Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

LM3875 and speaker impedance questions

Status
Not open for further replies.
Nigel Goodwin said:
however, why don't you upload them here rather than providing external links?.

it's a habit from other forums that they don't allow file uploading...
you want to avoid broken links right? i'll try to upload them here in the future.
 
hi the graphs look better now:D the slope is where the amplitude drops with increasing freq in a low pass filter (for the woofer) and when the amplitude drops with decreasing freq in a high pass filter (for the tweeter). Its the steep part of the graph in each case.
Have u checked the coil value? it seems my guess was far off :D from the looks of the response and the capacitor value (3.3u if u haven't changed it) i would say its between 0.4m and 0.6m :D.
 
hentai said:
Have u checked the coil value? it seems my guess was far off :D from the looks of the response and the capacitor value (3.3u if u haven't changed it) i would say its between 0.4m and 0.6m :D.

actually no.. :D i was waiting for you to explain what is slope...

if i am correct the slope is exactly at 2000Hz for the woofer and
and about 5000Hz for the tweeter...maybe? :confused:

I don't know if i should use this calculation "L=2.54/F" or the first one.
The photos in the first pages aren't helpful to understand which type of crossover i have?

so......what's next?! :)
 
Last edited:
the slope starts at those freq. Set the cursor at 2000 and the next cursor point at 4000 Read the diference in level it should be 12 dB for a second order filter (12db/octave). A closer look indicates that ur filter is a second order LR at 3kHz which means ur coil is about 0.8mH 0.85mH. See at what frequency the level is 6dB lower than the linear region (300 -1000 Hz). If it is around 3kHz then ur coil is likely to be 0.85mH.
 
if i zoom more i see that the slope starts at 1440Hz
anyway...too much detail probably! :D

from 1440Hz to 4015Hz it's -12,2dB
the program has 3 small windows which one i look? the second says "1,48oct -4,49dB/oct"

from 2018Hz to 4015Hz it's -12,1dB (0,99oct -6,21dB/oct)

i think that 6dB lower than the flat region (350-1000Hz) it's at 3700Hz :rolleyes:
 
Last edited:
look at the xx dB/oct and look after 3700Hz this time.
well my guess is that its 2nd order LR and ur coil should be about 0.85mH. Between 0.4mH and 0.9mH anyway hahaha. Well if u like the sound with those filters the next improvement is to add a Zobel network for the problematic woofer when u'll have the parts available. U said u have some 1st order filters? How many? and do u know the components' values on that one?
 
so...i can't cut my coils?
remind me what exactly improves the zobel network?
you suggest that i keep those 8ohm crossovers with the 4ohm speakers?
with those 3.3uF capacitors?

as i said the sound is more balanced from the speaker which has the 2nd order crossover...i think it's not my idea...the sound is indeed better by the time i reversed tweeter polarity.

the 1st order crossovers also have 3.3uF capacitor...but at the moment i don't have them both..


P.S the capacitors are at 100V and the resistor which is probably for protecting the tweeter...is at 5W how much power can withstand these crossovers? :rolleyes:

how did you found out that they are a 2nd Order Linkwitz Riley crossovers? by the mH of the coil?
 

Attachments

  • bass.JPG
    bass.JPG
    43.3 KB · Views: 200
Last edited:
Most comercial xovers are either LR or butterworth and are usually cut around 3kHz. Seeing that u have a 3.3u I figured it had to be a LR, butterworth uses bigger values for the capacitors. The fact that u found -6db at 3.7kHz further more confirmed that. If it would've been butterworth it should've had -3db around 3kHz...
If u want to addapt ur xover for ur drivers impedance u should cut the coils in half (the wire lenght i mean) and double the capacitors...
The zobel makes ur woofers impedance linear so at the xover freq the woofers impedance will be around 4 ohms not bigger. This will give u more balance because without it the xover point for the woofer wont be at 3k but at bigger freq.
It should withstand ur gainclone.
 
hentai said:
If u want to addapt ur xover for ur drivers impedance u should cut the coils in half (the wire lenght i mean) and double the capacitors...


i believe you don't encourage me to cut my coils... and you probably advice me to leave them like they are and just add a zobel network ouside the crossover in parallel with the woofer plus i change the tweeter polarity right?

from my side i don't have any view for the subject because i'm still learning basic things...yesterday i finally understood what is "corner frequency" you helped me without knowing that you helped! :D

anyway...i have to assemble back things... the one crossover is messed up
totaly unsoldered! :D but there is something i don't like.. it's the crossovers that still
are 8ohm and i don't know exactly what problems they cause.

instruct what to do...because as you understand a newbie easily gets confused! ;)
 
Last edited:
Dear friend, for ur further understanding of the intricacies of xover design (I just learned this word and i was dying to use it :D ), i will reveal to u the secret formulae for a 2nd order filter: the inductance is
L=Z/(2*pi*f*Q), the capacitance is C=Q/(2*pi*Z*f), where Z is the drivers impedance, f the xover cut frequency, Q is related to damping and is 0.707 for Butterworth (wich gives -3db at f) and aprox 0.5 for LR (which gives -6db at f). As u can see from this lowering Z by half will also lower f by half increasing Z also increases f.
The reason i dont encourage u to cut the coils is that u cant tell for sure where is the half of the wire's lenght. u could cut in steps and measure the response after each cut.
Lets see what we've learned until now. We have a xover that cuts at 3k (aprox) 2nd order at 8 ohms. With ur woofers impedance at that freq being bigger than 8 ohms its fair to say it will cut higher. However the tw impedance is quite linear so the tw is cut about 1.5k-2k which is 2 low. From the pics u posted it seems much easier to modify the tw coil so do it. cut the tw coil in half and double the capacitor. BUT I remember there was a 4 ohm resistor between the filter and the tw if this is so leave it right there for it will make the total Z around 8 ohms so the above modifications wont be necesary. Also that resistor provides a 3db loss in Spl also required for the tw is more sensitive than the woofer. So the tw part is done. Now for the woofer u should put a Zobel network so that the woofers impedance will result in 8 ohms. Yes 8 ohms so that the filter cut point will be at 3-4k same as the tw.
So the modifications u should do the simplest way i can think of:
for the tw if the 4 ohm resistor is still there leave it if not put one
for the woofer put a 8.2 ohm resistor mb even a 10 ohm resistor in series with a 33u capacitor, all this connected in parallel with the woofer so thats between the xover and the woofer.
Give it a try.
 
A 4 ohms resistor in series with the 4 ohms tweeter reduces the level in the tweeter 6dB, not 3dB.
 
yes indeed, my mistake. well then the attenuation might be unpleasent. Whiz should try it though if he doesnt like it then modifying the tw filter is the solution.
 
Last edited:
as you can see (both of you...) :D in the very first photos it says very clear
that the resistor already exists in series with the tweeter and it is ~8ohms...

ok guyz! i think "the Quest for better sound" somewhere here comes to an end!
Hentai your contribution is unbelievable unfortunately the only payment you can expect from helping people in forums...is a "Thank you!" probably and a pleasure for helping! or maybe a future help from those you helped! :D

i won't cut the poor crossovers, i'll re-assemble them to their boxes as they where before and i'll reverse the polarity of the tweeters plus i'll put a zobel network. (it's for the woofers only...right?)

i really hope that things will be allot better with that zobel network thing...
if not, i'll take those poor little crossovers and i'll through them out of the window together with the speakers and the zobel network! :mad: :D

maybe i'll post some results one of these days...
 
Last edited:
give up so easily huh? :D
whiz115 said:
the only payment you can expect from helping people in forums...is a "Thank you!"
never asked for more!

yes the zobel is for the woofer. u dont need to bother with the tw's impedance egalization. The 8ohm resistor is way too much put a 4 or 2.2 !
Well good luck and one more thing never throw speakers! (it may leave u scared for life :D ).
Keep us posted with ur progress ;)
 
**BREAKING NEWS!**

update...

i've made the following changes

1) i replaced the ~8ohm tweeter resistors with 2.2ohm as you instructed me

2) i reversed both tweeters polarity

3) i connected in parallel with the woofer a zobel network 8ohm/33uF


i) using an audio processing utility i created some tones to test speaker
response...at 10000Hz the tweeter reproduced the frequency and after 2 seconds it started to make strange noises like if i was scratching it. when i replaced the 2.2ohm resistor with the ~8ohm everything went back to normal.

ii) i think it makes a difference the highs are more balanced with the lows...

iii) i have connected a cable on the woofer connector and i took it out from the bass reflex hole so i can test zobel network while i was listening various songs. The zobel network 7ohm/33mf cuts voice detail and makes it blur.


P.S. checking old posts i've noticed one more misunderstanding... my amp's
power supply is +/-22V AC after rectification it is ~30V DC! :rolleyes: now do you think that 4ohm speakers affect my amp's performance?


hentai said:
give up so easily huh? :D

please tell me that you aren't the one who gives up! :p
 
whiz115 said:
**BREAKING NEWS!**
P.S. checking old posts i've noticed one more misunderstanding... my amp's
power supply is +/-22V AC after rectification it is ~30V DC! :rolleyes: now do you think that 4ohm speakers affect my amp's performance?
This thread is going around in circles.
Is the supply plus and minus 30VDC or is its total voltage only 30VDC?
The LM3875 amp will make strange noises as it protects itself from the high current into a 4 ohm speaker with such a high supply voltage as 60VDC. Keep the volume away from max output.
 
hey
no 1 issue is very strange... as audioguru said was ur amp at full output? and did u noticed a change in loudness between the 2 resistors? When u apply program do u still get the scratch?
no 2 is settled
no3 if u connected the zobel and the tw had the 8 ohm resistor in front of it its normal to loose detail in mid range for there will be a gap between the woofer's xo point and the tw's xo point. Try the zobel with the tw having the 2.2 resistor
 
audioguru said:
This thread is going around in circles.

yes that is true, but you know what? i don't like it either...
some things need very precise explanation, but neither all questions
were very specific...i don't want to blame someone with that! especially someone who tries to help me! like you and Hentai but do you thing it's all my fault?

audioguru said:
Is the supply plus and minus 30VDC or is its total voltage only 30VDC?

as i said the supply voltage is 22V AC (symmetrical supply) and after rectification
1.41x22=31.2V DC my knowledge and my English don't help me to explain it farther sorry...
 
Last edited:
hentai said:
hey
no 1 issue is very strange... as audioguru said was ur amp at full output? and did u noticed a change in loudness between the 2 resistors? When u apply program do u still get the scratch?

hey...so no one gives up? :D
yes it was at full output (to the point no serious distortion occurs when you listen to the music)...no i'm not sure about the loudness i've only noticed the specific problem...no i haven't noticed anything wrong from the other speaker while playing songs...also i haven't noticed any difference with the ~8ohm resistor but the volume was at the same level as wheni tested the crossover with the 10KHz tone.

hentai said:
no3 if u connected the zobel and the tw had the 8 ohm resistor in front of it its normal to loose detail in mid range for there will be a gap between the woofer's xo point and the tw's xo point. Try the zobel with the tw having the 2.2 resistor

i tryed it with the 2.2ohm resistors...(i also have some 4ohm resistors which i'm about to test)
 
ofc we wont give up ( i wont at least) we've got 2 far. The thing u have to consider though is that ur speakers are not top notch so even with a good xover u won be able to get ur speakers to sound like a B&W, u know what i mean? So what i'm trying to do here is to try a few tricks to obtain a coloration from the xover so that u find ur sound more pleasing. It's trial and error at this point but this is how experience builds up.

no 1. the thing that u didnt perceived a difference in sound level makes me think that is ur amps fault for that noise and that it was in clipping. try at low levels does the sound still exist?
no 2. try to increase the resistor value from the zobel until the mid range seems ok for u, after u have tried with the 4ohm tw resistor.

Two more things i want to know, where do u have this xover from? The speakers are store products or custom made?
 
Status
Not open for further replies.

Latest threads

Back
Top