I am still waiting for a definitive answer to this question---does anyone have one?
Let's say, for a "typical" 100 watt MONO amplifier---delivers 100 watts into 8Ω at ≤0.1% THD. 28.3 volts RMS, 3.5 amps. 40 volt peak; 200 watt clipping point. Let's say this has a power supply capable of delivering +/- 45 volts @ 5 amps. We have a speaker that has 90 db sensitivity (@1 watt/1 meter). If we allow for 14db of PEAK headroom, we can listen at 99 db SPL without clipping---fairly loud, but not overpowering. At this point, our amplifier is delivering 8 watts rms, or 8 volts into 8Ω. If we want the ripple to be 80 db below this (typically considered "inaudible"), this would require ripple to be ≤ 0.8 mVolts. Our 5 ampere supply, with 10,000uF filtering, will deliver about 600 mVolts of ripple. If our "typical" amplifier is capable of a PSRR of 60 db, then the ripple on the output would be 0.6mV---below the level of audibility.
Last edited:
I am very sensitive to femto-volt disturbances in the 1KHz to 2KHz band. Completely destroys my enjoyment of MotorHead so now you all have to design stuff that meets my criteria... and absolutely no Blue LEDs either for indication or as part of a voltage reference.
If God had meant us to use Blue LEDs he would not have made them so hard to make in the first place. Scourge of the Flying Spaghetti Monster if you ask me.
What is "the definitive audio power amplifier"?
In the stone-age, raw 120Hz ripple power, amp designers started with plans that promised good PSRR.
Car-sound, not so much (no 60Hz; and 50 pounds lead sucking alternator ripple).
AmpCamp *starts* with a regulated (switch-mode) power source. This freed NP to consider other plans with not-so-much PSRR. This one came out pretty good. MB's sequence-study of mods show you can find much worse, but 72dB is about what you get with NP's values. Which is really VERY good for such a simple amplifier. Comparable to an older voltage regulator.
So you gotta ask: is "linear power!!" really *better* for an amplifier *designed* for regulated power? (And promoted as a low-cost high-fun project?) Are we lipsticking a pig? A very pretty fair-winning piglet, but is lipstick an improvement?
The last amps I built I used a rule of thumb of output ripple 90dB below rated output and am happy that I did (one SS and one valve for bi-amping, appropriately sized for the speaker so they work together). This is after working its way through the amp so the ripple at the supply could be higher. I was careful to make simple (sinusoidal) ripple the only measurable (with my scope) noise from the supplies.
Last edited:
Had a few spare minutes to run a simulation on an amp I'm designing... with 1Vrms/120Hz (0dB) on each supply rail, 120Hz on the output was -127dB. I don't know how that ranks against other designs, but it seems like that should be enough that I don't have to go to great lengths to filter the supply.
If you really are concerned about ripple on the supply rails it might be better to go with a capacitance multiplier than a choke.
If you really are concerned about ripple on the supply rails it might be better to go with a capacitance multiplier than a choke.
Hello,
I just finished a single rail choke input power supply for the amp i use in my home cinema 2.1 set. Before it used to big 12 volt sealed lead batteries in series.
Now it uses an old Philips transformer bought in a surplus store. Says made in Holland and it could well be made 40 years ago probably made to measure but just the thing i needed.
After the four SBYV28-100 diodes i use a 128 mH choke on '' each side '' bought from a Belgian member here. Then 2 4700µF Kemet peh169 caps then a Lundahl LL2733 700mH 1A with two coils in parallel so 175mH 2A and then a 22000µF Kemet before going inside the amp where there is another cap on the circuit.
The photo is just a test situation to see if the diodes and chokes can handle the current. The diodes used in the end are rather small 3,5A continious current but i was told because of the choke input there are pulsing charges and everything stays cool.
The batteries had to be replaced because they didnt keep their charge as long as before. Because they are rather expensive a\nd need to be changed every 4 years i decided to go choke input with all the parts together less expensive than a set of batteries.
I was expecting going down in quality but it is at least equal quality. Maybe because the batteries were getting older. I remember starting with a choke input with a smaller transformer and a choke with less mH so needed a bigger bleeder and switching to batteries was a major improvement. Now with 4 times more mH for the input choke everything works perfect.
greetings, Eduard
I just finished a single rail choke input power supply for the amp i use in my home cinema 2.1 set. Before it used to big 12 volt sealed lead batteries in series.
Now it uses an old Philips transformer bought in a surplus store. Says made in Holland and it could well be made 40 years ago probably made to measure but just the thing i needed.
After the four SBYV28-100 diodes i use a 128 mH choke on '' each side '' bought from a Belgian member here. Then 2 4700µF Kemet peh169 caps then a Lundahl LL2733 700mH 1A with two coils in parallel so 175mH 2A and then a 22000µF Kemet before going inside the amp where there is another cap on the circuit.
The photo is just a test situation to see if the diodes and chokes can handle the current. The diodes used in the end are rather small 3,5A continious current but i was told because of the choke input there are pulsing charges and everything stays cool.
The batteries had to be replaced because they didnt keep their charge as long as before. Because they are rather expensive a\nd need to be changed every 4 years i decided to go choke input with all the parts together less expensive than a set of batteries.
I was expecting going down in quality but it is at least equal quality. Maybe because the batteries were getting older. I remember starting with a choke input with a smaller transformer and a choke with less mH so needed a bigger bleeder and switching to batteries was a major improvement. Now with 4 times more mH for the input choke everything works perfect.
greetings, Eduard
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.