kacernator said:
Are we talking guitar amp here, or HiFi?
I don't know what you mean by "stiffer sound". When designing capacitor input filters with SS diodes, I like to keep the initial capacitor low -- not over 47uF -- just as with a hollow state PS. The power xfmrs, such as the Hammond "Classic" series weren't designed to accommodate the Isurge values that silicon can easily handle. Working into a big capacitor, with an Isurge well above an amp, is something that a silicon diode will handle easily, but the xfmr wasn't intended to handle that.
As for subsequent ripple filtering with an LC lowpass filter, what's important is keeping the LF poles well away from the dominant LF pole (usually that of the OPT) to avoid such nasties as "filter bounce" and LF instability under gNFB.
For a HiFi amp, good voltage regulation is desireable. Maybe not so much in a gee-tah amp where you like overdrive distortion, and power supply collapse for increased compression and "sustain". (Which is why you frequently see undersized power xfmrs and smaller filter capacitors in these designs.)
kacernator said:
I've found that bass response and tightness has more to do with the quality of the OPT, the woofies you're using and how much damping they need to keep 'em under control. Adding gNFB helps by making the sonics less dependent on circuit and component peculiarities.
As for ripple filters, I design 'em to be quiet and stable. Whatever values of L and C make that happen are the ones I use. In push-pull designs, the PS is out of the signal path since cancellation occurs at the primary center tap, making it a virtual ground point. What more can you do? Other considerations are an invitation to veer off into the audiophoolery wilderness of unsubstantiated, overblown claims, old wives' tales and urban legends and other nonsense.
Geek said:
I didnt mean it like that. I am know the techie stuff of PSU, but I am asking how do you for example determine the last PSU cap for one channel of Aikido preamp? Generaly speaking if you have flabby bass response you go up by 100uF or so? Or is there a rule or something when have certain capacitance in you PSU and increasing it will not change much in a sound point of view? I mean going from 22uF to 220uF can be clearly heard. Can be also clearly heard 440uF to 1000uF?
Sadly, this is very true. I haven't seen much of it here but at AA there has been an abundance of dogmatic nonsense. Much of it flies in the face of common sense and the chief culprit is a self-confessed anti-engineer.
Flabby bass is more likely to be due to inadequate damping of the speaker - a combination of the speaker's own mechanical damping and the amp's electrical damping - than a problem with the final smoothing cap.
There is definitely an upper limit to what will help in the final cap - a law of diminishing returns, if you like. Salas's 'rule of thumb' seems reasonable to me. In my 15w + 15w stereo amp, with EL34s triode-strapped in PP and a common power supply, I have 470uF as the final cap. That's probably a bit more than I need and I'm sure that adding any more would be of no benefit.
In my Aikido preamp I have 47uF-choke-440Uf foreach channel.
I added 220uF parallel to 47uF for each channel and bass accuracy and clearnes is much better. However the sound is tighter but not as in your face.
So how do you know, concerning the sound, that you have enough filtering capacity or too much or too low?
I added 220uF parallel to 47uF for each channel and bass accuracy and clearnes is much better. However the sound is tighter but not as in your face.
So how do you know, concerning the sound, that you have enough filtering capacity or too much or too low?
kacernator said:
For technical evaluation, you have enough capacitance when you have low enough ripple.
For sonics, you have best capacitance when in a neutral sounding system the bass is extended and full and bouncy, without being slow nor hasty. The mids must be relaxed but not thick. In general, you must get neither tight and stressed nor sluggish music.
Use good neutral headphones to be sure, since you will avoid speaker & room deviations. This way you will know that you don't compensate for system anomalies elsewhere.
Its better to start low capacitance and add little by little. Also know that lytics have strong signatures and with another brand you may find that another value proves closer to what you need. My best recommendation is Mundorf M-Lytic.
It does not mention in the help menu of PSU II. I can only see that 'Max' in the results table, is what the top of the DC line on the graph represents for value. Also, when roughly filtered, the graph DC line result is thick and jagged, showing pk-pk ripple. But after some heavier filtering, the line gets thin enough. Unfortunately there is a visual end point that filtering more, does not make it any thinner. I wish that PSU II had a 'Ripple' result too.
If someone knows how to calculate ripple from the results table of PSU II pls enlighten us.
If someone knows how to calculate ripple from the results table of PSU II pls enlighten us.
Svein_B said:
Set the simulation delay to a value high enough that the voltages are stabilized. 1 second is usually high enough. Now the diff value is the ripple p-p.
Another thing to do is to zoom into the voltage plot using the mouse pointer to draw a square.
Svein.
For even 7000 msec I get Max=Diff. Strange.
Lowrider said:Very strange, I dont get any graphs now, zero...
Make sure that you have ticked any capacitor voltage(s). Check the voltage table min/max/diff.
If you run the sim with a delay you could get just a straight line at the bottom if there is no ripple. Try the zoom buttons also.
salas said:
For even 7000 msec I get Max=Diff. Strange.
You need to set the delay time in seconds (second box), and not increase the simulation period.
SveinB.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.