Anyone interested in a digital amplifier project? - Page 12 - diyAudio
Go Back   Home > Forums > Amplifiers > Class D

Class D Switching Power Amplifiers and Power D/A conversion

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 27th September 2003, 09:51 AM   #111
deandob is offline deandob  Australia
diyAudio Member
 
Join Date: Oct 2002
Location: Brisbane, Australia
The XR10 sounds like a great platform for tweaking, the newform research link above shows the potential of this unit.

Interesting that several web based reviews give the unit an average score for sound quality, and that the unit seems to run out of puff at higher levels compared to conventional amps. Perhaps this is due to the problems Brian has outlined in the SMPS at high volumes, which might be fixable.

I was considering ordering a few of the TI eval boards but buying a used XR10 on ebay or audiogon & hacking it looks like a more effective option.

Brian, can you tell if the power supply is switchable from 110v to 220 or 240v? Keep those posts coming, there are several of us on the forum keen to track your progress.

Regards,
Dean
  Reply With Quote
Old 27th September 2003, 01:18 PM   #112
diyAudio Member
 
ashok's Avatar
 
Join Date: Jun 2002
Location: 3RS
Default Pictures.

Hi Fcserei,
Would it be possible for you to put up pictures of your horn loaded Stax full range Electrostatic speakers ?
I am sure many of us would like to see it.
Thanks,
Ashok.
__________________
AM
  Reply With Quote
Old 27th September 2003, 05:53 PM   #113
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Quote:
Originally posted by deandob
Brian, can you tell if the power supply is switchable from 110v to 220 or 240v?
My SA-XR10 is a U.S. model. Its main power supply will only run off of 120VAC. It doesn't have a universal input range.

I believe that there were other versions of this unit in other countries for 240VAC. My documentation only covers the U.S. version.

I think they based both versions on the same board, but it would take some time to work out the details.

The main supply provides ~45VDC for the Equibit board, which has its own variable voltage switcher for the bus. If you only want to use the Equibit board by itself, then it probably doesn't matter which version you get.

Brian.
  Reply With Quote
Old 27th September 2003, 06:01 PM   #114
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Default SA-XR10 Update - clarification

Quote:
Originally posted by Brian Brown
I use a 10K log taper pot with the top side connected to 3.3V through a 453ohm resistor and the low side connected to ground through a 1.13K resistor. This produces a log taper control voltage at the wiper with a range of 0.3V to 3.2V, which corresponds to a supply voltage range of about 4V to 42V.
I forgot to mention that I needed to use a unity gain opamp between the pot wiper and the VOLUME input of the Equibit board.

Brian.
  Reply With Quote
Old 27th September 2003, 06:48 PM   #115
diyAudio Member
 
LineSource's Avatar
 
Join Date: Sep 2003
Location: SiliconValley
Default driving 1 ohm loads with digital amp

I want to buiild a digital amp that can safely drive 0.7 to 1.0 ohm speaker loads with 25-50 watts, and deliver high quality sound with very low noise. Any suggestions on modifications to the TI or Tripath, or ?? designs for this low impedance?

Is there more to the design than solving:
Robust power supply of about 24 volts.
Very low Ron Mosfets. Trade-off lower VDD_max for lower Ron.
Driving many parallel output Mosfets.
  Reply With Quote
Old 28th September 2003, 06:54 PM   #116
jwb is offline jwb  United States
diyAudio Member
 
jwb's Avatar
 
Join Date: Mar 2002
Location: San Francisco, USA
Send a message via AIM to jwb
50W into 1Ω load means greater than 7A output current ability. I think the main problem would be finding a reconstruction filter with low enough R. Perhaps you could also have parallel filters, one for each output FET?
  Reply With Quote
Old 29th September 2003, 02:10 PM   #117
diyAudio Member
 
Join Date: Sep 2002
Location: US
Brian,

Thanks for the clarification.
What I really don't get, why are they decided to drop bits first in the volume control and then decrease the supply voltage. The S/N ratio is not that great that you can drop 4 bits easily without any sonic penalties. It would be more logical to reduce the voltage first for the 0- -20dB range and then drop resolution bits, especially as you wrote, the variable supply is not noise free at higher voltages.
Most of my listening is happening in the -10 - -25 dB range so I'm thinking to short R621 , thus reducing the multiplication coefficient of the variable supply. This way I'd use the 0- -15dB range with about 15V supply voltage without resolution loss at high volumes and better supply regulation.

What do you think?
  Reply With Quote
Old 29th September 2003, 02:42 PM   #118
diyAudio Member
 
Join Date: Feb 2002
Quote:
Originally posted by fcserei
Brian,

Thanks for the clarification.
What I really don't get, why are they decided to drop bits first in the volume control and then decrease the supply voltage. The S/N ratio is not that great that you can drop 4 bits easily without any sonic penalties. It would be more logical to reduce the voltage first for the 0- -20dB range and then drop resolution bits, especially as you wrote, the variable supply is not noise free at higher voltages.
Most of my listening is happening in the -10 - -25 dB range so I'm thinking to short R621 , thus reducing the multiplication coefficient of the variable supply. This way I'd use the 0- -15dB range with about 15V supply voltage without resolution loss at high volumes and better supply regulation.

What do you think?

I have been thinking of the volume control question as well, particularly in light of the fact that on my XR25 (which uses ONLY digital attenuation, no variable supply) I _still_ get incredible low-level resolution even when listening at -35 or even lower. This is not what I would have expected from a unit with only ~96dB S/N running 35dB below full-scale.

I think the answer has to lie in understanding the noise/distortion mechanisms in the Equibit chipset, and unfortunately I don't really know enough about how they work internally to know. That won't stop me from speculating, though :-)

If the noise/distortion is direcly related to uncertainty/jitter in the PWM master clock, then it will be largely independent of signal level, and will behave like a normal analog amp in that the noise floor will be pretty constant.
However, if the noise/distortion is somehow proportional to the input level (ie percentage inaccuracies in the PWM signal or switching times), then the noise floor would drop with lower signals, and so the unit would preserve that 96 db S/N performance even when operating below full-scale.

Remember that the unit isn't really 'dropping bits', it's only 'shifting bits' since the input to the amps will (presumably) preserve 24 bit word length. This operation by itself won't hurt anything - it all comes down the the noise characteristics of the amp section.

Unfortunately, the most logical source of noise would be jitter in the pwm clock or other inaccuracies in the on/off switching of each pulse, which would seem to be more 'absolute' than 'proportional' in nature, since they would be related to the on/off transitions and not the width of the pulse.....
  Reply With Quote
Old 29th September 2003, 02:57 PM   #119
diyAudio Member
 
Join Date: Sep 2002
Location: US
Quote:
Originally posted by dwk123



I think the answer has to lie in understanding the noise/distortion mechanisms in the Equibit chipset, and unfortunately I don't really know enough about how they work internally to know. That won't stop me from speculating, though :-)
Where could we learn more about the Equibit coding method. I've never seen any details about the actual modulation.

Quote:
Originally posted by dwk123


Remember that the unit isn't really 'dropping bits', it's only 'shifting bits' since the input to the amps will (presumably) preserve 24 bit word length. This operation by itself won't hurt anything - it all comes down the the noise characteristics of the amp section.

Unfortunately, the most logical source of noise would be jitter in the pwm clock or other inaccuracies in the on/off switching of each pulse, which would seem to be more 'absolute' than 'proportional' in nature, since they would be related to the on/off transitions and not the width of the pulse.....
Yes, this is the question. If it is 'absolute', digital volume control is OK, if 'proportional, then shifting bits is bad because of the limited resoulution of the amp.
Anyway. I'll try to reduce the supply voltage gain and check the effect.
  Reply With Quote
Old 3rd October 2003, 04:30 PM   #120
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Quote:
Originally posted by fcserei
Brian,

Thanks for the clarification.
What I really don't get, why are they decided to drop bits first in the volume control and then decrease the supply voltage. The S/N ratio is not that great that you can drop 4 bits easily without any sonic penalties. It would be more logical to reduce the voltage first for the 0- -20dB range and then drop resolution bits, especially as you wrote, the variable supply is not noise free at higher voltages.
It's not that the supply is noiser at higher voltages. Rather it's that the supply is noiser at higher current loads.

If the supply was very stiff and low impedance in the audio frequency range, then it would probably be best to do the initial -1dB to -20dB attenuation by decreasing the bus voltage.

I suspect that with the SA-XR10's particular supply, TI found that it was less noisy to have a higher voltage with a lower current load for the -1dB to -20dB region.

I plan to do some more experimentation in this area.

Quote:

Most of my listening is happening in the -10 - -25 dB range so I'm thinking to short R621 , thus reducing the multiplication coefficient of the variable supply. This way I'd use the 0- -15dB range with about 15V supply voltage without resolution loss at high volumes and better supply regulation.

What do you think?
For the reason I stated above, I don't think it would be beneficial to drop the bus voltage.

Also, there might be a problem because the SA-XR10's control processor probably expects to see a verification of the bus voltage at the DEC pin. A mismatch might cause the receiver to shut down with a fault condition. You could give it a try and see what happens.

This problem could be overcome by rescaling the DEC feedback resistors (change R634 from 27K to 8.45K). This rescaling could possibly cause the DEC signal to overvoltage the A/D converter, so a diode from DEC to +5V (RESET pin) to clamp the signal would be a good idea if you wanted to try this experiment.

Brian.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Noise measurement amplifier - interested? jackinnj Solid State 15 27th March 2008 09:36 PM
New digital XO project borges Multi-Way 2 8th January 2008 06:49 PM


New To Site? Need Help?

All times are GMT. The time now is 04:27 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2