Low Drop Out Power Rail Filter - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

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 25th January 2013, 08:24 PM   #11
bravi is offline bravi  New Zealand
diyAudio Member
 
Join Date: Mar 2004
Location: Wellington, New Zealand
Default Actual performance

Hi Pete,

Interesting circuit idea for high currents. Have you had a chance to build a prototype and measure actual performance?

Cheers!
  Reply With Quote
Old 25th January 2013, 11:13 PM   #12
diyAudio Member
 
Join Date: Oct 2006
I will be ordering the parts to build it along with a 300W high bias class AB amplifier (up to 150W class A before reverting to AB). I hope to start construction very soon.
  Reply With Quote
Old 26th January 2013, 01:49 AM   #13
diyAudio Member
 
Sch3mat1c's Avatar
 
Join Date: Jan 2003
Location: Milwaukee, WI
Send a message via ICQ to Sch3mat1c Send a message via AIM to Sch3mat1c
~60dB is comparable to commercial LDOs (except they regulate as well; suppose you could get a low voltage one and bias it up to some fraction of Vin to use it for filtering). I'm guessing the higher frequency performance isn't as good, i.e., you wouldn't get away with it on a switching supply.

Two things bound such a design: 1. as the MOSFET saturates (voltage drop gets near the minimum of Rds(on) * Id), transconductance drops sharply and isolation similarly disappears. In this regime, you depend entirely on the performance of the driver/amplifier to cover low frequency response. 2. But that amplifier must be compensated at high frequencies, else it's just a big oscillator. So you might get, say, 100dB at 1Hz, but it's already tapering off by 60Hz (as shown, overall gain is probably more like 60dB, which is about your measured attenuation, so the cutoff frequency is probably a smidge over 100Hz), dropping to nil at fT (maybe a few hundred kHz). 3. A corollary to these: because the gain changes rapidly in saturation, compensation must also change, or be overcompensated enough to keep things in check. This can hurt performance further. (The high-voltage-drop case should really be worst case, but do remember to check it in all conditions, in any case! It's very easy to ignore low-line and startup conditions in simulations.)

Incidentally, for simulation purposes, there's no need to generate 17 painstaking seconds of data waiting for sources to stabilize -- just use a DC source with AC ripple superimposed. Use an AC (small signal) analysis to check the rough frequency response (being sure to check the DC operating points, of course).

The load you use is also noteworthy: it's true that, for a small change, almost any load looks like a current source, but since you're looking at small changes in current AND voltage, it still matters whether it's a resistor, current sink or what. Now, a solid state amplifier *does* tend to look like a current sink -- but I'm not going to make that assumption, because if you need this much filtering, your circuit's PSRR must be poor to begin with, or your signals very small. In either case, check for best results!

Tim
__________________
Seven Transistor Labs
Projects and Resources
  Reply With Quote
Old 26th January 2013, 03:28 AM   #14
diyAudio Member
 
Join Date: Oct 2006
Thanks for the thoughtful reply.

Here's a shot of the transfer function showing the 0dB rejection at DC (LDO output tracks average line level) with maximum attenuation at just over 100Hz and then slowly rising after that. But there's still 40dB attenuation of 10KHz input noise on the output. This transfer function is with a typical ampilfier load at idle (500mA bias draw).

I'm using a MOSFET with Rds at full on that won't exceed .06 Ohm. Even with 15A peak draw, the device won't be dropping even 1V. The only time the device would ever see that kind of current draw would be with 4 Ohm loads pulling 500W, which will be exceedingly rare. Even then, all that happens is the device simply acts as a .06 Ohm resistor for the duration of the peak current draw. During the amplifier operation in Class B, when the positive current draw goes to 0, there is some overshoot on the LDO recovery. I've included some other shots showing the input ripple using an AC source of 1V@120Hz and 70V offset, and .15 Ohm source impedance. The load is an amp using a multitone music simulation signal. The closeup of the LDO output looks a bit hairy, but in reality, the levels are quite small. I've included a spectral content of the rail under load as well. The two spikes you see are part of the test signal fed to the amplifier load. The spurious stuff is over 100dB below the output signal level (20dB).

The amp circuit I will be pairing with this LDO already has PSRR of around 80dB at 120Hz. Increasing that to a total of 120dB or more doesn't seem like a waste of time if you're trying to make the quietest amp you can, especially for use with high sensitivity speakers.
Attached Images
File Type: gif LDO Line Filter Xfer Function.gif (12.6 KB, 56 views)
File Type: gif Input - Output - Load current.gif (23.6 KB, 53 views)
File Type: gif output - load current.gif (22.4 KB, 55 views)
File Type: gif LDO rail spectrum under load.gif (12.6 KB, 54 views)
  Reply With Quote
Old 26th January 2013, 04:50 AM   #15
diyAudio Member
 
Join Date: Oct 2006
Thanks for the thoughtful reply.

Here's a shot of the transfer function showing the 0dB rejection at DC (LDO output tracks average line level) with maximum attenuation at just over 100Hz and then slowly rising after that. But there's still 40dB attenuation of 10KHz input noise on the output. This transfer function is with a typical ampilfier load at idle (500mA bias draw).

I'm using a MOSFET with Rds at full on that won't exceed .06 Ohm. Even with 15A peak draw, the device won't be dropping even 1V. The only time the device would ever see that kind of current draw would be with 4 Ohm loads pulling 500W, which will be exceedingly rare. Even then, all that happens is the device simply acts as a .06 Ohm resistor for the duration of the peak current draw. During the amplifier operation in Class B, when the positive current draw goes to 0, there is some overshoot on the LDO recovery. I've included some other shots showing the input ripple using an AC source of 1V@120Hz and 70V offset, and .15 Ohm source impedance. The load is an amp using a multitone music simulation signal. The closeup of the LDO output looks a bit hairy, but in reality, the levels are quite small. I've included a spectral content of the rail under load as well. The two spikes you see are part of the test signal fed to the amplifier load. The spurious stuff is over 100dB below the output signal level (20dB).

The amp circuit I will be pairing with this LDO already has PSRR of around 80dB at 120Hz. Increasing that to a total of 120dB or more doesn't seem like a waste of time if you're trying to make the quietest amp you can, especially for use with high sensitivity speakers.

I just got through a simulation that increased the bias current in the differential pair and it definitely improved the recovery of the MOSFET from the no-current to on-condition. I'll post an update to the configuration after I've played around some more.

Last edited by pete_schumacher; 26th January 2013 at 04:52 AM. Reason: Further analysis
  Reply With Quote
Old 27th January 2013, 08:45 PM   #16
diyAudio Member
 
Join Date: Oct 2006
OK, some progress and a thorough simulation. The power supply is fed with full bridge rectifier and AC voltage sources. The amplifier is fed with a 5 tone test signal simulating a real world music signal. Peak output power into 4 Ohm load resulting in 930W instantaneous peaks and and 15A peak current through the LDO.

The pictures show spectral content of the amp output, the LDO rail, and the rectifier side of the LDO. Also shown are the input and output of the LDO in the time domain as well as a shot of the amp output signal.

It appears that the use of the LDO greatly cleans up the power rail, far more than even doubling the rectifier cap bank would do at a much lower cost. This simulation is done with 30,000uF on the rectifier filter and 6000uF on the LDO output.

Click the image to open in full size.

Click the image to open in full size.

Click the image to open in full size.

Click the image to open in full size.

Click the image to open in full size.
  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
Low pass filter for simple sub on single rail? tubee Chip Amps 9 21st February 2008 02:28 AM
Filter Cap Directly from Rail to Rail, excluding Center Tap Workhorse Power Supplies 8 29th August 2007 12:33 PM
Need low power high beta transistors - please drop me a few part numbers... peufeu Solid State 37 3rd November 2006 12:22 AM
Rail voltage drop in preamp mmerig Solid State 4 25th May 2006 04:12 AM
Rail drop test darkfenriz Power Supplies 8 24th February 2006 01:38 PM


New To Site? Need Help?

All times are GMT. The time now is 12:36 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