1uf input cap, 5H choke, and 100uF 2nd cap . . .
That is similar to a pi filter for a vacuum tube transmitter output.
But the resonant frequency of a series 1uF, 5H, and 100uF is very close to the resonant frequency of
1uF and 5H.
With a small input capacitance, If the input filter still acts like a true choke input filter, then the resonance peak will be attenuated by the turned on diodes.
But the moment that the filter starts to act like a cap input filter, the resonance will not be as well attenuated.
Choke input filter = constant current
Cap input filter = transient current peaks.
Reminds me of a class C transmitter output stage.
Just my opinions.
That is similar to a pi filter for a vacuum tube transmitter output.
But the resonant frequency of a series 1uF, 5H, and 100uF is very close to the resonant frequency of
1uF and 5H.
With a small input capacitance, If the input filter still acts like a true choke input filter, then the resonance peak will be attenuated by the turned on diodes.
But the moment that the filter starts to act like a cap input filter, the resonance will not be as well attenuated.
Choke input filter = constant current
Cap input filter = transient current peaks.
Reminds me of a class C transmitter output stage.
Just my opinions.
One way of looking at it would be the that the power transformer secondary and all its inherent and reflected parasitics (stray capacitances, both to noisy primary and to chassis (often signal ground) yada, yada, and the noise that they generate, may be the most relevant or at least maybe important of issues in modern constructions. This is thanks to the publications of some whacky frickin guy named Jan D- something or other. See Linear Audio numbers 5 and 10, off the top of my head. I'll fix it tomorrow if the volume numbers are wrong. Just get them all, while avaiiable. Don't be that sad person who wishes he had the paper copies of Sound Practices from the 1990s - be the happy person with his own paper copies of Linear Audio.
All good fortune,
Chris
All good fortune,
Chris
The B+ level with choke input is 0.9x full load transfomer winding voltage - tube voltage drop
The B+ level with cap input is typical 1.3x full load transformer winding voltage - tube voltage drop.
A typical tube voltage drop is a small percentage of the total voltage. 1.4x voltage is still easily done even with tubes.
An error in your text-book.
With a capacitor input filter the RMS current is very high, resulting in larger transformer voltage drop, compared to resistive load.
The conduction angle is only 30 - 60 degrees.
Last edited:
No, the PEAK current is very high due to conduction angle, not RMS... Or is Hammond wrong?
Also, we were talking choke vs capacitor. Here's Hammond's resistive loading.
Here's the rest of it for reference: https://www.hammfg.com/electronics/transformers/rectifier
Exert:
Also, we were talking choke vs capacitor. Here's Hammond's resistive loading.
Here's the rest of it for reference: https://www.hammfg.com/electronics/transformers/rectifier
Exert:
Transformer Voltage:
A transformer's required secondary A.C. voltage varies greatly with the type of rectifier chosen and filter arrangement. Use the formulas below as a guide based on the D.C. voltage you require and the rectifier/filter chosen. All A.C. voltage references are R.M.S. Don't forget to take into account losses (not included in this guide), especially diode voltage drop. Leave an adequate safety margin for D.C. regulator voltage requirements and minimum operating line voltage.Transformer Current Ratings:
A transformer's A.C. current rating needs to be recalculated from the D.C. load current. The required current varies with type of rectifier chosen and filter type. Use the formulas below as a guide, shown for common D.C. supplies. Included in the formulas higher peak to peak capacitor charging current in the filter.
Last edited:
The discussion has shifted from input impedance to an easy way to trim a choke input PS upwards using a sub 1uf capacitor. So I thought I should post the two variations.
The ripple on both of these is very low. This is for the Salas 6V6 line amp which wants 340V @ approx 45ma. All capacitors are Kemet DC link (no electrolytics), all LC (no resistors).
Tradition raw power supply (for what I want to build now)...
Spot on output voltage 345V @50ma:
Future raw power supply to compensate for the insertion of a final shunt regulator (for future enhancement by adding a shunt regulator circuit TBD)...
Easy-peasy, cheap and cheerful, way to get 25 volts? 370V @50ma (for addition of a regulator by simply adding a small 0.47 uf capacitor while retaining mostly choke-input behavior?): But will it oscillate or be unstable or resonate? Does PSUD2 have a way to display if oscillations are occurring?
T1 is a 100VA Antek dual secondary toroid wired as CT 230-0-230 the 80 ohms is what I actually measured across 40-0-40.
The ripple on both of these is very low. This is for the Salas 6V6 line amp which wants 340V @ approx 45ma. All capacitors are Kemet DC link (no electrolytics), all LC (no resistors).
Tradition raw power supply (for what I want to build now)...
Spot on output voltage 345V @50ma:
Future raw power supply to compensate for the insertion of a final shunt regulator (for future enhancement by adding a shunt regulator circuit TBD)...
Easy-peasy, cheap and cheerful, way to get 25 volts? 370V @50ma (for addition of a regulator by simply adding a small 0.47 uf capacitor while retaining mostly choke-input behavior?): But will it oscillate or be unstable or resonate? Does PSUD2 have a way to display if oscillations are occurring?
T1 is a 100VA Antek dual secondary toroid wired as CT 230-0-230 the 80 ohms is what I actually measured across 40-0-40.
Last edited:
Now start the sim at 10 seconds for 20ms so you can see the ripple and RMS won't include the initial charging.
EDIT: Just a thought: Have you considered this cheap small simple replacement for those chokes?
I use this for my phono and line stages... adjust R4 to get your desired voltage. R3 = 5 watts
EDIT: Just a thought: Have you considered this cheap small simple replacement for those chokes?
I use this for my phono and line stages... adjust R4 to get your desired voltage. R3 = 5 watts
Last edited:
Ah thanks,. that's better to just chunk off 20ms of data, I'm still learning how best to leverage PSUD. The ripple is incredibly low swinging between 371.526491 and 371.526492. If there were oscillations on the peaks or instabilities, would PSUD2 be showing those in the final waveform? As expected the waveform at C1 is almost a sine wave, very low charging pulses, (it throws the PDUD scale off too much to show that with the ripple wave). Not seeing any oscillations on C1.
I did consider a solid state choke but wanted this to be a more vintage project point-to-point, choke-input, etc. The "modern" enhancements for my "phase one" of this project are limited (intentionally) to the following, (minimum silicon initially):
Modern stuff (that one would not have found in 1960):
1) Use DC links instead of electrolytics (so they last pretty close to forever, some guy after I'm dead won't be replacing electros).
2) Use DC solid state voltage regulator for the 6V6 filaments (with a polarity reversing switch on the back to average out the wear, flip it every few months).
3) Use two series SS diodes in front of the 5U4 as a protection should the 5U4 fail and start passing AC instead of DC (belt and suspenders).
3) In the future (after listening) maybe add a solid state B+ shunt regulator (TBD).
4) Use Broskie's "house ground" to separate the safety ground from the audio ground (back to back diodes, low voltage caps, a resistor, tiny PCB board)
5) Thats about it for the "modern" variations I wanted in this project.
Vintage stuff:
1) Use point to point wiring
2) Use choke-input PS
3) Tube rectifier
4) etc.
Last edited:
I have to admit the simple active solution Kobabmx posted above might be appropriate for a line amp.
It is sort of nice to have a choke input filter on something like a class A amplifier. You do not waste as much power and have somewhat nicer power factor compared to capacitor input.
The other catch is capacitor life is likely improved when you avoid large current spikes on the peaks.
I say this knowing I have a collection of boat anchor transformers and chokes to play with.
It is sort of nice to have a choke input filter on something like a class A amplifier. You do not waste as much power and have somewhat nicer power factor compared to capacitor input.
The other catch is capacitor life is likely improved when you avoid large current spikes on the peaks.
I say this knowing I have a collection of boat anchor transformers and chokes to play with.
Also, try increasing 10S to 30S because the system is still charging by the look of it. Still, that's ridiculously low ripple 🙂 Much less than 1µV by the look of it. It's effectively an electric battery 🙂
daqvin_carter
I t0o have a boatload of chokes, I need to use them up somehow. 🙂kodabmx
Good idea, it seems to be fully done charging after 30 seconds In practice the load will be less than 50ma probably 42-46 so ripple will decrease more. My initial build wont have the 0.47 up front (jury is still out on that method's stability). I'm thinking if the 0.47 is used up front, maybe it should be a safety capacitor so it fails "open"?
Last edited:
I just try to build a different large amplifier every few years and rotate chokes and transformers so they do not feel neglected.
That started about 1991 or so.
That started about 1991 or so.
I default to X2 MKP safety caps anyway, but an electro can fail short, too. If you're worried about it, use a fuse.
Good idea. The downstream failure mode of a capacitor going open is that everything still works but as a choke-input rather than a short across the 5U4. The shunt regulator (if used) will just drop out probably. Salas uses a shunt regulator but I'm not sure yet, so I won't use the first capacitor method of upping the voltage anyway. Also I figure the music is going through that last 150uf MKP, even though it is already much less ESR than an electro, I will bypass it with an additional small MKP like a Solen. PSUD doesn't let you put two capacitors in a row.
Last edited:
It does it like I do. 1 + 0.1 = 1.1 🙂 Bypassing electros for audio is a waste of a capacitor IMHO. Radio OTOH...
PSUD2 can indicate transient step load response by using the 'stepped load' option in the load component - eg. go from 50mA to 60mA at 5.1 sec, and delay the plot to start at 5 sec. The filter capacitor current waveform is typically the most sensitive way to indicate resonant behaviour subsequent to the step. For your 'over the top' LCLCLC filter, there is no resonance characteristic of concern imho. Keep in mind that a step load change is a worst-case form of transient, and that I'm guessing you haven't measured cap ESR.