Thankfully, just about any of the drivers I sort of want are available - you know, the W100/130/170/200S 4 or 8 ohm. I had a chat with the sales person at PQ Sound. Kit's are just too expensive, so that's why I'm avoiding them. Weird how I can buy the same stuff for a kit, but it works out cheaper if I just buy the drivers and make the crossovers.
A friend of mine has a pair of F15s, but I don't think I'll pry them from his tight grip. They're fantastic!
My main hobby and passion is designing and building amplifiers. I'm an audiophile in the sense that I listen to my sound with music, and furthermore a keen DIYer in the sense that I listen to my amplifiers with music. (As opposed to listening to music with my equipment).
I will definitely get into those F20's soon. I just haven't had the chance yet. They're particularly good for classical music. I've attended a few orchestral performances, and the F20's provide a very realistic reproduction. But I listen far more to metal and rock, and the F20's aren't the best for voices or guitars. Hence why I'm so keen on a strong, detailed midrange (but not quite as overpowering as Klipsch R-26F or MA PL500 II). Something more like the Kef IQ9's. I complemented my F20's with a DIY sub, which works extremely well to get those very low notes out. Ideally, I want two speakers to do everything without a sub.
I have those two Philips 9710M's that I want to add to the Celestions to see if they can improve things a little. I also have a pair of Pioneer speakers (salvaged them from an all-in-one system), and they actually perform well, so I've connected them as rears for when I do watch movies in 4.1.
And then, a long shot, how do you think these would do in a hi-fi setup? They're certainly loud, but is it possible to use them for, say, 200-4kHz? I have three, two are in guitar cabs, and the third is about to go into another one. They're mind-blowingly awesome as guitar speakers!
Celestion G12H-75 Creamback
Celestion G12H-75 Creamback
We are ranging far and wide here. PA speakers are a whole different ball-game, IMO.
I do think a good speaker is built around the midrange. It's what the three way theory is saying:
FWIW, I thought the weakness of the 8"/4"/1" idea was it pushed the mid quite hard. That's what the SPL limits were saying.
So the Wharfedale E70 is a good idea:
Technology is getting silly these days. This is DSP filtering with Class D amps.
Don't let the cat or children near it!
I do think a good speaker is built around the midrange. It's what the three way theory is saying:
FWIW, I thought the weakness of the 8"/4"/1" idea was it pushed the mid quite hard. That's what the SPL limits were saying.
So the Wharfedale E70 is a good idea:
Technology is getting silly these days. This is DSP filtering with Class D amps.
Don't let the cat or children near it!
Of course! You mentioned that E70 idea (probably in another thread I read). It completely slipped my mind. Ok, tonight when the children are asleep...
I'm not keen on DSP filtering / pre-filtering and class D. I've done that, and it's just not right for me. Nothing in this world has sounded better than a pair of B&W 802 D3's, driven simply by monoblocks with no DSP.
I think we are making some progress here. If you can get the drivers for about R4000, which I make to be about £260, that's not bad at all.
I'm slightly intimidated by the prospect of competing with a £22,000 B&W 802 D3 speaker that weighs a hundredweight, or a 1/8 of a ton!
Bowers & Wilkins 802 D3 Diamond loudspeaker | Stereophile.com
I'm the World's worst for rambling off-topic, but I was struck by how similar the Visaton Starlet is to the Wharfedale E50. A WMT design below. Or WMMT in the E70 crossover below.
And here's a WMTMW design by those enthusiastic forum amateurs at Visaton:
https://boxsim-db.de/experience-v20/
TBH, Marc, I need to sleep on this, for about a month.
And I wouldn't touch this cone TW 6 NG tweeter with a bargepole:
TW 6 NG - 8 Ohm
What I would say, is that you'll need to do some serious modelling and line up your crossover parts. This sort of project is a huge labour of love and enthusiasm. Nobody builds a good speaker in 2 days, because there is far more to it than frequency response.
BW1, LR2, BW3, LR4, room gain. Lobing. Power response. Loudness without distortion. Flat and easy impedance.
I regularly input to a thread called Classic Monitor designs, because I am interested in what designs seem to work well. Even simple ones. Seems like there are some good ideas in loudspeakers. And, IMO, they are the most important part of the chain.
Here's a guy I admire, Roy Allison. He thought about EVERYTHING, but especially the room. His speakers were designed to mount close to the wall or nearest boundary. But the main thing is to have good ideas, IMO.
I'm slightly intimidated by the prospect of competing with a £22,000 B&W 802 D3 speaker that weighs a hundredweight, or a 1/8 of a ton!
Bowers & Wilkins 802 D3 Diamond loudspeaker | Stereophile.com
I'm the World's worst for rambling off-topic, but I was struck by how similar the Visaton Starlet is to the Wharfedale E50. A WMT design below. Or WMMT in the E70 crossover below.
And here's a WMTMW design by those enthusiastic forum amateurs at Visaton:
https://boxsim-db.de/experience-v20/
TBH, Marc, I need to sleep on this, for about a month.
And I wouldn't touch this cone TW 6 NG tweeter with a bargepole:
TW 6 NG - 8 Ohm
What I would say, is that you'll need to do some serious modelling and line up your crossover parts. This sort of project is a huge labour of love and enthusiasm. Nobody builds a good speaker in 2 days, because there is far more to it than frequency response.
BW1, LR2, BW3, LR4, room gain. Lobing. Power response. Loudness without distortion. Flat and easy impedance.
I regularly input to a thread called Classic Monitor designs, because I am interested in what designs seem to work well. Even simple ones. Seems like there are some good ideas in loudspeakers. And, IMO, they are the most important part of the chain.
Here's a guy I admire, Roy Allison. He thought about EVERYTHING, but especially the room. His speakers were designed to mount close to the wall or nearest boundary. But the main thing is to have good ideas, IMO.
Attachments
Lojzek, thanks. I like it, but the only thing is the lack of midrange. I think 2 W100S's will be needed. Based on your suggestion of the WWMT, and combining that with the E70, I came up with that WMTMW. It's much like the Experience V20 Steve linked to, but it uses 2 W170S's. I also did the MTM configuration similar to the E70.
Steve, the cone tweeter isn't ideal. I've put the SC10N in, but I put that cone tweeter there in the hope you would comment on it I wasn't sure if it was a good thing or not.
And yes, this is no rush job. I'm mainly looking for direction. I think I've found that I really like the 2xW170S, 2xW100S and 1xSC10N combo, and I can refine that in time.
Thanks for all the help and advice so far Lojzek and Steve! I've taken a few steps forward, especially in learning. Hopefully I can eventually turn this into a build thread.
Steve, the cone tweeter isn't ideal. I've put the SC10N in, but I put that cone tweeter there in the hope you would comment on it I wasn't sure if it was a good thing or not.
And yes, this is no rush job. I'm mainly looking for direction. I think I've found that I really like the 2xW170S, 2xW100S and 1xSC10N combo, and I can refine that in time.
Thanks for all the help and advice so far Lojzek and Steve! I've taken a few steps forward, especially in learning. Hopefully I can eventually turn this into a build thread.
No problem, you are welcome to do your own modelling but in regard to midrange SPL there is no need to introduce a second midrange and then pad it down, and doing it in a D'Appolito fashion causes further off-axis cancelations. The small response hole made in crossover region was on purpose to smooth out the horizontal off-axis curves but it's not definite what one might prefer better. That's how professional designers do things.
I have a buddy who owns Experience V20 and according to him this kit sounds rather poor considering his initial expectations and when you observe the project in Boxsim the graphs on and off-axis look pretty odd. No doubt, the XO filter needs improving.
Cheers!
My reasoning was doubling up W100S drivers in series was going to improve power handling, Lojzek. Thus less distortion. The midrange is the main event IMO. And didn't someone suggest that lower quality midranges can work well a while back. Just coils and cones and magnets in the end.
The physics says that each driver is only outputting a 1/4 of the SPL for the same overall level. Thus lower distortion. I also quite like Roy Allison's idea of putting the (single) bass near the floor below the bafflestep cutoff of 500Hz on a 22 cm baffle. You then effectively double up the bass response again with the floor gain. It's clever.
If it was me, with all the supply issues in SA, and availability of coils is the biggest one it seems. I'd be looking into availability of enamelled copper wire and wind them round a bobbin to correct inductance value.
This is just a hobby, in the end. To build a speaker, I have to be excited about the idea being tested. I liked everything about this DeVore Gibbon 88 clone idea:
Add the BBC impedance and passive RL bafflestep, and it all falls into place. Knowing Boxsim a bit, it overdoes the bass response when designed flat, but this 22cm wide, 32L design is a peach IMO. The W170S is actually more of a closed box driver, IMO, so a sock in the port might be good too. Joachim Gerhard used to talk vaguely about his diffraction module which adjusted for the cabinet. Well, if that isn't the BBC RL bafflestep circuit, I'll eat my hat. And John DeVore and likes easy to drive flat impedance. Ah, the secrets of high-end audio...
My last thought on this, is that I want to try some air-coils rather than ferrites in the bass filter. The measurably worst components in a crossover. But as Sreten wisely pointed out, don't expect miracles. After all, the speaker is one big ferrite coil too.
The physics says that each driver is only outputting a 1/4 of the SPL for the same overall level. Thus lower distortion. I also quite like Roy Allison's idea of putting the (single) bass near the floor below the bafflestep cutoff of 500Hz on a 22 cm baffle. You then effectively double up the bass response again with the floor gain. It's clever.
If it was me, with all the supply issues in SA, and availability of coils is the biggest one it seems. I'd be looking into availability of enamelled copper wire and wind them round a bobbin to correct inductance value.
This is just a hobby, in the end. To build a speaker, I have to be excited about the idea being tested. I liked everything about this DeVore Gibbon 88 clone idea:
Add the BBC impedance and passive RL bafflestep, and it all falls into place. Knowing Boxsim a bit, it overdoes the bass response when designed flat, but this 22cm wide, 32L design is a peach IMO. The W170S is actually more of a closed box driver, IMO, so a sock in the port might be good too. Joachim Gerhard used to talk vaguely about his diffraction module which adjusted for the cabinet. Well, if that isn't the BBC RL bafflestep circuit, I'll eat my hat. And John DeVore and likes easy to drive flat impedance. Ah, the secrets of high-end audio...
My last thought on this, is that I want to try some air-coils rather than ferrites in the bass filter. The measurably worst components in a crossover. But as Sreten wisely pointed out, don't expect miracles. After all, the speaker is one big ferrite coil too.
Attachments
Lojzek, yes, it's the power handling I was after with two mids. But all merit to the design.
Steve, are you hinting I should build that gibbon clone to share results? I must say, I'm very much in favour of it! I have been since you posted it at first. The only thing is the low power handling ability. I'm pretty sure those speakers won't compete with my F20s.
As for the enclosure, is it ported? Would you put the tweeter in it's own chamber? If ported, would you tune it to 39Hz?
Steve, are you hinting I should build that gibbon clone to share results?
I must say, I'm very much in favour of it! I have been since you posted it at first. The only thing is the low power handling ability. I'm pretty sure those speakers won't compete with my F20s.As for the enclosure, is it ported? Would you put the tweeter in it's own chamber? If ported, would you tune it to 39Hz?
Steve, you made a big effort to help me in the past. It was a few months ago, and it gave me the extra confidence that I can maybe build my next speakers. My friend was, and still is extremely happy with what I built for him. Skytec based bookshelfs
Those speakers sounded unexpectedly good. I did nothing to the crossovers - I never got the time. Baby also arrived early. Things were crazy at the time.
Anyway, I can get hold of a similar driver in 6.5". Would that be a good candidate in the Gibbon clone? I can get the 902.193 or the 902.423. TS linked here.
The 193 looks like a bad idea. The 423 looks like it can't really give enough bass. I could get a 902.214, but I'll need a midrange with it. In a huge ported box, it should reach 24 Hz or so? It'll probably work better in a sealed 65 litre. I'm not sure.
Those speakers sounded unexpectedly good. I did nothing to the crossovers - I never got the time. Baby also arrived early. Things were crazy at the time.
Anyway, I can get hold of a similar driver in 6.5". Would that be a good candidate in the Gibbon clone? I can get the 902.193 or the 902.423. TS linked here.
The 193 looks like a bad idea. The 423 looks like it can't really give enough bass. I could get a 902.214, but I'll need a midrange with it. In a huge ported box, it should reach 24 Hz or so? It'll probably work better in a sealed 65 litre. I'm not sure.
Ok, so I also like sealed cabinets. There's something to be said about how a port can just warm up that bass, but if you can get it in a sealed thing, there's something nice about it.
Peach of a Gibbon 88 clone, I'll just take it one step further. Make it 160 cm high, 22 cm wide and 45 cm deep. I like that height; it's nice and eye-level when I'm sitting. Put the W170S and SC10N where they would normally be at the top of the box, and then add another W170S symmetrically below the tweeter. Both low's 4 ohm, tweak the crossover a little, and this 100 L sealed box should sound quite nice, I would guess.
The phase for elevation angles is a bit bad, but I suspect this won't be a problem in my room. I sit about 4 meters from the speakers.
Peach of a Gibbon 88 clone, I'll just take it one step further. Make it 160 cm high, 22 cm wide and 45 cm deep. I like that height; it's nice and eye-level when I'm sitting. Put the W170S and SC10N where they would normally be at the top of the box, and then add another W170S symmetrically below the tweeter. Both low's 4 ohm, tweak the crossover a little, and this 100 L sealed box should sound quite nice, I would guess.
The phase for elevation angles is a bit bad, but I suspect this won't be a problem in my room. I sit about 4 meters from the speakers.
Attachments
I'd expect a 6" bass to go louder than a 5" actually.
W 170 S - 8 Ohm
It doesn't like reflex at all, but then it's Qts 0.52:
mh-audio.nl - Vented System
3dB lump at the bottom end with reflex. Ideally you'd want a golden ratio (0.618:1:1.618) cabinet to avoid standing waves. Or 1:1.414:2 is another reasonable one. Something to think about there. I don't know the easy answer to that with a tower closed box. Maybe a lot of heavy stuffing in the box. Maybe divide the cabinet internally to 30L because overly big boxes worsen power handling. DeVore uses a double chamber reflex, you see.
FWIW, I don't like that 10R resistor on the bafflestep.
mh-audio.nl - Home
Seems to worsen the impedance. It's too much. 3.5R and 1mH seemed about right for 4dB.
I know how to model things like the SC5 these days. I think it's just the optimiser that needs more data with those old driver files. Boxsim has obviously been updated over the years and is not 100% compatible with old files. But the FR and phase works. I seemed to get on OK with projects from the databank anyway. https://boxsim-db.de/kategorie/systeme/drei-wege/
Those skytronic drivers look a possibility too, but a whole new ball game.
W 170 S - 8 Ohm
It doesn't like reflex at all, but then it's Qts 0.52:
mh-audio.nl - Vented System
3dB lump at the bottom end with reflex. Ideally you'd want a golden ratio (0.618:1:1.618) cabinet to avoid standing waves. Or 1:1.414:2 is another reasonable one. Something to think about there. I don't know the easy answer to that with a tower closed box. Maybe a lot of heavy stuffing in the box. Maybe divide the cabinet internally to 30L because overly big boxes worsen power handling. DeVore uses a double chamber reflex, you see.
FWIW, I don't like that 10R resistor on the bafflestep.
mh-audio.nl - Home
Seems to worsen the impedance. It's too much. 3.5R and 1mH seemed about right for 4dB.
I know how to model things like the SC5 these days. I think it's just the optimiser that needs more data with those old driver files. Boxsim has obviously been updated over the years and is not 100% compatible with old files. But the FR and phase works. I seemed to get on OK with projects from the databank anyway. https://boxsim-db.de/kategorie/systeme/drei-wege/
Those skytronic drivers look a possibility too, but a whole new ball game.
Last edited:
Hi. Thanks! I'll keep in mind that golden ratio - that's a cool thing. I actually changed the two 10 ohm resistors to 4.7 just after posting. I preferred what it did for the impedance. I was thinking of that tall box and using many deflectors in it - jutting pieces of bracing here and there. Or I could make two 50 L chambers.
When I tweak a crossover, I look at peaks and dips in FR and impedance, and then I can identify where it's coming from, and then I play around with the value of the suspected component. I imagine that one can get quite obsessive and tweak each component after building.
When I tweak a crossover, I look at peaks and dips in FR and impedance, and then I can identify where it's coming from, and then I play around with the value of the suspected component. I imagine that one can get quite obsessive and tweak each component after building.
Oh, I see! Now we are in the middle of something new, namely monkey business.
@ Steve: The TS parameters of W170S-8 are considerably worse than specified by Visaton or at least it was so 15 years ago when I did a project with these. I'll try to dig up impedance measurement if there still is one on my ancient PII-300 Mhz machine, overclocked to 450Mhz for the last 20 years and still working.
AFAIK, the W170S has changed quite a lot over the years.
I spent some time looking up the DeVore Gibbon 88 last night, since we may be interested in something like it. Looks like a SEAS 6" driver not far removed from this: H1215-08 CA18RNX. Actually a straightforward reflex if I can trust my eyes.
I also found out quite a lot about the huge Orangutan 96 speakers that fascinate a lot of people here:
https://parttimeaudiophile.com/2016/11/27/john-devore-talks-larger-drivers-and-experimental-speaker-designs/
How hard are they to clone? Actually, not easy, and certainly not cheap. I know a SEAS Exotic motor when I see one.
WOOFERS
Seems like John spends years on his designs:
https://parttimeaudiophile.com/2017/01/20/ces-2017-video-devore-fidelity-to-unleash-20000-flagship-orangutan-loudspeaker/
I spent some time looking up the DeVore Gibbon 88 last night, since we may be interested in something like it. Looks like a SEAS 6" driver not far removed from this: H1215-08 CA18RNX. Actually a straightforward reflex if I can trust my eyes.
I also found out quite a lot about the huge Orangutan 96 speakers that fascinate a lot of people here:
https://parttimeaudiophile.com/2016/11/27/john-devore-talks-larger-drivers-and-experimental-speaker-designs/
How hard are they to clone? Actually, not easy, and certainly not cheap. I know a SEAS Exotic motor when I see one.
WOOFERS
Seems like John spends years on his designs:
https://parttimeaudiophile.com/2017/01/20/ces-2017-video-devore-fidelity-to-unleash-20000-flagship-orangutan-loudspeaker/
Hi Steve. I need to build something which will be able to handle the potential of my amplifier(s) - 150 W and 180 W (8 ohms). My room is big, and I often need loud sound, but I also want the room filled with very low volume. 100 W speakers should be happy on these amplifiers.
I really like that Gibbon clone you posted. So what I think I'm going to do is build something using the W170S and SC10N, but I'll make it something much smaller for a secondary system. I need something in my dining room, where I spend most of my evenings cooking, building or chilling by the fire (fire is in the family room, which is basically part of the dining room). At the moment I put my main system loud and can barely hear it from the kitchen. I don't want it too loud, else it will bother my mom-in-law who lives in a flat near the living room.
So I'm going to see if I can modify the Gibbon 88 clone to a bookshelf size. I'll compare them to my Celestions, and just judge the overall sound, and if I like what I hear, I'll build the bigger speakers as a main system.
You've helped me massively, and if I can help you with anything (amplifiers are my strong point), let me know.
I really like that Gibbon clone you posted. So what I think I'm going to do is build something using the W170S and SC10N, but I'll make it something much smaller for a secondary system. I need something in my dining room, where I spend most of my evenings cooking, building or chilling by the fire (fire is in the family room, which is basically part of the dining room). At the moment I put my main system loud and can barely hear it from the kitchen. I don't want it too loud, else it will bother my mom-in-law who lives in a flat near the living room.
So I'm going to see if I can modify the Gibbon 88 clone to a bookshelf size. I'll compare them to my Celestions, and just judge the overall sound, and if I like what I hear, I'll build the bigger speakers as a main system.
You've helped me massively, and if I can help you with anything (amplifiers are my strong point), let me know.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.