Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Multi-Way Conventional loudspeakers with crossovers

celestion 66 needs Mid-range
celestion 66 needs Mid-range
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools Search this Thread
Old 25th September 2013, 09:08 AM   #1001
nigel pearson is offline nigel pearson  United Kingdom
diyAudio Member
Join Date: Mar 2007
Location: Oxfordshire
The Caruso I put in yesterday was no diversion from the balancing of a speaker . It needs everything to be correct so as not to the diminish the ember of the original performance . One thing I do find with the 1903 Caruso is it makes my ears do what they do in real life . No idea how that is possible . My son noticed the same . I hate to get mixed up in the cult of Caruso . Problem is it would be stupid to say it's not hi fi thus I can not listen .

When I first heard 78's through correctly driven Quad 63's I found it possible to go into a time machine . If as little as possible distracts , the mind invents the missing octaves . Bass is one that usually is OK .

If 78's or ancient LP's sounds bad and the condition of the discs is OK it usually is things wrong that are universally wrong that makes it sound strange . Get the EQ right is a priority . The cheap way is to buy a Leak Varislope and build a buffer ( OPA 2604 ) .

An extreme example of little things being right is what I am listening to right now . My hairdresser brought her boom box made hi fi to repair . It has drive units seldom seen in the cheapest transistor radio these days . However they are not one million miles removed from mid range units . The boxes the thinnest MDF which gives them a chance ( not plastic ) . Two applications of PVA glue ( 10 % water ) as thin as possible ( not on the surround ) . Some Thomas Fast special foam offcuts I couldn't bring myself to throw away and bingo they sound OK . In fact the BBC R3 voice is unusually good . Listening now the Bach Toccata & F . Even that is reasonable . It probably has taken as long to write about it as do it ! This is not a universal fix . I found out that ripped valve radio speakers sounded better after this repair . The trick is to use as little as possible ( a Rizla paper to the cone rear to look almost like new when dry ) . Lets be clear this way of doping usually sounds better than the plastic cone speakers and has equally strengths where they do ( clout ) . I suspect 2 octaves of distrotion reduction with this boom box . The likelihood is 200 to 6 kHz now ( 100 to 10 kHz - 10 dB at a guess ) . Absolutely no tizz nor chestiness ( Thanks Thomas ) !

Poulenc . I went to a R3 recording of the Gloria a few months back . Bliss . The Organ concerto I heard in York Minster before the fire . His life a bit like mine . He stopped talking and did .
  Reply With Quote
Old 26th September 2013, 07:23 AM   #1002
DennyG is offline DennyG  Australia
diyAudio Member
Join Date: Feb 2010
Location: Sydney
Hi Nigel, Any advice on removing a ding in a metal dome tweeter? Fortunately, I was able to buy a spare unit but repairing the old one would be nice
  Reply With Quote
Old 26th September 2013, 11:34 AM   #1003
nigel pearson is offline nigel pearson  United Kingdom
diyAudio Member
Join Date: Mar 2007
Location: Oxfordshire
If the dome is held by screws the dome often is very easy to remove . If you find a soft surface which has the right shape use a pencil rubber to ease the dome out into the form . Pay attention to the coil staying in shape . If you preserve the ferro-fluid that damps the dome . It isn't always fitted . I would risk the minutest amount of candle wax on the dome back to prevent brake up modes . I suspect the micro mass will not reduce transient response . In fact a system pulled out of resonance is faster .

Best of luck .
  Reply With Quote
Old 28th September 2013, 06:55 PM   #1004
alan-1-b is offline alan-1-b  United Kingdom
diyAudio Member
Join Date: Apr 2006
Location: Midlands
Default LT-Spice ; frequency bandwidth ; domes and cones ; turntables ;

G'day Denny ,

indeed it seems a good music festival you were fortunate to be able to attend !

As you have the LT-Spice , and if you have time available ,
it will still be interesting and likely useful to see the effect of the upper midrange crossover filter on the dome-mid.
I refer you to my Post #952 on Page 96 for the details of how to proceed.

*** *** ***

'ullo Nigel ,

I agree , that when the midrange is correct from a recording that for the bass and treble one's mind can invent the missing octaves ,
however that may only be for those of us who were conditioned since childhood by hearing only limited frequency range recordings and/or single cone wide-range driver loudspeaker systems.
Perhaps the children of the 1980s and since have heard more Bass and Treble , even if low quality through Boom-Box audio systems and similar in-car audio systems ,
and as such their brains are formed around hearing Bass and Treble , and even accentuated over the midrange ... as some of those replay systems are that their parents have owned.
I do not know how they hear - I know only how I hear , and how those people with whom I engage in comparative listening sessions seem to hear.

Very clever about putting a tiny amount of Wax on the back at the Apex of a metal dome !
I should have thought of that ...
With a few of the old Paper dome tweeters , when there was no damping material in the cavity behind , I put some soft Fibreglass wool.
That substantially reduced treble resonances.
I was reluctant to apply PVA to paper tweeters , because the addtional stiffness resulting can cause a larger break-up resonance around a narrow frequency band ,
and likely in the audible spectrum.

I have used PVA with very thin paper on the backs of mid and bass paper cone drivers for repairs to good effect.

Audibly much better than the entirely Plastic cone drivers are the recent very thin dual layer of plastic + metal that KEF used for their iQ* , XQ* , iQ** , XQ** series of loudspeakers ,
however KEF have now changed that for the new iQ*** series ,
and I do not know to what , nor have I heard any of this new series.

I should have expected a lover of Rim-drive , Idler-wheel turntables to state that a Belt is a primary problem ... and I do not entirely disagree ,
and perhaps with a very low vibration motor the Idler wheel will not feed significant vibration into the platter to either interfere with the stylus tracking the groove nor to modulate the rotational speed ,
however , one is going to need a notionally perfect circle idler wheel with perfect centre fixing to the motor shaft to commence to achieve stable speed ,
and as you do not like high mass platters , then how does one maintain sufficient Rotational Inertia for the multi-varying stylus' load drag through the vinyl to the platter to not modulate the speed of the platter ?

I have not been able to inspect the Loricraft Garrards 501 and 601 to see the quality of idler wheels , etc ...
nor have I heard either in use yet , however you will know what is in both , etc ... any comments about ?

Last edited by alan-1-b; 28th September 2013 at 06:58 PM. Reason: to add a word and change a few others
  Reply With Quote
Old 29th September 2013, 08:52 AM   #1005
nigel pearson is offline nigel pearson  United Kingdom
diyAudio Member
Join Date: Mar 2007
Location: Oxfordshire
Some times the dome is held by three wedge shaped clips . A couple of flat blade screw drivers will ease them off . Rotate until free . I think a spoon will make a good form . The rubber refereed to is the tip of the pencil type . A new one would be best .

With a belt drive getting the power to the platter is the question . I have no objection to belt . Idler is hard work and has many almost impossible problems . Lenco's are very good . The idea of a stiff belt like a leader tape is practical . Verdier's smooth resistance via the oil also .

PVA for other than bass units seems to add too much mass except when the cheaper units .

I am hooking up my Magneplanar bass and Eminence 12 LTa mid top when I get a moment ( tax return next ) . I a looking for dynamics an verve more than accuracy . The idea is to estimate how the open baffle best be configured ( 4 x 2 ply , top 1/3 central or offset 12Lta + 15 inch bass QTs 1.2 to bottom 1/3 ) . A frame and cardboard at stage one . As I have highly accurate speakers I can tolerate what by physics alone can not be accurate as a party speaker ( It must play Caruso well ) . My friend John says at hi fi levels I might get a pleasant surprise from a drive unit looking bad by even transistor radio standards . <1 g balls ( 3 ) of cotton wool to tame the whizzer cone that has mid-range boost .

Eminence Beta-12LTA 12" Full-Range/PA Driver 290-409
  Reply With Quote
Old 3rd October 2013, 05:33 AM   #1006
DennyG is offline DennyG  Australia
diyAudio Member
Join Date: Feb 2010
Location: Sydney
Default Re: post 952 - frequency bandwidth

The spoon idea is a beauty.

I've had a go with the LTSpice software to produce the plots 2 and 3 as per post 952 in this thread.

Unless I am intepreting the design incorrectly plot 1 will be a line at 0db. I don't know yet how to combine the plots so plots 2 & 3 are on separate charts. The circuits used to generate the plots are included. I've used separate components in the circuits for clarity. Can these be further refined?




  Reply With Quote
Old 4th October 2013, 07:06 PM   #1007
alan-1-b is offline alan-1-b  United Kingdom
diyAudio Member
Join Date: Apr 2006
Location: Midlands
Default LTSpice plots ; Tweeter Dome damping ; Open Baffle woofer

G'day Denny ,

Great stuff !

Both circuits as you have drawn and used are correct.

About the plots:

Likely the 0dB line is for a no filter prediction.
To test that , take out the 1.8 ohm resistor from the Input to the filter , and have no resistor there ,
unless your program requires a resistor with that capacitor , and if it does then put 0.10 ohm there.
Try the above with the 0.34mH + 3.9uF + 3.9 ohm circuit.
We can then see the action of the 1.8 ohm resistor.

The .34mH + 3.9uF + 3.9 ohm circuit simulation prediction for 5kHz is very close to what I have all along hoped for following my own estimations.
To try an alternate which will show us further , please do a simulation with 3.9uF + 2.7 ohms there ,
and with the 1.8 ohms in at the Input so that we can compare the new simulation directly to the original with the 1.8 ohms in.

The 4.2mH + 4.7uF circuit attenuates too much at 5kHz ,
and indicates that Celestion allowed for the specific Impedance characteristic of their driver at 5kHz better than was predicted by a self-proclaimed crossover designer here in the UK
who seems to trust only his particular computer program but seems to not know how to correctly program it.
Unfortunately he has ruined at least two Celestion owners' 66s , and seemingly ruined their enthusiasm for their 66s ,
and unfortunately he is still advertising his Service !

I have two ideas , of which one may work well for owners of the MD500 version of the dome
for samples which have a degree of rough frequency response around 5kHz or a little lower ,
however so as to not confuse readers of this thread , let us look at the two I have described above before we introduce any other variations.


After some time in my mind on the modification idea that Nigel described to damp a resonance in a metal dome tweeter I have remembered a potential problem if one uses Candle Wax.
For some metal domes , the metal itself is directly joined to the cylindrical Former around which the voice-coil is wound.
The voice-coil heats up when the music is playing , and will heat quite a lot with loud music , and more so with modern high-power tweeters that some users of play to very loud.
This heat will conduct from the coil into the former and into the dome , and thus soften the wax , and will melt it in a high-temperature application.
The molten wax , even though it is only a very small quantity ,
will in part flow into the gap between the voice-coil and the surrounding magnet pole ... and solidify there after the play session is finished.
Sure , it will melt again during the next session , but it will be where one does not want it - clogging the gap ,
and where if anything is useful it is only the specific chosen viscosity Ferro-fluid one wants there in some tweeters , and nothing there in others.

Better than candle wax to damp a metal dome will be a version of the Silicone adhesive filler that is used to seal around metal roof panels and metal pipes.
Ideally the formulation that is for outside roofing metal , because that is a high-temperature stable type , and very long life.
There are several formulae of these Silicone sealants.
Some are specifically for Plastics , some for Ceramics and Masonary { concrete cement } , some for Rubber , and some for Metals , etc ...
Most types will bond to several different types of materials , so look at all the alternates of them in a Hardware store and choose one that includes Metal.
If it will also bond to Ceramics and/or Masonary { concrete cement } that will be useful , because:
you will only need a very tiny quantity to make two tiny centre blob smears behind the domes of your two tweeters ,
and will have almost an entire tube remaining ...
which can be used if you buy the large low DCR inductors for your crossover's woofer filter that I have posted to Reggie about recently.
Use a Ceramics/Masonary bonding silicone to hold the inductors onto a board { of either wood or a suitable plastic I posted about for Reggie }
and tie the inductors tightly with Nylon cable ties through the board.
The Silicone will damp the inductors to reduce resonance which can otherwise modulate the magnetic field around the inductors to audible consequence.
These Silicone sealants cure to flexible , and can be cut with a sharp blade ,
thus the inductors can later be removed with no damage to their Ceramic insulation that other adhesives may grip too tightly to { or not grip to at all }.


'ullo Nigel ,

that Eminence woofer has too high Fs , 50 Hz , to be useful for low bass.
There are some open-baffle enthusiasts who use recent and current model Eminence Pro woofers in the application you want to try.
I have forgotten the particular Model they use , but I will find it and post that here then.
For now , if you are looking through the Eminence catalog ,
choose a woofer that has Fs at 40 Hz or lower in frequency ... down to 30 Hz will be audibly useful.
For its Qt to be useful for Open-baffle with no Equalizer circuit needed , choose a model with Qt between 0.9 and 1.2 .
The lower the Fs , the larger the Qt can be with little audible degredation.
You seem to like under-damped loudspeakers , but anything with larger Q than a 30Hz/Qt 1.2 or 40Hz/Qt 1.0 will really resonate a lot with low bass !

Last edited by alan-1-b; 4th October 2013 at 07:11 PM. Reason: to change one word
  Reply With Quote
Old 5th October 2013, 05:43 AM   #1008
DennyG is offline DennyG  Australia
diyAudio Member
Join Date: Feb 2010
Location: Sydney
Default More LTSpice plots

I'll try putting the new plots inside a .PDF file. Also included is a plot of the combined 1+2+3 from the previous post.

Post 1007.pdf

I use that high temp silicone around a metal flue. It is very heat resistent. I measured the damaged tweeter and it looked bad so I tracked down a replacement but will attempt a repair if I can get it apart.
  Reply With Quote
Old 8th October 2013, 12:25 PM   #1009
alan-1-b is offline alan-1-b  United Kingdom
diyAudio Member
Join Date: Apr 2006
Location: Midlands
Default Looking for the Optimum resistance

Great stuff Denny !

First we see that the 1.8 ohm input resistor attenuates the signal by a fraction more than 1dB ,
thus I think it likely that when the lower DCR inductors are used in the woofer filter this 1.8 ohm resistor will not be needed in the midsrange filter.

The resistor in Series with the 3.9uF capacitor will still be needed , because that resistor reduces Resonance.
See in the 2.7 ohm versus 3.9 ohm plots:
Normally one would expect that when less Impedance , 3.9uF + 2.7 ohm , in Parallel with the driver ,
the Series inductor , 3.4mH , would filter more of the treble off ,
but here we see the 3.9uF + 2.7 ohms filters less , and the 3.9uF + 3.9 ohm filters more ...
See the wide , banana shaped area between the 2.7 ohm and 3.9 ohm plots.
That shows the broad-band resonance caused by the 3.9uF cap resonating with the .179mH inductance of the driver's voice-coil
- broad-band because the 8.97 ohm of the driver's voice-coil has reduced the Q of the resonance
-{ but plays the resonance out through the dome to the listener }.
The 3.9 ohms reduces that Q further , more so than 2.7 ohms , { and both reduce the resonance before it can be played by the dome }.
One might then think that the larger Resistance in Series with the 3.9uF cap the more the Q of the resonance is damped ... yes ...
BUT , when the Resistance there gets too large the Impedance of that Parallel arm will be high enough to reduce the amount of filtering the 3.4mH inductor can do.

What is the optimum resistance for that arm to reduce the Q of the resonance but not reduce the filtering off of the treble above the crossover point ?
From those two plots we can only conclude that the optimum is larger than 2.7 ohms ,
and that 3.9 ohms is closer to the optimum than 2.7 .

Let us try to zoom in on the optimum , but with regard to the limited resistance values readily available in 5 watts.
{ 4.3 ohms in 5 watts is only available from one seller.}

Denny , please plot two more for comparison.
Have the 1.8 ohms input in the circuit.
Use 3.4mH for the inductor and 3.9uF for the capacitor , so as to be the same as for the 2.7 and 3.9 ohms circuits ,
but plot with 3.3 ohms and 4.7 ohms as alternates for the resistor ,
because I do not know which side of 3.9 will be the optimum resistance.

The optimum resistance may be larger than 4.7 ohms , such as 5.1 or 5.6 , 6.2 , 6.8 even ,
but 7.5 or 8.2 is unlikely because when it gets close to the 8.97 ohm effective resistance of the driver's voice-coil the effect of the filter will be less.

If you don't get a maximum filtering with 3.3 ohms , then try some values larger than 4.7 , but only if 4.7 filters more than 3.9 did.

Last edited by alan-1-b; 8th October 2013 at 12:31 PM. Reason: to add some numerals and two phrases.
  Reply With Quote
Old 9th October 2013, 07:45 AM   #1010
DennyG is offline DennyG  Australia
diyAudio Member
Join Date: Feb 2010
Location: Sydney
This is a plot for the mid range circuit with a range of resistances from 2.7 to 8.2:

Just for the record the basic circuit for the LTSpice model is the first circuit in post #1006 (circuitforplot2.jpg). Note I'm using 0.34mH in series with the mid unit not 3.4mH as you have written (L4).
  Reply With Quote


celestion 66 needs Mid-rangeHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

New To Site? Need Help?

All times are GMT. The time now is 09:25 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.79%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio