the die, leading to failure.
As soulmerchant and EL504 said, damper diodes will delay B+. They are single plate. Find some quality ones.
Or, if the B+ windings are not too high of a voltage, use a 5AR4 or similar indirect heated 2 plate rectifier.
For any and all tube rectifiers, be sure to pay attention to the maximum input capacitor specification of the tube you use. That cap spec is dependent on the applied AC plate voltage, DC current draw, and DCR of the secondary; as well as the reflected DCR of the primary as it transforms upward to a higher resistance that is equal to the turns ratio secondary/((1/2)primary) x DCR of the primary). Or, for a choke input filter, be sure to use a choke that is made for that kind of service, and be sure it is at least as large as the critical inductance, Henrys = 350/(mA DC current) of the amp.
Solid state slow start circuits, and relays too are only as good as the parts quality and the quality of the design.
"You should make things as simple as possible, but no simpler" - Albert Einstein
Or, if the B+ windings are not too high of a voltage, use a 5AR4 or similar indirect heated 2 plate rectifier.
For any and all tube rectifiers, be sure to pay attention to the maximum input capacitor specification of the tube you use. That cap spec is dependent on the applied AC plate voltage, DC current draw, and DCR of the secondary; as well as the reflected DCR of the primary as it transforms upward to a higher resistance that is equal to the turns ratio secondary/((1/2)primary) x DCR of the primary). Or, for a choke input filter, be sure to use a choke that is made for that kind of service, and be sure it is at least as large as the critical inductance, Henrys = 350/(mA DC current) of the amp.
Solid state slow start circuits, and relays too are only as good as the parts quality and the quality of the design.
"You should make things as simple as possible, but no simpler" - Albert Einstein
Last edited:
I never understood why some people put series RC across 1N4007 diodes for an Audio amp in a home stereo application. I do not know how similar the UF 1007 is.
We are not in a Ham Radio shack with an RF field of 100s of volts/meter impinging on the AC mains cords. The same thing can apply in government communication installations too. In RF applications and some others, "Grounds are commonly misunderstood" - Me
If there is a transient from power down of the audio amp, or mains transients that exceeds the 1N1007 diodes PRV, then get better rated diodes. One particularly tough situation I had was fixed by replacing those diodes with 8A 1500V HEXFRED diodes. The large area chip handled any transients I could throw at it.
I use power strips that have transient protection for line transients. The power strips also have a power switch which saves the power switches on amps, signal sources, etc. Having power amps, signal sources, preamps, etc all on the same power strip helps to reduce the chance of ground loops between all those items. If the power strip goes bad, I replace it.
We are not in a Ham Radio shack with an RF field of 100s of volts/meter impinging on the AC mains cords. The same thing can apply in government communication installations too. In RF applications and some others, "Grounds are commonly misunderstood" - Me
If there is a transient from power down of the audio amp, or mains transients that exceeds the 1N1007 diodes PRV, then get better rated diodes. One particularly tough situation I had was fixed by replacing those diodes with 8A 1500V HEXFRED diodes. The large area chip handled any transients I could throw at it.
I use power strips that have transient protection for line transients. The power strips also have a power switch which saves the power switches on amps, signal sources, etc. Having power amps, signal sources, preamps, etc all on the same power strip helps to reduce the chance of ground loops between all those items. If the power strip goes bad, I replace it.
Last edited:
I don't know about stereo's, but I put a capacitor across 1N400X diodes in low voltage power supplies as a matter of routine, because it eliminates the diode as a source of noise that some radio receivers will easily respond to ( direct conversion receivers are notorious for this ).
Whether better diodes would make that unnecessary, I don't know - I have better diodes, I just never seem to use them ...
Delayed B+ is not on my radar screen as something to worry about.
While one diode in a bridge that is connected to secondary lead #1 is turning Off, Another diode in the bridge that is connected to secondary lead #1 is turning On. I suppose capacitances, plus the lead inductance of the leads of the diodes and leads to the transformer primary are ringing. But then does not the series lead inductance of the capacitor and its capacitance also ring.
I figure use more expensive diodes, in able to use less parts (no RC across the diode).
Just asking.
My hi fi amp designs with solid state diodes have less than 100uV hum and noise at the output (and I do not use negative feedback).
I wish my amplifiers were the worst noise source interfering with all the AM radios in my house.
You want interference, just put a sub microvolt LF receiver in a fully shielded screen room with finger-stock door ground all around. The only penetration is the AC power coming through the shield wall, and it employs a very expensive filter, and complete electrical ground unity with the shield wall. Do this in a shipyard full of welders . . . good luck!
I figure use more expensive diodes, in able to use less parts (no RC across the diode).
Just asking.
My hi fi amp designs with solid state diodes have less than 100uV hum and noise at the output (and I do not use negative feedback).
I wish my amplifiers were the worst noise source interfering with all the AM radios in my house.
You want interference, just put a sub microvolt LF receiver in a fully shielded screen room with finger-stock door ground all around. The only penetration is the AC power coming through the shield wall, and it employs a very expensive filter, and complete electrical ground unity with the shield wall. Do this in a shipyard full of welders . . . good luck!
There is a delay period when all diodes in a bridge are off (for typical capacitor input filter power supplies) - which allows the secondary winding to float in voltage relative to ground, being only restrained by the need to charge stray capacitances in order to change voltage with respect to ground.
It is the leakage inductance in the secondary winding that is the source of energy that can cause ringing in the power supply. The capacitance through/across a diode allows a path for ringing current to flow out in to the B+ filter and circuitry. Poor layout and wiring is typically the reason that some amps may be prone to that noise/hum being heard - well made amps don't exhibit it.
The chance of noise getting in to the B+ circuitry is suppressed if the diode is faster, and the diode current rating is smaller (less junction capacitance/charge), and if the peak B+ charging current through the diodes is lower (less first filter capacitance), and if valve diodes are used instead of ss diodes (higher series resistance).
It is the leakage inductance in the secondary winding that is the source of energy that can cause ringing in the power supply. The capacitance through/across a diode allows a path for ringing current to flow out in to the B+ filter and circuitry. Poor layout and wiring is typically the reason that some amps may be prone to that noise/hum being heard - well made amps don't exhibit it.
The chance of noise getting in to the B+ circuitry is suppressed if the diode is faster, and the diode current rating is smaller (less junction capacitance/charge), and if the peak B+ charging current through the diodes is lower (less first filter capacitance), and if valve diodes are used instead of ss diodes (higher series resistance).
Not in our usual cap-input rectifier. The diodes conduct maybe 10% of the time.
PRR,
So true.
I forgot about the effects of using cap input filters, what was I thinking. A great generator of large current transients. Voltage drops in ground leads, and magnetic fields too. And the wave-shapes are not friendly either.
I use choke input filters whenever possible. Then I use CRCRC. And I use solid state diodes, not tube rectifiers. Solid state diodes versus tube rectifiers makes up for Part of the difference of B+ voltage of a cap input versus choke input supply (1.414 x Vrms versus 0.9 x the Vrms).
I return the center tap of the HV secondary to the negative of the first filter cap (which has its + side connected after the choke). Then I connect the 1st filter cap negative to the 2nd filter cap negative. That attention to detail means the first 2 major hum loops are self contained from the rest of the amp wiring. Of course lead length, aperture, and where all this first circuitry is located on the amp becomes important. But the nature of the transients are smoother than that of cap input filters. All the + connections from cap to cap are resistors.
At that point, I connect the negative of the 2nd filter cap to the rest of the amp (the central ground point). And that central ground point is connected to the chassis there. the 3rd filter cap connects there too.
The RCA input jacks are insulated from the chassis. The jack common connects to the bottom of the input stage's cathode self bias R and bypass C. Then from there, I connect that to the central ground point. Grid stopper(s) on the grid(s) helps to maintain the stability of this stage, but short leads are important too.
I use a 3-wire power cord from a 3 wire power outlet to the amplifier. The ground of it connects to the amp chassis. All the signal sources: phono preamp, turntable, tuner, and CD player have 2 wire power cords. There is a ground wire from the turntable to the phono preamp. All other system grounds are intrinsic in the shielded phono cables.
There are other factors, but they may differ from amp to amp, depending on the circuit and the real estate of the chassis, and the parts used, particularly the 'magnetics'.
Most of my amps do not use negative feedback, the UL connection is the exception (Well, I also consider a screen to plate triode wired mode as having negative feedback too).
Less than 100uV to the amp speaker output is the goal, and is most often met.
It works for me, but may not for other amplifier's layouts and topologies.
So true.
I forgot about the effects of using cap input filters, what was I thinking. A great generator of large current transients. Voltage drops in ground leads, and magnetic fields too. And the wave-shapes are not friendly either.
I use choke input filters whenever possible. Then I use CRCRC. And I use solid state diodes, not tube rectifiers. Solid state diodes versus tube rectifiers makes up for Part of the difference of B+ voltage of a cap input versus choke input supply (1.414 x Vrms versus 0.9 x the Vrms).
I return the center tap of the HV secondary to the negative of the first filter cap (which has its + side connected after the choke). Then I connect the 1st filter cap negative to the 2nd filter cap negative. That attention to detail means the first 2 major hum loops are self contained from the rest of the amp wiring. Of course lead length, aperture, and where all this first circuitry is located on the amp becomes important. But the nature of the transients are smoother than that of cap input filters. All the + connections from cap to cap are resistors.
At that point, I connect the negative of the 2nd filter cap to the rest of the amp (the central ground point). And that central ground point is connected to the chassis there. the 3rd filter cap connects there too.
The RCA input jacks are insulated from the chassis. The jack common connects to the bottom of the input stage's cathode self bias R and bypass C. Then from there, I connect that to the central ground point. Grid stopper(s) on the grid(s) helps to maintain the stability of this stage, but short leads are important too.
I use a 3-wire power cord from a 3 wire power outlet to the amplifier. The ground of it connects to the amp chassis. All the signal sources: phono preamp, turntable, tuner, and CD player have 2 wire power cords. There is a ground wire from the turntable to the phono preamp. All other system grounds are intrinsic in the shielded phono cables.
There are other factors, but they may differ from amp to amp, depending on the circuit and the real estate of the chassis, and the parts used, particularly the 'magnetics'.
Most of my amps do not use negative feedback, the UL connection is the exception (Well, I also consider a screen to plate triode wired mode as having negative feedback too).
Less than 100uV to the amp speaker output is the goal, and is most often met.
It works for me, but may not for other amplifier's layouts and topologies.
In my place, the ENTIRE AM dial is just noise and 60Hz. I literally can not receive ANY am stations in here at all. Using a loop antenna or a ferrite bar antenna on older equipment. Same marde.
... and the October issue of AudioXpress which will be published early September has an article from me: "A High-Voltage Delay for Tube Amplifiers".
I am introducing a small half-kit for it which will be available through the diyaudio store.
Just sayin' ;-)
Jan
I listen to a local station from San Francisco on my R390A/URR. Short waves are also polluted. Only rare religious preachers on them.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.