SI amp, high impedance phones, and ultrasonic noise
panomaniac,
A number of people are (or are considering) using the SI amp to drive headphones. With a 120 ohm or 300 ohm load, the SI's filter becomes pretty ineffective. There is a rise in it's response, reaching a peak of +10db at about 53kHz. I've been cautioning people about the ultrasonics and possible hearing damage that may result from listening to phones with this amp -- until we collectively answer a bunch of questions. For example, the Sennheiser HD-650 phones (300 ohm) are pretty much flat up to 35kHz, so you'd be listening to 35kHz noise that's 6db higher than the music level.
Check out these threads to see how the various arguments have developed:
http://www6.head-fi.org/forums/showthread.php?t=87285
http://www6.head-fi.org/forums/showthread.php?t=103035
If you can investigate this further, or shed some light on these issues, please let us know your results.
- Eric
panomaniac,
A number of people are (or are considering) using the SI amp to drive headphones. With a 120 ohm or 300 ohm load, the SI's filter becomes pretty ineffective. There is a rise in it's response, reaching a peak of +10db at about 53kHz. I've been cautioning people about the ultrasonics and possible hearing damage that may result from listening to phones with this amp -- until we collectively answer a bunch of questions. For example, the Sennheiser HD-650 phones (300 ohm) are pretty much flat up to 35kHz, so you'd be listening to 35kHz noise that's 6db higher than the music level.
Check out these threads to see how the various arguments have developed:
http://www6.head-fi.org/forums/showthread.php?t=87285
http://www6.head-fi.org/forums/showthread.php?t=103035
If you can investigate this further, or shed some light on these issues, please let us know your results.
- Eric
Re: SI amp, high impedance phones, and ultrasonic noise
Prior to dealing with issues of the load dependancy of the 2nd order lowpass output filter [see the tripath 2024 pdf for an explanation and equations] and its effect on the filter's 'Q', confirm that both the headphones and cabling to be used do not have a common ground. The amp is a btl (bridge tied load) with a single rail, and none of its ouputs sits at ground; it will not work if the phones or cable are wired 'common ground'.
Eric Weitzman said:
Prior to dealing with issues of the load dependancy of the 2nd order lowpass output filter [see the tripath 2024 pdf for an explanation and equations] and its effect on the filter's 'Q', confirm that both the headphones and cabling to be used do not have a common ground. The amp is a btl (bridge tied load) with a single rail, and none of its ouputs sits at ground; it will not work if the phones or cable are wired 'common ground'.
Bass responce at differeing loads
I did check the bass roll off at some higher levels. At near max output (8V P-P), there is a little more bass drop, it is down -10.5 dB versus -9 dB at levels around 1 watt.
There seems to be only the very slightest difference in level between 4, 8 or 16 ohms. You'll have more residual ultrasound at the higher impedance.
I agree with Motherone that it is probably the input caps and resistors that are making a hi-pass filter. I will try to bypass the stock units with outboard components of the correct values.
Maybe the input pot has something to do with it? Acting as part of the LC hi-pass, perhaps?
Aloha,
MM
I did check the bass roll off at some higher levels. At near max output (8V P-P), there is a little more bass drop, it is down -10.5 dB versus -9 dB at levels around 1 watt.
There seems to be only the very slightest difference in level between 4, 8 or 16 ohms. You'll have more residual ultrasound at the higher impedance.
I agree with Motherone that it is probably the input caps and resistors that are making a hi-pass filter. I will try to bypass the stock units with outboard components of the correct values.
Maybe the input pot has something to do with it? Acting as part of the LC hi-pass, perhaps?
Aloha,
MM
I'm almost certain it's a function of the input capacitor, or something related on the input of the amp. There is an input inductor (L1 on top and L2 on the bottom), as well as C3 (bottom) and C4 (top) before it hits the 20k input resistors, R2 (top) and R1 (bottom).
If you look at page 7 of the TA2024 PDF, you'll see that it's on .2 to .4 db down, depending on the impedance it's driving.
I suspect that the input capacitors are the culprit, though. They're so small that I have a hard time believing they're anywhere close to what's on the example on the datasheet.
I would suggest jumpering C3 and C4, and adding your own outboard input capacitors and see if that makes a difference.
Also, it's probably possible to add an external lowpass filter for the amp if ultrasonics are a concern.
If you look at page 7 of the TA2024 PDF, you'll see that it's on .2 to .4 db down, depending on the impedance it's driving.
I suspect that the input capacitors are the culprit, though. They're so small that I have a hard time believing they're anywhere close to what's on the example on the datasheet.
I would suggest jumpering C3 and C4, and adding your own outboard input capacitors and see if that makes a difference.
Also, it's probably possible to add an external lowpass filter for the amp if ultrasonics are a concern.
Re: SI amp, high impedance phones, and ultrasonic noise
This is (mostly) reasonable advice. However,
1. If you review the referenced threads, you will note that the need to recable the headphones to separate the common ground was made known to people who were unaware of this.
2. You will also note that the spice model and results that I posted contain the reference circuit from the data sheet. Also, the data sheet is referenced (with it's URL) in several posts.
3. Rewiring headphones is not a prerequisite for modeling the circuit, understanding the effects of ultrasonics below 100kHz on human hearing, or making measurements using dummy 120 ohm loads that don't share a common ground.
4. The amp does indeed work if the two channels are tied together. What it sounds like is another matter entirely, known only to those who have tried it
pmkap said:
This is (mostly) reasonable advice. However,
1. If you review the referenced threads, you will note that the need to recable the headphones to separate the common ground was made known to people who were unaware of this.
2. You will also note that the spice model and results that I posted contain the reference circuit from the data sheet. Also, the data sheet is referenced (with it's URL) in several posts.
3. Rewiring headphones is not a prerequisite for modeling the circuit, understanding the effects of ultrasonics below 100kHz on human hearing, or making measurements using dummy 120 ohm loads that don't share a common ground.
4. The amp does indeed work if the two channels are tied together. What it sounds like is another matter entirely, known only to those who have tried it
Re: Re: SI amp, high impedance phones, and ultrasonic noise
Well I have tried it, by accident. With the grounds tied together into an 8 ohm load, the amp is very unhappy. Mostly it clicks strongly, a sort of cry for help. Is that what you mean by "the two channels tied together." ?
I do not know what happens with a common ground on a high impendance load, but I can't see why it would be different.
Aloha
MM
Eric Weitzman said:4. The amp does indeed work if the two channels are tied together. What it sounds like is another matter entirely, known only to those who have tried it
Well I have tried it, by accident. With the grounds tied together into an 8 ohm load, the amp is very unhappy. Mostly it clicks strongly, a sort of cry for help.
Is that what you mean by "the two channels tied together." ?I do not know what happens with a common ground on a high impendance load, but I can't see why it would be different.
Aloha
MM
Re: Bass Shy
http://www.diyaudio.com/forums/showthread.php?s=&threadid=50664&highlight=
Turns out the required filter is almost exactly what's shown by your testing (a 50Hz HPF around .5 Q).
The current demands into a 4-ohm load are suprising as well, and suddenly make my 1700ma supply look a little anemic.
Thanks for the testing, though!
That's funny! I built a couple of 4-ohm fullrange speakers for my SI, and solicited help on this forum to design the required highpass filter to prevent overexcursion on these speakers:panomaniac said:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=50664&highlight=
Turns out the required filter is almost exactly what's shown by your testing (a 50Hz HPF around .5 Q).
The current demands into a 4-ohm load are suprising as well, and suddenly make my 1700ma supply look a little anemic.
Thanks for the testing, though!
Re: Bass Shy
You may be alright with 1.7 amps. It depends on what you listen to.
The bass response chart is now on my site, along with a page called "Musical Power." Check it out to see how much power you might actually be using. As long as your PSU has a big cap on the output for peaks, you should be OK. Please have a look and let me know if you have any questions.
http://www.michael.mardis.com/sonic/start.html
Aloha
MM
bwbass said:
You may be alright with 1.7 amps. It depends on what you listen to.
The bass response chart is now on my site, along with a page called "Musical Power." Check it out to see how much power you might actually be using. As long as your PSU has a big cap on the output for peaks, you should be OK. Please have a look and let me know if you have any questions.
http://www.michael.mardis.com/sonic/start.html
Aloha
MM
Musical Power
I have added a page to my Sonic-Tripath site called
Musical Power
http://www.michael.mardis.com/sonic/start.html
On this page I explain how little musical power you can really get out of the Tripath chip, if you want to avoid clipping. Some nice waveforms are included showing average levels in various types of music.
Please have a look and let me know if you have any questions, ideas or disagreements.
Aloha
MM
I have added a page to my Sonic-Tripath site called
Musical Power
http://www.michael.mardis.com/sonic/start.html
On this page I explain how little musical power you can really get out of the Tripath chip, if you want to avoid clipping. Some nice waveforms are included showing average levels in various types of music.
Please have a look and let me know if you have any questions, ideas or disagreements.
Aloha
MM
Erukian said:
As noted above, it all depends on how loud you go. You will get a good 4.5 watts before clipping, that should be very loud on your klipsch. You'll still only get ~1/8 watt average musical power per channel without clipping. Unless you have a huge room, that will be pretty loud at 94dB/watt.
You may not want to run the SI pot all the way open. Try 1/2 way first and make some tests. Let us know what you find.
FYI, the 10 o'clock position gives unity gain .
From reading about the SI amp, i've picked up from somewhere that the pot does some bad things to the sound if it's not at 100%. Because i've read that if it's at 100% and if you hook up a good pre-amp then it sounds better. Is that true then?
It's pretty easy for me to add a trim pot to my output stage on my professional soundcard. So add like a -30 db trim and then it should all be righteous. right?
It's pretty easy for me to add a trim pot to my output stage on my professional soundcard. So add like a -30 db trim and then it should all be righteous. right?
If you're not afraid to open the SI up and do some soldering, you could just bypass the potentiometer altogether. To do this, you would need to open up your SI, pop out the volume board, and solder the volume in for each channel to the volume out.
To do it right, though, I would just suggest popping the main board off it's support and solder pin 3 to pin 7 and pin 4 to pin 6 on the volume board connector. Then just remount the board, maybe with some hot glue, or remelt the plastic, and you should be good to go.
Volume connector pinout:
(from Left to Right, looking down at the top of the board):
1 2 3 4 5 6 7
1 = switch 12volt out
2 = switch 12volt in
3 = to volume from right channel
4 = to volume from left channel
5 = ground
6 = from volume control to amp left channel
7 = from volume control to amp right channel
To do it right, though, I would just suggest popping the main board off it's support and solder pin 3 to pin 7 and pin 4 to pin 6 on the volume board connector. Then just remount the board, maybe with some hot glue, or remelt the plastic, and you should be good to go.
Volume connector pinout:
(from Left to Right, looking down at the top of the board):
1 2 3 4 5 6 7
1 = switch 12volt out
2 = switch 12volt in
3 = to volume from right channel
4 = to volume from left channel
5 = ground
6 = from volume control to amp left channel
7 = from volume control to amp right channel
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.