Mix and match 4 and 8 ohm speakers

[Hyun]

what would be the result if u run 2 speakers, originally both 4 ohms coming out of a 2-way passive crossover, but now you swop one of them to become 8 ohm?

 

since it's a crossover, i expect that at the lower frequences, let's say the 8 ohm one is playing, the amp will be facing an 8 ohm load. and at the higher freqs, let's say that's the 4 ohm one, the amp will be facing an 4 ohm load.

 

that means the response at these 2 freq ranges would be difference since the amp load is different. the amp output at 8 ohms is about half of 4 ohms. however, 8 ohm drivers are said to be "more efficient". and the speaker sensitivity is also important. but what will happen in such a combo? let's assume sensitivity to be EQUAL for both speakers.

 

will the amp output less automatically when faced with 8 ohms, so since sensitivity is the same, the dB output will be half the power, so that's about 3 dB less. so it's a softer output. then when playing higher freqs facing with 4 ohms, the output suddenly increases, and since sensitivity is the same, it sounds 3dB louder for higher freqs?

 

or is it standard procedure for such combos to require impedance matching, so it requires both speakers to be matched with something like an L-pad to convert all into 8 ohms? that way it effectively "attenuates" the 4-ohm speaker by half (to prevent the 3dB louder effect) - using a 4 ohm resistor in series, and a 4 ohm one in parallel across the 4 ohm speaker for attenuation.

 

what is the standard procedure?

[Lionheart7004]

use a multi-meter.. and you will know the actual impedance.

[Ractis]

If sensitivities are the same then the 4ohm will get double the power of the 8ohm and be 3db louder. Will need some attentuation like you said. In general it's not a good idea to use a crossover network designed for a particular component set for another, especially if the impedance is different.

 

You are probably familiar with martin electronics? Maybe see if he is willing to give you some tips on crossovers..

[Hyun]

lionheart:

the unknown isn't the impedance - i got the values of nominal impedance and DC resistance etc from the published sites. the question is how it would sound like together when one is low-pass and one is high-pass, compared to using matching 4ohm-4ohm speakers.

 

ractis:

hmm that's my fear - the 4 ohms draws more power get gets louder, and needs attenuation. what i plan to implement is similar in concept but different in values:

 

at 50-200hz 2 speakers both 4 ohms are playing (that's effectively 2 ohms, being in parallel). beyond 200hz just 1 4-ohm will play. so in that band 50-200hz it'll be about 3db louder assuming flat response.

 

good thing is most speakers dun play 50-100hz very loud, so who knows, this might just balance the rolloff!

[Lionheart7004]

Post your question to the following.

 

http://www.diymobileaudio.com/forum/index.php?

 

I saw a writeup similar to your concerns.

[Hyun]

hehe i already did that. the one u saw might be my post! heh. this is called KIASU SINGAPORE!

[Lionheart7004]

Not recent.. 1 yr back posting.. Please do a search.

[Ractis]

oic, issit you bought another 4ohm midbass driver and want to parallel to using the same amp that is currently driving your passive components?

 

In that case no need LPad, just low pass the new driver, i suggest lower than 200hz, maybe 100hz-120hz might be better? Have to see where your passive start to rolloff in your car

 

If the price of good caps is a concern, you can still use electrolyte and parallel it with better ones of smaller values for a compromise in quality

 

edit: you intend to run the wires from the new speakers back to the amp right? this is preferable so that you wont have to mess with the crossover, which is already tuned to the passive components. If you are thinking of connecting the new speakers to the crossover outs, then its complicated...

Edited August 25, 2006 by Ractis

[Hyun]

i was advised to bypass caps as well from another forum. this is the explanation:

 

The idea is that they do a better job of passing very high frequencies due to a lower effective series inductance and other effects that appear at those frequencies. So usually they're recommended for the tweeter branch of the passive crossover. Personally, I don't usually bother.

 

and these bypass are very small in value so as not to alter the net (summation of parallel caps) capacitance. well i got 4 of these bypasses at 10nF lying at home. time to put them to use, heh. but they'll probably end up in my tweeter section since the advice was more for tweeters. i'll just use the lousy caps for the 200hz LPF.

 

actually my final plan is to just do a one-sided LPF about 200hz or 160hz on the midbass, at about -6 or -12db/oct and let it roll off to match the midrange rolloff below 200hz. so the midrange wont have a HPF. (the midrange was the one i bought.)

[Hyun]

to feedback on the effect of my experiment:

 

i ran a midrange from 50hz to 3khz -12slope. i ran the midbass from 50hz to 200hz -6slope. there is overlap at the entire midbass range. at that point due to 2 speakers being active and in parallel, despite the fact that the impedance of the midrange rose by nature of its Fs=84hz, the system sounded high output at that region. the midrange wasn't playing that much already since it was naturally rolling off below 200hz. but the midbass was disproportionately louder.

 

so perhaps the two 4ohms in parallel here net a 2ohm+ load which caused an increase in output. my amp is supposed to be stable output across 1 to 4 ohms maybe by voltage switching but i guess it can't react in a similar manner when it's a variable 2 to 4 ohm load depending on frequency played rather than a fixed 2 ohm or 4 ohm throughout.

 

result after implementing LPF of midbass at 200hz -6 slope, and HPF of midrange at 380hz -6 slope, it was much more natural. EQ ended up rather flattish across that range, but a bit boosting the midrange probably due to the lower-sensitivy of the midrange by about 2-3 db/W/m.

 

moral of the story? hehe dun play stunts, ya?

[Ractis]

What you really need is a active crossover.. you can use it to find your ideal crossover points and slopes, then go ahead and build the passive network.

 

Tweaking from the HU to find the optimum points is already so difficult, I cannot imagine anyone tweaking the passive xover network :P

[Hyun]

i actually have one, and the name's quite impressive leh - ZEUS active cross (2-way) but when i searched the net, only got GOD called zeus leh, no active cross. this one my fren pass me one.

 

so partly i didn't dare to use, cos i dun even know the type of cross employed, and the cross point dial is one of those analogue turn dial, the "hope u turn to the right number" wheel of fortune. and as all active crosses go, it's at the pre-amp level so i need to amplify (and i dun have enough amp channels).

 

that's why i ended up taking this route.

i'm surprised my combo of 200hz LPF for midbass and 390 HPF for midrange with both -6 slope works quite nicely, without feeling any significant gap in between. quite a reasonable transition.

[Kwongwing]

In theory what you suggested would seem to work - but in practice, the active x-over settings do seem to differ quite a lot from any passive x-over settings!!!

Anyway Hyun is experimenting and seems to like taking the difficult route

But believe that he is finally reaching his initial goal, and would soon be shifting the goal posts again

[Hyun]

dun say until like that leh ... tempting me to change tweeter ah?

 

no lah currently got enough concerns to keep me busy. planning a DIY fibreglass enclosure for midrange and tweet, possibly at A-pillar or nearby.

 

and haven't buy my aircon foam to form a focus cone for my midbass yet.