Square wave

[Prune]

How low a frequency square wave should look square? 20 Hz?

[till]

my square waves look pretty square at 20kHz

[phase_accurate]

I thought the same at first !!!!

I think he is talking about the square-wave rsponse of some device. Since we don't know of which one (amp, speaker , ......) we can't give an answer.

Regards

Charles

[Prune]

No, no! I'm asked whether it should be square at 20 Hz, not 20 KHz!
As for the application, I'm asking because I was told the Aleph-X should be AC coupled to the source. However, using 10 uF (which is about as large quality film caps as tend to be reasonably priced) around a 10 K load (representing the Aleph-X input) in the simulator gives a pretty nasty square wave at 20 Hz. Using far larger caps means electrolytics... So, is 20 Hz too low to worry about square wave, or not?

[till]

ahhh, this is some more information.

What do you think how a sqare wve may look if coming from your speaker?

Go and try if you hear a difference between one 10uf cap or two or some more in parallel....

[Prune]

Oops, I just noticed it's 10K between a line and ground, so 20K between the balanced lines. Still crappy 20 Hz square, see attached image.

[phase_accurate]

This is much better than what most (> 99.9%) of all the speakers would reproduce when they are fed with a perfect squarewave !!

Regards

Charles

[Sch3mat1c]

That's normal low frequency drop off. Most of my tube amps show the same thing and they sound perfectly fine. Just don't tell them to reproduce any power under 50Hz.

Tim

[sam9]

That's just what a 10uF cap does to a LF square wave. If you are trying to test an amp with 20hz square wave you have to insert the signal after the cap, "it don't mean nothin". It tells you something about the cap but little or nothing about the amp.

Also it doesn't say much about whether the 10uF cap will adversley affect actual listening. Since you are simulationg it do seruies of AC analyses with a range of cap values to see if the frequency response is acceptable to you. Next do a series of transient analyses at 100Hz and 1kHz and use a .DISTO directive (or however your simulator gets you a THD or THD+N figure) with the same cap values. See if differences in distortion are enough to concern you.

10uF IMO get resonable results in most cases. 20uF gets better, but not a lot. Perhaps not enough to justify the hassle of using the higher value.

[Prune]

No, I'm just trying to figure out how big a DC blocking cap I should use, which means I need to know what square wave performance is supposed to be acceptable.