System, actually I agree completely with Sir John.Kgrlee, I've gotta say that it is very enjoyable to talk to you online. Evidently you are one of the last people who believed that Britain was founded as a manufacturing nation, and that Mrs. Thatcher (Gawd bless her in most respects, a politician who had actually had a real job at some stage...) was totally wrong in killing off our factories.
I used to bump into Sir John Harvey-Jones, Chairman of ICI and lover of bright ties, on a regular basis. It was always his belief that REAL WEALTH is created by MANUFACTURING alone. Sound man.
Baroness Thatcher's pogrom against British manufacturing happened during the time I was privileged to be involved in helping the Wharfedale, Leeds factory make the most consistent drive units in the world ... and still make a profit.
It was my most satisfying and humbling experience. It taught me it wasn't the British worker that led to the demise of British manufacturing but British management & government.
Sadly, Thatcher's policies benefited only those who played with wealth, the yuppie generation, ... and decimated the CREATORS of wealth.
My involvement with the Chinese factories that now make Celestion & KEF products was cos the UK management of their UK factories were still in the 19th century and considered workers, QA etc far beneath their notice.
However, there are Chinese factory managers who are far more enlightened, both about imbuing their workers with pride and in understanding quality is BUILT into a product, not inspected. I was pleased to help them make better speakers.
But I still think Wharfedale, Leeds, in God's own county, could be a viable operation in today's market and I mourn the loss of the expertise and pride. [/rant]
Thanks for your simulations, System.
Unfortunately, what is missing is the pretty awful frequency response and pretty awful impedance curve of the Wharfedale treble unit.
I'm not sure if the Visaton software can accept such information but if it doesn't, it is little more than a toy. (BTW, are you actually making a speaker with the Visaton & DT94? If so, there is a small mod that will probably improve the sound a lot.)
Can I urge you to get a copy of Schuck's AES paper which explains loadsa stuff.
Also to buy or make a cheap measurement mike and try something like ARTA or REW.
If you are playing with these, all my comments will make sense.
What happens is ... if you combine the awful response of the Laser 90Bs treble unit in the box, its awful impedance curve and the awful xover response, (which BTW, will be differently awful from the awful one you sim, the end result is surprise, surprise, not so awful after all.
_____________________
I note from Robin's interview that he has high regard for Drs. Don Barlow (note spulin!) Peter Fryer and other luminaries. I worked with Don & Peter and share his regard.
I won't join in on the relative merits of various treble units as you are obviously expert on such matters.
My only concern (apart from a bit of wanking
) is to help Mike with what poor knowledge & experience I might reasonably provide as a beach bum._________________
Mike, have you found the 0.2mH inductor in the treble xover circuit. If its gone, you MUST replace it.
Last edited:
"Mike, have you found the 0.2mH inductor in the treble xover circuit. If its gone, you MUST replace it"
kgrlee, yes, I found it... even for an electronic 1st grader like me it wasn't hard to find, measurement awfully difficult until I purchased a multimeter which would measure inductance, which I've done and the reading I get is 1.8hH.
Off-topic, but you mentioned Celestion- I've obtained 6 of the 5" extended-range (neo magnet) jobs for future projects. What little is in print about them is very positive. They are said to be "designed in GB" but they're made in China.
kgrlee, yes, I found it... even for an electronic 1st grader like me it wasn't hard to find, measurement awfully difficult until I purchased a multimeter which would measure inductance, which I've done and the reading I get is 1.8hH.
Off-topic, but you mentioned Celestion- I've obtained 6 of the 5" extended-range (neo magnet) jobs for future projects. What little is in print about them is very positive. They are said to be "designed in GB" but they're made in China.
Hi,
What would be useful is some real information, not I know stuff from
working for wharfedale that you don't, such as the acoustic target
for the "horrible" 3rd electrical filter to fix the tweeter issues.
What tweeter issues ? Fs Q ? response ? both ? what sort of correction ?
This would allow some intelligent estimations for substitute tweeters.
rgds, sreten.
High consistency low variability manufacturing is a sign of "quality",
in terms of quality control, but that doesn't mean for an instant
the driver design is high quality, only you get decent matching.
What would be useful is some real information, not I know stuff from
working for wharfedale that you don't, such as the acoustic target
for the "horrible" 3rd electrical filter to fix the tweeter issues.
What tweeter issues ? Fs Q ? response ? both ? what sort of correction ?
This would allow some intelligent estimations for substitute tweeters.
rgds, sreten.
High consistency low variability manufacturing is a sign of "quality",
in terms of quality control, but that doesn't mean for an instant
the driver design is high quality, only you get decent matching.
Sreten, this is all previous century so I can't answer any of your questions. I DON'T KNOW any of the 'real information' anymore. And the unit is now Unobtainium so what's the point?
System believes all treble units are the same but believing you can find another unit which is a drop in replacement is nearly as foolish. eg the response will vary ALOT depending on the box you mount the unit in.
You really need to design a new xover from scratch to match the new unit, its mounting ... and this has to integrate with the bass unit too.
There are other web sites which tell you how to do this, some good and some laughable but I've not really researched this.
I'm sorry if you don't like this wanking but I wouldn't have posted at all if system hadn't implied all treble units are the same. I'll go back to my hole now. It's likely that you (sreten) will be at least as useful as anything I can provide.
What you say is true but there really isn't any point in a beach bum arguing over the quality of Unobtainium Wharfedale units, previously made in God's own county.
I HAVE told you what I'd be looking for if I did abandon my present lifestyle. I'm sure you have your own ideas.
I'd be quite interested to hear what it was that finally killed off the Leeds Wharfedale factory. Workers wages in Britain AREN'T all that high compared to our competitors. What is high is profit expectations and Directors and Managers wages. The tax burden of a huge state spend and high house prices don't help either. Nor does a banking sector that does NOT act in its customers interest. The geniuses who move manufacturing abroad forget they destroy the local wage income that creates purchasing power. This is why the EU is doomed.
As for tweeters being tweeters, I can justify that. Look at these three basically identical 19mm models, the only difference being dome material and efficiency:
Vifa D19TD-05 3/4" Poly Dome Tweeter 264-500
H0737-08 19TFF 1
H0532-08 19TAFD/G
Not hard to adjust the input resistor to the filter to compensate. This would do with a 6" paper bass. If you make the bass bafflestep coil bigger, you just make the shunt components smaller but preserve the product. If attenuation becomes severe, the tweeter coil gets a bit smaller, say 0.25mH becomes 0.2mH. But not hard.
As for tweeters being tweeters, I can justify that. Look at these three basically identical 19mm models, the only difference being dome material and efficiency:
Vifa D19TD-05 3/4" Poly Dome Tweeter 264-500
H0737-08 19TFF 1
H0532-08 19TAFD/G
Not hard to adjust the input resistor to the filter to compensate. This would do with a 6" paper bass. If you make the bass bafflestep coil bigger, you just make the shunt components smaller but preserve the product. If attenuation becomes severe, the tweeter coil gets a bit smaller, say 0.25mH becomes 0.2mH. But not hard.
Attachments
Agree with all that, system. Please petition Baroness Thatcher on behalf of all those who lost their jobs including ex-Wharfedale employees in Yorkshire.
system, if you are really interested in the subject, I urge you to buy a cheap Behringer ECM8000 and try the free version of ARTA or REW.
Looking at datasheets is all very well and so is reading the comics about speaker design. But when you actually measure something you've put together, you quickly find what is solid & liquid BS.
You might also like to ask your hero, Robin Marshall, whether he thinks 'tweeters are tweeters'.
sreten may have some experience with measuring speakers & some of the free speaker design stuff available. I haven't tried any of them.
Please note the priority. Some means of measuring speakers FIRST. Supa dupa software to automate design later when you have a better idea of the need for this.
There's also stuff about formalising Listening Tests as part of the design process which I won't go into.
______________
Mike, can you take a picture of your xover? Can you trace the circuit and post a pic of this too?
The 1.8mH you measured is almost certainly the bass unit inductor. Did you unsolder one end of the inductor to measure it?
Last edited:
Hi,
FWIW the amateurs priority is drivers available with reliable measurements,
and then using some decent simulation and then some tweaking by ear.
Boxsim seems a good sim, if you can trust Visatons data,
but not very good for anything else, unless drivers are
similar, which generally in the critical regions are not.
Anything beyond numskull speaker design is fine by me, at
least your thinking about it, which solves half the problem.
Solving the other half properly or intelligently is the problem.
rgds, sreten.
FWIW the amateurs priority is drivers available with reliable measurements,
and then using some decent simulation and then some tweaking by ear.
Boxsim seems a good sim, if you can trust Visatons data,
but not very good for anything else, unless drivers are
similar, which generally in the critical regions are not.
Anything beyond numskull speaker design is fine by me, at
least your thinking about it, which solves half the problem.
Solving the other half properly or intelligently is the problem.
rgds, sreten.
I'd be quite interested to hear what it was that finally killed off the Leeds Wharfedale factory. Workers wages in Britain AREN'T all that high compared to our competitors. What is high is profit expectations and Directors and Managers wages. The tax burden of a huge state spend and high house prices don't help either. Nor does a banking sector that does NOT act in its customers interest. The geniuses who move manufacturing abroad forget they destroy the local wage income that creates purchasing power. This is why the EU is doomed.
I agree in many respects. As a student of history (another hobby- another BA late in life) I've often wondered why British mine and factory owners kept wages at subsistence level while trying to find markets for their manufactured goods, usually abroad because few British workers could afford to buy the fabulous products they were making. Over-generalisation, I know, but it has come full-circle in a way, now there are fewer workers required while housing is ridiculously priced (housing is no better in Australia, in fact Sydney is one of the worse places in the world for over-priced homes).
Well, Mike, if you listen to my learned friends here, you will abandon all hope of knocking that simple little speaker into shape, obviously far too difficult to try at home....
Tina TI is an op-amp simulator circuit that would interest me if I ever picked up amplifier design again. It does have a good filter modeller in it, but not one that has much relevance to loudspeakers, lacking FRD (Frequency Response and Phase) and ZMA (Impedance) loudspeaker modelling data.
My simple-minded approach relies on well-behaved standard sorts of drivers with published data and frequency response plots. Cabinet solutions are based on simple optimal stuff that works. It is not simple-minded at all, of course. More clear-minded. I leave it to the forum windbags to debate 0.618:1:1.618 or 0.8:1:1.25 cabinet ratios.
To solve a problem, the best approach is to simplify it. That is what I suggest. Treble filters are really not hard. I use Zobels and an input resistor because they just work with most stuff and it's not hard to look up tweeter inductance.
Bass filters vary a little and I assume you're working with the usual 0.6-0.9mH units. Plastic is a slow material which responds to simple filters. A simple coil and capacitor circuit will work passably well with little need for a notch as with the Wharfedale Laser 90B design. With paper bass, it helps to notch the AAL3 Resonance to reduce the annoying bass cone-breakup sound which is around 5kHz for 6" paper units.
Yes, Robin Marshall IS a hero of mine. I love his stuff. Maybe that's because I understand it. And like Lynn Olson, Joachim Gerhard and John Linsley-Hood, he can explain it simply. Good players make any game look easy. So does Troels Gravesen with this SEAS CURV plastic bass design:
SEAS CURV
Tina TI is an op-amp simulator circuit that would interest me if I ever picked up amplifier design again. It does have a good filter modeller in it, but not one that has much relevance to loudspeakers, lacking FRD (Frequency Response and Phase) and ZMA (Impedance) loudspeaker modelling data.
My simple-minded approach relies on well-behaved standard sorts of drivers with published data and frequency response plots. Cabinet solutions are based on simple optimal stuff that works. It is not simple-minded at all, of course. More clear-minded. I leave it to the forum windbags to debate 0.618:1:1.618 or 0.8:1:1.25 cabinet ratios.
To solve a problem, the best approach is to simplify it. That is what I suggest. Treble filters are really not hard. I use Zobels and an input resistor because they just work with most stuff and it's not hard to look up tweeter inductance.
Bass filters vary a little and I assume you're working with the usual 0.6-0.9mH units. Plastic is a slow material which responds to simple filters. A simple coil and capacitor circuit will work passably well with little need for a notch as with the Wharfedale Laser 90B design. With paper bass, it helps to notch the AAL3 Resonance to reduce the annoying bass cone-breakup sound which is around 5kHz for 6" paper units.
Yes, Robin Marshall IS a hero of mine. I love his stuff. Maybe that's because I understand it. And like Lynn Olson, Joachim Gerhard and John Linsley-Hood, he can explain it simply. Good players make any game look easy. So does Troels Gravesen with this SEAS CURV plastic bass design:
SEAS CURV
system, perhaps you are actually a guru masquerading as a beach bum. A lot of what you say is so nearly true.
(sorry. Couldn't help that)But does Robin Marshall share your opinion on treble units?
____________
In the face of all this expert knowledge, I'm just going to try and ensure, Mike has the correct value inductor at the proper place in the treble xover.
I won't say anything more unless someone proposes a xover which might cause damage to whatever unit Mike chooses.
Well, Mike, if you listen to my learned friends here, you will abandon all hope of knocking that simple little speaker into shape, obviously far too difficult to try at home....
Tina TI is an op-amp simulator circuit that would interest me if I ever picked up amplifier design again. It does have a good filter modeller in it, but not one that has much relevance to loudspeakers, lacking FRD (Frequency Response and Phase) and ZMA (Impedance) loudspeaker modelling data.
My simple-minded approach relies on well-behaved standard sorts of drivers with published data and frequency response plots. Cabinet solutions are based on simple optimal stuff that works. It is not simple-minded at all, of course. More clear-minded. I leave it to the forum windbags to debate 0.618:1:1.618 or 0.8:1:1.25 cabinet ratios.
To solve a problem, the best approach is to simplify it. That is what I suggest. Treble filters are really not hard. I use Zobels and an input resistor because they just work with most stuff and it's not hard to look up tweeter inductance.
Bass filters vary a little and I assume you're working with the usual 0.6-0.9mH units. Plastic is a slow material which responds to simple filters. A simple coil and capacitor circuit will work passably well with little need for a notch as with the Wharfedale Laser 90B design. With paper bass, it helps to notch the AAL3 Resonance to reduce the annoying bass cone-breakup sound which is around 5kHz for 6" paper units.
Yes, Robin Marshall IS a hero of mine. I love his stuff. Maybe that's because I understand it. And like Lynn Olson, Joachim Gerhard and John Linsley-Hood, he can explain it simply. Good players make any game look easy. So does Troels Gravesen with this SEAS CURV plastic bass design:
SEAS CURV
Thanks for the info, Steve. Actually some of the stuff I find on this site makes my head spin! While I am a technical officer (working in meteorology) my academic leanings are to the humanities, the electronics are a challenge for another part of the brain perhaps, but it has to be in small doses or I faint
Mike, I'm going to conquer my fear of system and try & take you through what I did for speaker design before I joined the industry and had facilities to measure response, bla bla.
After more than half a lifetime doing this for real, I still think its a good procedure.
But if system or anyone else wants to take over, I'll bow out as it's impossible to follow 2 gurus.
__________________
In order of importance.
1) Have you identified the elusive 0.2mH inductor?
2) Have you chosen your replacement treble unit? If not, Jaycar CT2005 looks good value. Will it fit?
3) Can you use a router?
4) Can you borrow a working 'good' speaker of similar size? What is it?
5) Do you have or can beg borrow a microphone with associated preamp?
6) What model & brand
After more than half a lifetime doing this for real, I still think its a good procedure.
But if system or anyone else wants to take over, I'll bow out as it's impossible to follow 2 gurus.
__________________
In order of importance.
1) Have you identified the elusive 0.2mH inductor?
2) Have you chosen your replacement treble unit? If not, Jaycar CT2005 looks good value. Will it fit?
3) Can you use a router?
4) Can you borrow a working 'good' speaker of similar size? What is it?
5) Do you have or can beg borrow a microphone with associated preamp?
6) What model & brand
Last edited:
Hi kgrlee,
1) Yes, there is only one, I measured at 0.18mH
2) Yes, it was fitted around six months ago. This is it: PSB Speakers G Design 1" Aluminum Tweeter s GT1 GB1 GC1 | eBay
It wasn't a "good fit", but then I hadn't expected any tweeter to go in easily (although the Dayton truncated 8Ohm would have been the closest fit).
3) Yes, with a high degree of inexperience. Never owned or used one until this year.
5) No. I live in the outback although possibly not as remote as you, being 9.5 hrs by road from Sydney.
6) Not applicable.
Cheers,
Mike
That's much more like what I'd expect from a DMM measuring inductance of the 0.2mH
This is actually quite a good choice. The front plastic plate is designed (like the Wharfedale unit) to be mounted directly on the box front panel.
Can you post a pic?
Jaycar CT2005 for instance, would need to be recessed into the front panel so the plastic plate was flush. That's the reason for asking if you could use a router.
__________________________
Now we have the treble unit mounted properly and the xover bits in good order, the next thing is to decide what phase the treble unit should be.
For this you need a source of Pink Noise. Download http://www.realtraps.com/nti_minirator.exe This free application is useful for loadsa other stuff too.
On your 2 speakers, play the PNOISE signal through each speaker & check that they sound similar.
Then change the phase on ONE treble unit and compare the 2 speakers again.
The correct phase will sound like "shuuush". The wrong phase will sound like "sssssssss".
What is happening is that with the wrong phase, the outputs of bass & treble cancel out & there will be a big dip at the crossover point (the "shuuush" frequencies) leaving the "sssss" high frequencies to stand out. On music, the wrong phase will make voices sound laid back and whispy.
___________________________
The reason for asking
4) Can you borrow a working 'good' speaker of similar size? What is it?
is to determine the correct level for the treble unit.
Compare the speaker (with correct phase) with a known accurate speaker and adjust the resistor at the input of the treble xover until they sound similar. Use music to do this rather than Pink Noise.
With a lot of experience, and a really good 'reference' speaker, you can do a lot of xover work by ear too.
___________________________
That's as far as you can go without some means of measuring response.
BTW, all the stuff I've described will be done by a good speaker designer even when he has all the fancy gear & automatic programmes too ... just to check the supa stuff is making sense.
Hi kgrlee,
The link came up with: "REALTRAPS PAGE NOT FOUND". I'll keep trying tomorrow though.
I know about recessing tweeters (and other drivers) because I built a large 3-way (using an off-the-shelf crossover) with Peerless India (Bought out of UK- Willy's HiFi) mid and woofers with Dynavox tweeters:
Dynavox TD2801XL 1-1/8" Silk Dome Tweeter 275-200
Posting a pic is above my ability level, the url has several pics:
PSB Speakers G Design 1" Aluminum Tweeter s GT1 GB1 GC1 | eBay
The speaker build was very testing and taxing as my electronics knowledge is quite low, and my woodworking ability almost non-existent. I am one of those people who set themselves challenges (not the sort which can be fatal, however, like "extreme sports") because life is boring without surprises and challenges.
Not sure what qualifies as a "good" speaker. My Wharfedale 510s are "good'' albeit with new tweeters (Vifa ring radiator). Same impedance as originals and I didn't play with crossovers. My ignorance of things Hifi is such that I ran the 510s for 23 years on an NEC amp rated 8-16 Ohm (510s are 4 Ohm) and eventually the tweeters blew. While they are still "good" I've moved them to the AV amp for their own safety (and the safety of my amp which I like very much- cost an arm and a leg, before amps got cheap).
You've probably noticed that I have an assortment of tweeters (also have some horns sitting in a box), I saw all the different sorts and thought I'd try them. To be honest, I like the ring radiators the best so far, but anyway it has been a revelation listening to speakers with good, functioning tweeters.
Cheers,
Mike
The link came up with: "REALTRAPS PAGE NOT FOUND". I'll keep trying tomorrow though.
I know about recessing tweeters (and other drivers) because I built a large 3-way (using an off-the-shelf crossover) with Peerless India (Bought out of UK- Willy's HiFi) mid and woofers with Dynavox tweeters:
Dynavox TD2801XL 1-1/8" Silk Dome Tweeter 275-200
Posting a pic is above my ability level, the url has several pics:
PSB Speakers G Design 1" Aluminum Tweeter s GT1 GB1 GC1 | eBay
The speaker build was very testing and taxing as my electronics knowledge is quite low, and my woodworking ability almost non-existent. I am one of those people who set themselves challenges (not the sort which can be fatal, however, like "extreme sports") because life is boring without surprises and challenges.
Not sure what qualifies as a "good" speaker. My Wharfedale 510s are "good'' albeit with new tweeters (Vifa ring radiator). Same impedance as originals and I didn't play with crossovers. My ignorance of things Hifi is such that I ran the 510s for 23 years on an NEC amp rated 8-16 Ohm (510s are 4 Ohm) and eventually the tweeters blew. While they are still "good" I've moved them to the AV amp for their own safety (and the safety of my amp which I like very much- cost an arm and a leg, before amps got cheap).
You've probably noticed that I have an assortment of tweeters (also have some horns sitting in a box), I saw all the different sorts and thought I'd try them. To be honest, I like the ring radiators the best so far, but anyway it has been a revelation listening to speakers with good, functioning tweeters.
Cheers,
Mike
Mike, I wanted to see how you mounted the PSBs in your Diamonds.
The 510s if original would have been a good choice as reference speakers. Unfortunately, we have no idea if the treble response of the system with the Vifa's is flat or even at the right level.
Can I suggest you get some Polyswitch Speaker Protectors: Jaycar RN-3460 and put them at the input to the treble sections of your xovers.
They should be in series with the resistor on the Diamond xover. I think there is a similar resistor on the 510. If not, the Polyswitch should be in series with the first capacitor on the input of the treble xover.
These won't guarantee the treble units won't die but they will help a lot.
The 510s if original would have been a good choice as reference speakers. Unfortunately, we have no idea if the treble response of the system with the Vifa's is flat or even at the right level.
Can I suggest you get some Polyswitch Speaker Protectors: Jaycar RN-3460 and put them at the input to the treble sections of your xovers.
They should be in series with the resistor on the Diamond xover. I think there is a similar resistor on the 510. If not, the Polyswitch should be in series with the first capacitor on the input of the treble xover.
These won't guarantee the treble units won't die but they will help a lot.
Thanks again kgrlee. The Diamonds, I'm pretty sure, have a polyswitch in the crossover circuits. I'm almost certain the 510s do not. Unfortunately the 510s have the worst crossover access known to man, in fact they were the cause of me joining this forum as I was completely stumped as to the method of accessing the cabs. The tweeters have a plastic cover which hides the screws (glued on) and the mids and woofers aren't screwed on- they "cam" onto the baffle, presumably using a special tool like the small ones used for removing watch backs, for each different sized woofer which, needless to say, I do not have. Not an experience to be repeated! The 8" sound woofers wonderful, and I believe the frames are cast aluminium- which is nice, but after 23 years, often in harsh environments, the frames were covered in surface corrosion.
Since I'm at work photos of the PSBs in place will have to come when work finishes this afternoon, sorry, misunderstood your request.
Since I'm at work photos of the PSBs in place will have to come when work finishes this afternoon, sorry, misunderstood your request.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.