
Home  Forums  Rules  Articles  diyAudio Store  Blogs  Gallery  Wiki  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 
Analog Line Level Preamplifiers , Passive Preamps, 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 

Thread Tools  Search this Thread 
23rd October 2012, 05:02 PM  #281 
diyAudio Member
Join Date: Oct 2011

Very nice ...of full moons and capacitor rolling
__________________
onaudio.herokuapp.com 
23rd October 2012, 08:18 PM  #282  
diyAudio Member
Join Date: May 2012

Quote:


7th December 2012, 01:56 PM  #283 
diyAudio Member
Join Date: Dec 2007
Location: Italy

Yer right! We want to use discretes
__________________
"The total harmonic distortion is not a measure of the degree of distastefulness to the listener and it is recommended that its use should be discontinued." D. Masa, 1938 
12th August 2014, 11:01 AM  #284 
diyAudio Member
Join Date: Mar 2012

The book is very well written and provides essential information on the subject. Have read articles by R. Elliott and the book by Mr. Self completes the picture on active XO nicely. Thanks a lot for the comprehensive work !

20th September 2015, 10:21 AM  #285 
diyAudio Member
Join Date: Feb 2008

Phase oh the sum for LR4

28th September 2015, 07:11 PM  #286 
diyAudio Member
Join Date: Jul 2009
Location: Milano

I own all your books, since I'm a pro audio designer, they are usefull to me. I've already used come of your solution in design of mine.
Mik 
16th April 2016, 02:34 AM  #287 
diyAudio Member
Join Date: Jul 2006

I am working on a piece of code to answer the question if we allow the resistors to differ "somewhat" in a SK filter, can we arrive on a more accurate set of resistors to build an adjustable crossover with. Also, what sort of tradeoffs are there between various resistor ranges (e96 versus e24).
My design is fixed gain, using resistors that are "approximately" a ratio of 1 and capacitors that are a ratio of 2, each with 3 choices (100nF, 10nF, 1nF, 47nF, 4.7nF, .47nF). This results in a target Q = 0.707. I'm also assuming I have a list of target frequencies (I have ten values, then I multiply by 10 to cover 25Hz to 20kHz). Here's an example of what my code has found. t: 0.1%, f error range: 0.015% 1.296%, Q error range: 2.578% 3.542%, number of unique resistor values used (e96): 10 t: 6.1%, f error range: 0.015% 0.443%, Q error range: 2.571% 3.535%, number of unique resistor values used (e96): 18 So, t is the percentage that two resistors are allowed to vary from each other (with the constraint that the resulting resistors are available in the series (e96) specified), then a range of frequency errors (from target) and a range of Q errors (from target) and the total number of different resistor values needed to produce the combinations. So, my question is  is there any thing else I should be interested in considering, during these "experiments"? How much variance in Q is acceptable, at the expense of a more accurate frequency value? Are there other unintended consequences of having the resistor values "differ somewhat"? (Obviously as the two values diverge, it impacts Q, so there's a very clear limit to what combinations should be considered) 
16th April 2016, 10:05 AM  #288 
diyAudio Member
Join Date: Jul 2004
Location: Scottish Borders

the errors in capacitor value are not mentioned.
__________________
regards Andrew T. 
16th April 2016, 01:03 PM  #289 
diyAudio Member
Join Date: Jul 2006

That could be easily modeled. As it turns out, the easiest way to deal with this might be adding a fourth capacitor at yet one more power of 10 down. This could then be placed in parallel with the other values as needed. Once actual values were chosen, then they could be measured and then a chart could be made on when exactly to engage more than one capacitor, including this fourth "adjustment" capacitor. I could also run the simulation with a random element. The capacitors I plan on using is a type not mentioned in the book (but I would be curious if perhaps Doug would measure some!)  Rifa/EVOX MKI which are metalized polypropylene sulfide capacitors which are +/ 10%, which could be matched. For some values I would use EXFS polystyrene made by LCR components are +/ 1%. 
19th May 2016, 02:50 AM  #290 
diyAudio Member
Join Date: Jul 2006

Similarly, for the adjustable HPF. To maintain unity gain and a target Q of .707, we need resistors in a ratio of 2. But what if we let them vary slightly? Obviously the more the resistors vary, the further from our target Q we will be. However, we can get closer accuracy to a given frequency target, if there is some wiggle room.
In this output, t is the amount the two values can differ from the ideal ratio of 2:1 as a percentage of the second resistor. So, 0.001 means the values vary from 2:.99 to 2:1.01, subject to the constraint of a valid resistor in the series. Things converge nicely around about 5% allowable variance. t: 0.001, f error range: 0.014% 60.014%, Q error range: 0.000% 0.113%, unique resistors (e24): 9 t: 0.001, f error range: 0.017% 34.032%, Q error range: 0.000% 0.113%, unique resistors (e96): 18 t: 0.006, f error range: 0.014% 60.014%, Q error range: 0.000% 0.113%, unique resistors (e24): 9 t: 0.006, f error range: 0.017% 1.010%, Q error range: 0.000% 0.266%, unique resistors (e96): 15 t: 0.011, f error range: 0.014% 60.014%, Q error range: 0.000% 0.551%, unique resistors (e24): 8 t: 0.011, f error range: 0.017% 1.010%, Q error range: 0.000% 0.266%, unique resistors (e96): 15 t: 0.016, f error range: 0.014% 60.014%, Q error range: 0.000% 0.551%, unique resistors (e24): 8 t: 0.016, f error range: 0.017% 1.010%, Q error range: 0.000% 0.266%, unique resistors (e96): 15 t: 0.021, f error range: 0.014% 32.472%, Q error range: 0.000% 1.159%, unique resistors (e24): 13 t: 0.021, f error range: 0.017% 0.906%, Q error range: 0.000% 1.055%, unique resistors (e96): 14 t: 0.026, f error range: 0.014% 32.472%, Q error range: 0.000% 1.255%, unique resistors (e24): 14 t: 0.026, f error range: 0.017% 0.517%, Q error range: 0.000% 1.196%, unique resistors (e96): 14 t: 0.031, f error range: 0.014% 15.590%, Q error range: 0.000% 1.482%, unique resistors (e24): 16 t: 0.031, f error range: 0.017% 0.517%, Q error range: 0.000% 1.338%, unique resistors (e96): 15 t: 0.036, f error range: 0.014% 7.742%, Q error range: 0.000% 1.626%, unique resistors (e24): 15 t: 0.036, f error range: 0.017% 0.517%, Q error range: 0.000% 1.338%, unique resistors (e96): 15 t: 0.041, f error range: 0.003% 6.217%, Q error range: 0.000% 1.869%, unique resistors (e24): 16 t: 0.041, f error range: 0.017% 0.517%, Q error range: 0.000% 1.976%, unique resistors (e96): 17 t: 0.046, f error range: 0.003% 5.828%, Q error range: 0.000% 2.342%, unique resistors (e24): 14 t: 0.046, f error range: 0.003% 0.517%, Q error range: 0.000% 2.283%, unique resistors (e96): 18 
Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Douglas Self's Active Crossovers book  forr  Analog Line Level  8  31st December 2010 05:20 PM 
active crossovers  zimjj  MultiWay  9  7th November 2010 01:58 AM 
Active Crossovers  zenmasterbrian  Subwoofers  77  21st September 2006 05:14 PM 
New To Site?  Need Help? 