No doubt 2-ways are the most popular design, but they're also quite difficult to do well because you have to get significant overlap between the woofer and tweeter. Since they're both rolling off near there, the crossover design is never straightforward. Though some like a low order crossover, that gives you a power handling problem with the tweeter. You also want to keep signal out of the woofer in the breakup area. No matter what you use, the acoustic response of the drivers is a major portion of the design.
Here's the schematic for Klipsch KG 4.5:
An externally hosted image should be here but it was not working when we last tested it.
.
LF: single 10" woofer, ported cabinet.
HF: Tractrix horn, 7.5" width x 4" height.
That RDE 070 is a polyswitch resettable fuse with internal resistance spec'd at 0.5 ohm.
--
SPECIFICATIONS
DIMENSIONS: 37.5" (95.3cm) x 12.25" (31.1cm) x 12.75" (32.4cm)
ENCLOSURE MATERIAL: Medium density fiberboard construction (MDF)
ENCLOSURE TYPE: Bass reflex via front-mounted port
FREQUENCY RESPONSE: 36Hz-20kHz(+-)3dB
HIGH FREQUENCY HORN: 90(o)x60(o) Tractrix(r) Horn
NOMINAL IMPEDANCE: 8 ohms
POWER HANDLING: 100 watts maximum continuous (500 watts peak)
SENSITIVITY: 95dB @ 1watt/1meter ------- [More like 93dB/1w/1m from what I can tell]
TWEETER: K-85-K 1" (2.54cm) Polymer dome compression driver
WEIGHT: 54 lbs. (24.5kg)
WOOFER: K-1005-K 10" (25.4cm) Injected Carbon Graphite cone
--
Last edited:
http://forums.klipsch.com/forums/storage/74/1311651/Klipsch%20kg%234.jpgI had picture searched Klipsch KG-4 apparently the .5 got dropped in the shuffle
my thoughts about crossing that low still remains
essentially this is a a psuedo 1st order crossover with a Zobel on both drivers more agressively done on the woofer to tame any cone breakup. were they cross exactly depends mostly on the acoustic responses.
You could run a signal generator around 1-2 KHz and let a microphone tell you.
my thoughts about crossing that low still remains
essentially this is a a psuedo 1st order crossover with a Zobel on both drivers more agressively done on the woofer to tame any cone breakup. were they cross exactly depends mostly on the acoustic responses.
You could run a signal generator around 1-2 KHz and let a microphone tell you.
Last edited:
Thanks! Silly me, I had looked at that crossover and thought it was 2nd order electrical for both LF and HF but couldn't figure out what the 2 ohm resistor in the LF and the 12 ohm resistor in the HF were. But if those form Zobel networks, then you have an electrical (only) 1st order network (series L on the LF, series C on the HF). Interesting.
I only feed the speakers from a 6wpc PP DHT amp. I'd imagine that if I tried to push these speakers to their supposed 100+ watt max that I'd get some truly nasty sounds around the crossover point. Fortunately, I'm not allowed to push these speakers that hard, being an apartment dweller and all...
--
I wouldn't think of them as Zobels, especially the horn network. A Zobel will not go in the middle of the network (with a C before and after), plus it is only a Zobel if it truly flattens impedance, something that we don't know.
It really is a damped third order. The simulation higher in the thread looks very plausible. The long rise from a low frequency would be just right for equalizing the horn and its naturally falling response.
David S.
a Zobel is usually a network is that provides a cancellation of the opposite (dual) reactance. A Horn loaded compression drivers impedance is a double humped beast similar to a woofer in a bass reflex, so many good crossovers will target the second hump cause it's closest to the tweeters BW of operation. Maybe it's more accurate to say conjugate impedance match.
Last edited:
But again, that is not what that part of the HF network is doing. It is a third order network with a Q softenng resistor added. Same with the woofer, I use the R in the shunt leg a great majority of the time. It allows you to Use a fairly large shunt capacitor to get the shape you want without having a "Q bump". The clue is the resistor value, 2 ohms is too low for a Zobel since the voice coil DCR will be higher. Can't count the number of times I've put a 1 or 2 ohm resistor in that position.
Zobels are not just a matter of topology but of value. If it is specifically added to flatten impednace and values are so adjusted, then it is a Zobel. In the case of these networks they are auxiliary components used to effect the filter shape, not Zobels.
David S.
I'm talking about the HF network.
The LF networks resistor is usually done to tame the woofers increased response due to cone break up, the resultant order is 1 within the pass band.
I like to call this technique an aggressive Zobel , sadly the woofers rising impedance is not due to a single pure inductance value.
The LF networks resistor is usually done to tame the woofers increased response due to cone break up, the resultant order is 1 within the pass band.
I like to call this technique an aggressive Zobel , sadly the woofers rising impedance is not due to a single pure inductance value.
Last edited:
But the HF network has nothing that looks at all like a Zobel. It is clearly a third order high pass that happens to have a Q modifying resistor in one leg.
Perhaps pedantic, but the woofer shouldn't be called a Zobel either, unless we know for sure that it had been adjusted to flatten impedance. If the R value drops from 2 to 0 ohms is it a really, really agressive Zobel? (or just a capacitor?)
David S.
Last edited:
Degenerate 2nd order. It becomes first order some distance into the stop band.
David S.
I should know better than to post anything while I'm distracted with other things... 😉
I could see the idea behind the LF section being a 1st order network with Zobel added instead of seeing a 2nd order filter, whether or not that's what's actually happening there. But the HF network does have that series-C, parallel-L, series-C circuit. Definitely 3rd order. Oops.
Thanks for posting the conversation. I do learn a lot from lurking around these exchanges...
--
I could see the idea behind the LF section being a 1st order network with Zobel added instead of seeing a 2nd order filter, whether or not that's what's actually happening there. But the HF network does have that series-C, parallel-L, series-C circuit. Definitely 3rd order. Oops.
Thanks for posting the conversation. I do learn a lot from lurking around these exchanges...
--
- Status
- Not open for further replies.