MJL21193 said:
That's nothing compared to what it has to deal with in operation under load. With music that supply will be jumping all over the place - magnitide shifts, changing ripple amplitudes, subharmonics of low frequencies when driven to clipping. It's a lot harder for the amp to deal with all that trash than a simple periodic waveform. The effects of that will usually be masked by whatever music is playing. But on an adjacent channel on the same supply - you'll "hear" how good/bad your PSRR really is. If the other channel stays dead quiet, then you've got a good design and layout. With most amps, it won't stay dead quiet.
wg_ski said:
Ok, now I get it. Yes, I feel embarrassed - slightly. I'll recover .
I did as wg-ski instructed:
I connected another amp to the same supply (OS' Frugal2) and drove it with my signal generator. It was connected to the 8 ohm dummy load. I drove this amp to clipping at different frequencies.
The amp being tested (Abomination) was connected to the scope and a fullrange 8 ohm speaker. Even though the scope did show some ripple, especially at higher frequencies, there wasn't enough to be audible through that speaker. Not depending on my hearing, I asked my 18 year old son to have a listen - nothing. To be sure I switched to a square wave and then and only then could he hear a very faint buzz from the speaker.
Oh well, it must be below the noise floor, and apparently I have a good design AND layout.
KLe said:
Hi KLe,
The sim may show some DC on the output (it does: -3.5V) but that can't be relied on. Real measurements are needed and to justify extra DC offset, it would need to "sound" better, so some critical listening would be in order.
Not my ears though - I strained them listening for sound of bad PSRR (or was that the sound of one hand clapping?)
MJL21193 said:
Lets see... that both rails together or one at a time? Let's try one - 72.6VDC rails into 8 ohm load ....equals ~9 amps...or 654 watts...getting quite warm here!
Oh, the 4 ohm, I forgot. that's only 1318 watts into my trusty dummy load.
No problem.
I think Jacco got it right.
I suggested driving the dummy load for the amplifier driving the floating ground to 10V peak (a pretty brutal PSRR test). That's 12.5W into 4 ohms.
Thanks MJL,
Yes, DC offset adjustment would be required (-3.5V is not good
).
I was wondering if it might reduce the 3rd order harmonic distortion artefacts. If this was the case it would sound better, providing the PSRR, or the PSU, or both, are good enough to allow it to be heard. Your designs are very good, so ...
I suggested it, because those who have done this, have always said that the sound improves.
Yes, DC offset adjustment would be required (-3.5V is not good
). I was wondering if it might reduce the 3rd order harmonic distortion artefacts. If this was the case it would sound better, providing the PSRR, or the PSU, or both, are good enough to allow it to be heard. Your designs are very good, so ...
I suggested it, because those who have done this, have always said that the sound improves.
G.Kleinschmidt said:
No doubt. I, unlike you and jacco, am here to learn. Your main focus lately seems to belittle those that know less than you and display your multi-million part count 40 watt and less amps for the fawning attention of those few that are left here that you haven't driven off yet. The time is coming though...
What's that old saying? It's lonely at the top?
G.Kleinschmidt said:
And wg_ski proposed a more practical method. I used his to a satisfactory conclusion.
MJL21193 said:
To be sure I switched to a square wave and then and only then could he hear a very faint buzz from the speaker.
Oh well, it must be below the noise floor, and apparently I have a good design AND layout.
It's obviously go decent rejection at low/mid frequency, falling at 6dB per octave, like theory would suggest. You get the harmonics coming through with square wave.
Now let's take it a step further... try it with a crappy power supply - like the one you'll find in a $200 HT receiver. A 300 VA trafo and a pair of 4700 uF's. Give the amp something to reject
. Hit the other channel 6-10dB into clip with a 2 ohm load using Metallica's "Unforgiven". See if you get any offset voltage shifts in your amp every time the kick drum (?) hits and the causing the supply to drop (none too gently) from +/-70 to +/-45 on every kick. A lot of cheap PA amps don't fare so well when you do this.
KLe said:
I was wondering if it might reduce the 3rd order harmonic distortion artefacts. If this was the case it would sound better, providing the PSRR, or the PSU, or both, are good enough to allow it to be heard. Your designs are very good, so ...
I suggested it, because those who have done this, have always said that the sound improves.
Hi Kle,
I will look at it more closely. The sim suggests a reduction in THD, though I don't know if it specific to the third.
I think (think) the large offset is the result of the "constant power" cascode (going to the tail of the LTP) but I'm not sure. I can play around some more to see if I can get some concrete answers.
MJL21193 said:
John, I'm a fan of your enthusiasm and your ability to turn out products very quickly. However, I would urge you to go back and read your posts in many threads I have seen from you and re-evaluate your attitude to others. You claim to want to learn, but your actions do not support this claim. It may just be an 'internet communication' thing, but you're not coming across in a very good light.
Steve.
wg_ski said:
Ah, but very faint. I couldn't hear it myself, and it was more audible at mid frequencies.
I'm not ready for your "extreme" test yet - the interference amp is only held onto a temporary heatsink with some spit and boogers. I'd need to robustify the arrangement.
No problem to downgrade the PS, but to what end? Surely if this amp is built, it will be mated to a good power supply and, if at all feasible, to a separate front end supply. There is also the fact that when the amp is amplifying, it will be feeding a speaker that will be making sounds, potentially loud sounds. This will raise the noise floor and anything below it won't be heard.
Doesn't this make sense?
Oh and the only Metalica i have is legally downloaded, low quality MP3's
And I prefer "One"
MJL21193 said:
What I suggested was a practical and easy way to actually measure and plot an amplifier PSRR performance over a broad frequency range.
If you were here to only learn then you would have actually read what I had written, and maybe even though about it a little, instead of dismissing it with deliberately antagonistic nonsense and then complaining with personal invective when snubbed.
spind said:
Sure, let's start in this very thread.
How about post #17 or post #36 or post #40 or post #43 or post # 45 or post #48 or post #50 or post #51 or post #53 .
My treatment of roender was singled out a page ago or so. Those posts above are my "rough" attitude. Go ahead, pick out the offensive parts.
Don't worry spind, I do review my attitude and the things i say here all the time. The fact that even though I have a few posts in Texas, I have never been binned says something. I'm bad, but I'm not that bad.
Anyway, it would be pretty boring here if we were all cast in the same mold - a utopia as exciting as watching paint dry. Glen has his faults, but can be amusing at times. I'm not usually critical of others - live and let live for me. I'm not into personality modification, especially not someone else's.
I hear most of my criticism from my "fellow" Canadians. If you are that ashamed of my behavior, don't read my posts (ignore list). Be democratic about things and circulate a petition to submit to the mods to have me expelled, that kind of thing.
G.Kleinschmidt said:
You started it!
I'm sorry Glen, I am guilty of not fully understanding your suggestion. I should have asked for clarification, but you have been so mean to me lately...I was...afraid
Seriously, clarify your method. I missed something the first time. Like I said, I thought you were suggesting that I drive the power supply into a 4 or 8 ohm load.
I am an open vessel - pour the knowledge in.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.