designing a 1-Hz lowpass with no DC error - Page 2 - diyAudio
Go Back   Home > Forums > Source & Line > Analog Line Level

Analog Line Level Preamplifiers , Passive Pre-amps, Crossovers, etc.

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 11th July 2013, 05:02 AM   #11
Zero D is offline Zero D  United Kingdom
diyAudio Member
 
Join Date: Aug 2009
@ CharlieLaub

Well how about that ! I was next going to suggest removing the Active LPF & input OpAmp, & instead passively stringing @ least 3 x 6dB Oct 1Hz LPF's in series, @ the input to 2 x gain stages as in my screenie. The reason was due to your not wanting to do any offset trimming, so cutting out 1 stage would help.

If it works ok with TL072's, then If not the OPA177's or similar Will do the trick

Keep us posted.
  Reply With Quote
Old 11th July 2013, 06:54 PM   #12
diyAudio Member
 
Join Date: Mar 2007
Location: California
Quote:
Originally Posted by Zero D View Post
@ CharlieLaub

Well how about that ! I was next going to suggest removing the Active LPF & input OpAmp, & instead passively stringing @ least 3 x 6dB Oct 1Hz LPF's in series, @ the input to 2 x gain stages as in my screenie. The reason was due to your not wanting to do any offset trimming, so cutting out 1 stage would help.

If it works ok with TL072's, then If not the OPA177's or similar Will do the trick

Keep us posted.
After trying lots of different amplifiers, it looks like I do need to use a "zero-drift" or chopper stabilized type. The low tempco of a drift-stabilized amp means I only need to calibrate once and the input offset is very low as well.

So far, a good choice seems to be the ICL7652 from TI. I will post some more info on the circuit, and the results of sims, shortly.

-Charlie
  Reply With Quote
Old 11th July 2013, 08:31 PM   #13
diyAudio Member
 
Join Date: Mar 2007
Location: California
Here are more details on the circuit, and application.

GOAL:
The goal is to dynamically measure the voice coil resistance of a driver during use. A "sense" resistor is connected after the driver, to ground, and the driving amplifier has an intentional DC voltage offset of a few millivolts. The ratio of the voltage drop at DC across the driver and across the sense resistor can be used to determine the voice coil DC resistance, Re. The changes in Re with power input due to VC heating can be tracked by the circuit described below, with a time resolution of about 1 second. This should make it suitable for use with woofers and larger midrange drivers.

CIRCUIT DESCRIPTION:
The circuit that I have been using to simulate this is attached below as "Driver_VC_resistance_measurement_circuit" and "Arduino_processing_of_VC_measurement". To sense the DC voltage I use three single pole RC low-pass filters, each having a corner frequency of about 1 Hz. The filtered signal is then sent through one or more gain stages. One gain stage operating on the voltage coming from Rsense is used to multiply by the ratio of Re/Rsense, which is 35 times using the loudspeaker model shown in the circuit. There are two separate measurements taken, one at the amplifier output and the other at the sense resistor. As shown in the attachment "Arduino_processing_of_VC_measurement" the gain-scaled voltages will be acquired by an Arduino, DC offset removed from each voltage feed, and then the ratio calculated. Because of the 35x gain stage mentioned above, the ratio of the voltages is now equal to Re(T)/Re(T=To) where T is the current temperature of the voice coil and To is some reference temperature (e.g. room temperature) to which the system was calibrated. Although temperature is not directly (or indirectly) measured it could be calculated using the tempco of the voice coil material. The Arduino is useful as way to monitor and display information and also because it is a simple matter perform the division operation in software. The Arduino can sample 0-5V with a 10-bit (4.7mv) resolution. In order to make better use of the range and resolution, the gain stages are used to bring up the nominal voltage to around 3V when a 10mV DC offset is supplied by the amplifier. The full 5V range could be used by increasing the amplifier's DC signal to 15mV, or by using more gain.

CIRCUIT SIMULATIONS (see attached figures):
  • "offset_calibration" shows the response of the circuit with the amplifier output grounded (no DC offset voltage). For this sim, the offset correction voltages in the Arduino were set to zero. The voltages stabilize to the DC offset of the circuit, and these are subsequently programmed into the Arduino to remove the effect of the small offset on the voltage ratio calculation.
  • "Response_with_10mV_DC_input" is a sim of the circuit response with the DC offset corrections applied, and the amplifier set to have a 10mV DC offset (no AC signal present). The calculated ratio "V_diff" is very close to 1.0, meaning that the current value of Re is the same as when Re is at its "cold" value of 3.5 ohms (see loudspeaker model).
  • "Response_with_10mV_DC_input_zoom" is the portion of "Response_with_10mV_DC_input" lying between 4 and 5 seconds, to more clearly show that V_diff is very close to 1.0.
  • "Response_to_music_input" is a sim of the circuit to a music input. This was done by reproducing the first 15 seconds of Tracy Chapmans' "Fast Car" by the amplifier with voltage swing up to about 40Vpk-pk. The response of V_diff still shows the correct value of 1.0.
  • "Response_to_music_input_with_Re_doubled" shows the effect of doubling Re in the loudspeaker model using the same "Fast Car" sim as above. V_diff has increased to approximately 2, indicating that Re has doubled.

THOUGHTS:
It looks like this circuit could work for dynamically measuring Re. In the sim, I have used a low-offset, chopper-stabilized op-amp that also has a very low DC-offset temperature dependence. This may make it possible to do a one-time calibration only. I need to check the influence of the tempco of the resistors in the gain stages, since these can also effect the performance. This may only be of concern if all the circuitry was located in the same housing as the amplifier or other heat-producing circuitry. If temperature dependence is not of concern, it may be possible to use other amplifiers depending on the amount of DC offset that will be produced. I have not make any effort to reduce DC offsets in the gain stages, and some changes in component values or topology (inverting gain stages?) might reduce the offset further. Any suggestions are welcome in this area.

One issue that I have not yet tackled is how to get the power amplifier to generate a reasonably stable DC offset of around 10mV. I thought that a DC servo could be used, but instead of trying to zero out the DC offset the servo would bring the offset voltage to the desired setpoint. Since I can directly measure the DC offset, I could also use one of the Arduino's output pins to send a voltage through a resistor to the amplifier's inverting or non-inverting input to control the DC offset. Both of these will require a dedicated amplifier design, e.g. you probably won't be able to use a commercial amplifier since few are DC coupled, unless by itself it happens to have an appropriate and stable DC offset...

That's where things stand for now...

-Charlie
  Reply With Quote
Old 12th July 2013, 01:20 AM   #14
expert in tautology
diyAudio Member
 
bear's Avatar
 
Join Date: Apr 2002
Location: New York State USA
Perhaps a silly question, but the changes in Re due to heating could be measured directly by merely heating the VC directly?

Or is there some factor in the heating when driven with "music" that may yield some other information?

The other thing is, capacitively couple an amplifier - maybe even with the F3 point being fairly higher than usual, >20Hz or more, and then inject a stable DC offset?

With nothing below >20Hz it might be easier to measure the DC?

_-_-
__________________
_-_-bear
http://www.bearlabs.com -- Btw, I don't actually know anything, FYI -- [...2SJ74 Toshiba bogus asian parts - beware! ]
  Reply With Quote
Old 12th July 2013, 03:10 AM   #15
diyAudio Member
 
Join Date: Mar 2007
Location: California
Quote:
Originally Posted by bear View Post
Perhaps a silly question, but the changes in Re due to heating could be measured directly by merely heating the VC directly?

Or is there some factor in the heating when driven with "music" that may yield some other information?

The other thing is, capacitively couple an amplifier - maybe even with the F3 point being fairly higher than usual, >20Hz or more, and then inject a stable DC offset?

With nothing below >20Hz it might be easier to measure the DC?

_-_-
No questions are silly...

To be more specific, the goal is to measure how a music signal is modulating Re, so directly heating the VC won't really be helping in this regard.

Yes, you MUST AC couple the amplifier or any signals below the corner frequency of the filters will pass, and then go through the gain stage. This would cause the op amps to saturate (bad). We only want to see the "DC" passing through Re and the sense resistor.

-Charlie
  Reply With Quote
Old 12th July 2013, 02:59 PM   #16
expert in tautology
diyAudio Member
 
bear's Avatar
 
Join Date: Apr 2002
Location: New York State USA
AC couple the *output* of the amplifier, then inject DC after the cap.

_-_-
__________________
_-_-bear
http://www.bearlabs.com -- Btw, I don't actually know anything, FYI -- [...2SJ74 Toshiba bogus asian parts - beware! ]
  Reply With Quote
Old 12th July 2013, 03:21 PM   #17
diyAudio Member
 
Join Date: Mar 2007
Location: California
Quote:
Originally Posted by bear View Post
AC couple the *output* of the amplifier, then inject DC after the cap.

_-_-
That's already been done. See:
Hot Stuff: Loudspeaker Voice-Coil Temperatures | Stereophile.com

I think that my approach is simpler, since it uses the power amplifier itself to generate the DC signal.

-Charlie
  Reply With Quote
Old 12th July 2013, 04:00 PM   #18
expert in tautology
diyAudio Member
 
bear's Avatar
 
Join Date: Apr 2002
Location: New York State USA
Charlie,

I am curious what you are hoping to reveal through this experiment?
Seems to me that the conversion efficiency of loudspeakers is well known, as well as the "power compression" effects at high outputs (also high VC heating). Wondering what else you are looking for?
__________________
_-_-bear
http://www.bearlabs.com -- Btw, I don't actually know anything, FYI -- [...2SJ74 Toshiba bogus asian parts - beware! ]
  Reply With Quote
Old 12th July 2013, 05:25 PM   #19
diyAudio Member
 
Join Date: Mar 2007
Location: California
Quote:
Originally Posted by bear View Post
Charlie,

I am curious what you are hoping to reveal through this experiment?
Seems to me that the conversion efficiency of loudspeakers is well known, as well as the "power compression" effects at high outputs (also high VC heating). Wondering what else you are looking for?
Once you can quantify something, then you are free to manipulate it in a prescribed way or compensate for it, and that can be advantageous in some applications. That non-answer will have to suffice for now.
  Reply With Quote
Old 13th July 2013, 03:20 AM   #20
Zero D is offline Zero D  United Kingdom
diyAudio Member
 
Join Date: Aug 2009
@ CharlieLaub

Aha The plot thickens !

I had a feeling there was more to it. I'm looking forward to the "non" becoming Known

All the best with it
  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
DC Coupling for MC Cartridges - how much error voltage before coil damaging tiefbassuebertr Analogue Source 20 10th June 2013 08:09 AM
Hypex SMPS DC Error detection and DC blocking cap bypass MaartenJ Class D 4 7th January 2011 10:58 AM
Most critical frequency response area? 192 Hz to 384 Hz hebrer Full Range 19 24th August 2010 07:07 AM
Designing a simple dc filament supply? Dr.Frankenstein Tubes / Valves 5 27th September 2004 10:31 AM
Seeking advice on my first shot in designing a 5842 DC'ed to 45 arnoldc Tubes / Valves 2 18th August 2003 05:30 PM


New To Site? Need Help?

All times are GMT. The time now is 06:56 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