Any easy way to implement a decent low pass filter (an ultrasonic cut) for the output of a preamp? I'm concerned about potential buildup of stuff if there is already noise or other junk in 96khz material and would like to be safe just in case. Any common implementations? A cut off above 20k but under 30k would be preferable. Thanks!
I always recommend an RF filter at the input of every bit of audio gear, (except a tuner).
A competent source should not emit excessive RF.
A competent source should not emit excessive RF.
Hi,
series R in the signal line followed by a C from line to signal ground.
Usual R values vary from 100r to 3k3.
Usual values for C vary from 100pF to 1n5F.
I find that I cannot hear any difference taking RC lower than ~0.7us.
Therefore I use 1k0 & 680pF giving RC=0.68us. This has an F-3dB ~230kHz
There is a big subjective variation in acceptable RC values. Experiment.
I also add a small 47pF across the input RCA phono socket to attenuate the VHF before it gets into the "box".
series R in the signal line followed by a C from line to signal ground.
Usual R values vary from 100r to 3k3.
Usual values for C vary from 100pF to 1n5F.
I find that I cannot hear any difference taking RC lower than ~0.7us.
Therefore I use 1k0 & 680pF giving RC=0.68us. This has an F-3dB ~230kHz
There is a big subjective variation in acceptable RC values. Experiment.
I also add a small 47pF across the input RCA phono socket to attenuate the VHF before it gets into the "box".
does this not add to the effective output impedance of the preceding stage? or all is fine provided the next stage input is high impedance enough?
At higher frequencies the input of the receiver sees the 680pF as source impedance.
At lower frequencies it sees the 1k0+Rs in parallel with 680pF.
Those should be low enough source impedances to satisfy most Power Amplifiers and low gain Pre-amps. If Rs=200r, then over much of the audio band the effective source impedance <=1200r
At lower frequencies it sees the 1k0+Rs in parallel with 680pF.
Those should be low enough source impedances to satisfy most Power Amplifiers and low gain Pre-amps. If Rs=200r, then over much of the audio band the effective source impedance <=1200r
right got it, so the simple answer is yes and yes 😉 quite a pertinent thread really, as i'm toying with setup for the main output of my dac for active monitors, I cant find the spec online, but I would assume most active studio monitors will have a reasonably high input Z. I have been eyeing them off lately for an upgrade, so will find out for sure what is at the input soon enough
If you need an ultrasonic filter which is 2nd order or above you might want to consider a bessel filter which is usually the best solution as it has excellent phase characteristics. Take a look at this datasheet:
http://www.ti.com/lit/an/snoa586d/snoa586d.pdf
On page 11 you will find an ultrasonic filter. You can also get filterpro from Texas Instruments which is free software for designing filters including bessel filters.
Huh?
Why would you need or want such a thing?
The so called 96kHz source material that is commercial is going to be free of any high level "garbage". The other thing is what speaker do you have that will reproduce any of it to begin with?
It would take a rather high level signal above 20kHz to do anything much to your amplifier.
Seems like a non issue.
Have you found a problem that has been measured by you or anyone else, or are you just worried about it?
_-_-
Why would you need or want such a thing?
The so called 96kHz source material that is commercial is going to be free of any high level "garbage". The other thing is what speaker do you have that will reproduce any of it to begin with?
It would take a rather high level signal above 20kHz to do anything much to your amplifier.
Seems like a non issue.
Have you found a problem that has been measured by you or anyone else, or are you just worried about it?
_-_-
NOS DACs often omit the required low-pass reconstruction filter so then you could get lots of ultrasonic image frequencies. Normally-engineered DACs will not, so no further filter would be necessary.
NOS?
You mean the DAC chip alone? Who's got that, except in the case where the person building with a DAC chip knows something about what they are doing?
Lost me here.
You mean the DAC chip alone? Who's got that, except in the case where the person building with a DAC chip knows something about what they are doing?
Lost me here.
The web is full of NOS DACs; some even appear in this forum. Some of them have little or no output filtering, thus guaranteeing that the output will not be a faithful reproduction of what went in at the other end.
I once had a problem with the output of a cheap *** keyboard hooked to a digital mixer via a cheap ADC, turns out the keyboard didn't exactly get the reconstruction filter right and the ADC was a bit defficient in the input filtering.
The result was a very audiable 'birdie', not at all what you need on stage setting up for a gig.
The cure was to mic the keyboard amp which did not have a response extending to multiple MHz.
NOS is asking for problems, NOS without a nightmare of an analogue filter is just stupid.
73 Dan.
The result was a very audiable 'birdie', not at all what you need on stage setting up for a gig.
The cure was to mic the keyboard amp which did not have a response extending to multiple MHz.
NOS is asking for problems, NOS without a nightmare of an analogue filter is just stupid.
73 Dan.
You might listen to this 3 minute AES audio/video clip.
"Bill Whitlock talks about audio pioneer Deane Jensen"
AES Oral History Project Gallery Learn From Audio Experts
I'll stick with the idea that it would take a fairly substantial amount of energy outside the GBW of ur electronics to have an audible effect.
It's much more likely to have that happen in a studio setting than in a home playback setting, which is where most of the Jensen audio transformers are used (the studio/live sound).
But I think the idea Bill Whitlock put up of Jensen's Spectral Contamination is valid. It's just multi-tone IM in essence. If it were weighted in some manner, that might make it even more interesting. I seem to recall reading somewhere that something along those lines is or has been done?
I guess the Deane Jensen idea is to look for non-harmonically related signals, which is also an interesting idea, IF such a thing actually happens.
So, yeah, I guess throwing a fair amount of HF energy that exceeds the signal chain's slew rate/bandwidth would create a problem. But back to the OP, is there actually a problem?
And, Jensen's transformers are fairly wide bandwidth items, usually 60kHz---> ~100kHz or better. A 96kHz digital source is coming in with signal under those bandwidths, so the advice of using a transformer as a "filter" is maybe questionable.
But is the OP talking about unfiltered DAC output or not??
_-_-
It's much more likely to have that happen in a studio setting than in a home playback setting, which is where most of the Jensen audio transformers are used (the studio/live sound).
But I think the idea Bill Whitlock put up of Jensen's Spectral Contamination is valid. It's just multi-tone IM in essence. If it were weighted in some manner, that might make it even more interesting. I seem to recall reading somewhere that something along those lines is or has been done?
I guess the Deane Jensen idea is to look for non-harmonically related signals, which is also an interesting idea, IF such a thing actually happens.
So, yeah, I guess throwing a fair amount of HF energy that exceeds the signal chain's slew rate/bandwidth would create a problem. But back to the OP, is there actually a problem?
And, Jensen's transformers are fairly wide bandwidth items, usually 60kHz---> ~100kHz or better. A 96kHz digital source is coming in with signal under those bandwidths, so the advice of using a transformer as a "filter" is maybe questionable.
But is the OP talking about unfiltered DAC output or not??
_-_-
this has been discussed briefly in many Threads here.
Maybe a new Thread specifically to discuss this is what is now required.
An example of a passive 21kHz lowpass is attached🙂.
The filter has 40 dB attenuation at 100kHz. and zero noise.
The preamp output impedance (100 Ohm) is important for correct filter response.
C3 is the line capacitance (about 100pF / meter)
Attachments
That quantity of treble attenuation will probably be rejected by most Members here.
I aim for 200kHz single pole passive filtering.
Other members generally use between 1MHz and 100kHz for the passive input filter of their Power Amplifiers.
I aim for 200kHz single pole passive filtering.
Other members generally use between 1MHz and 100kHz for the passive input filter of their Power Amplifiers.
- Status
- Not open for further replies.