I just looked and indeed I have one, so yes will go ahead with that one too. My prediction is that it will be similar to the MFC. We shall see.
Parts Connexion got the Duelund caps that they had made for the Elsinores. Mine have been shipped out and should arrive soon:
Here is the link to them:
https://partsconnexion.com/duelund-capacitor-1-8uf-100vac-jdm-sn-cu-series-tinned-copper-foil-wpio/
Here is the link to them:
https://partsconnexion.com/duelund-capacitor-1-8uf-100vac-jdm-sn-cu-series-tinned-copper-foil-wpio/
Other than not being radial leads which have been nice, it should fit on the crossover PCB. Not a bad price either and I note the link to the Elsinore Project webpage.
I just placed an order for 2 of them for my ULD elsinores as well, so will update as well when I have them in. Thanks to Mike for getting this coordinated. Couldn't swallow the cost for a 3rd for my Hamlet center for the moment, but will see how it goes.
Did the following tests by setting it up and running one driver plus the next driver so that the conditions were the same.
First time I have attempted doing a table here. Now figured out how to use it.
What the above shows me is that the quality of the 'motor' in dynamic drivers is so important. When we look at the speaker from the outside, or looking at the cones or domes, that in terms of distortion, the most important thing is that which we cannot see. I am reminded what a renowned manufacturer told me, that most $10.000 speakers have rubbish drivers in them. They might look flash, but they are not.
The second column will be controversial to some. The parallel resistor here simulates what "current EQ" used in the Elsinores. All measurements were done using a 50 milli-Ohm current sense resistor, so the above is actually recording what the amplifier is doing, on the current side, when exposed to the nonlinear impedance of the driver. That measured distortion is reduced in the second column.
.
| 2V RMS - 0.5 Watt Nom. | IMD Distortion | IMD w/parallel 8R |
|---|---|---|
| B&W DM303 MidBass | 1.143% | 0.593% |
| Vifa P17WJ-8 | 0.792% | 0.359% |
| SB17NBAC35-8 | 0.230% | 0.113% |
| SB17NRXC35-8 | 0.226% | 0.110% |
| SB17MFC35-8 | 0.181% | 0.89% |
| Satori MW19P-8 | 0.134% | 0.063% |
| ULD PTT6.5W08-NFA-01 | 0.093% | 0.049% |
First time I have attempted doing a table here. Now figured out how to use it.
What the above shows me is that the quality of the 'motor' in dynamic drivers is so important. When we look at the speaker from the outside, or looking at the cones or domes, that in terms of distortion, the most important thing is that which we cannot see. I am reminded what a renowned manufacturer told me, that most $10.000 speakers have rubbish drivers in them. They might look flash, but they are not.
The second column will be controversial to some. The parallel resistor here simulates what "current EQ" used in the Elsinores. All measurements were done using a 50 milli-Ohm current sense resistor, so the above is actually recording what the amplifier is doing, on the current side, when exposed to the nonlinear impedance of the driver. That measured distortion is reduced in the second column.
.
@Joe Rasmussen won't the resistor halve the output therefore reducing distortion by default. Shouldn't you be measuring at the same output spl?
I received the Duelund caps from Parts Connexion today. I was able to get the 0.47mH foil inductors for the midrange on the board as well as the Duelund caps onto the board. The foil inductors have been installed for a while but with smaller caps. You can see the claritycap/Russian film that was in the circuit before.
The fact that they are radial is actually a really nice thing. They can be installed to one side of the axial footprint which allows you to run a large foil inductor on the midrange. So you can have your cake and eat it too. Axial caps with these dimensions would not fit with the large inductor... Also, if you wanted to place the cap on a small dab of silicone, that would secure it and allow you to pluck it off of the board if needed.
Because you do not dare snip Duelund leads, I simply soldered them to pads following the traces on the bottom of the PCB. So if for some reason I want to remove this cap, I can desolder the leads, then melt the solder on the tweeter pads and pull the Duelund leads through.. Very nice! And I will have the full leads intact.
It is still early but the sound is definitely very nice!
The fact that they are radial is actually a really nice thing. They can be installed to one side of the axial footprint which allows you to run a large foil inductor on the midrange. So you can have your cake and eat it too. Axial caps with these dimensions would not fit with the large inductor... Also, if you wanted to place the cap on a small dab of silicone, that would secure it and allow you to pluck it off of the board if needed.
Because you do not dare snip Duelund leads, I simply soldered them to pads following the traces on the bottom of the PCB. So if for some reason I want to remove this cap, I can desolder the leads, then melt the solder on the tweeter pads and pull the Duelund leads through.. Very nice! And I will have the full leads intact.
It is still early but the sound is definitely very nice!
Attachments
It's better to use a voltage that is safe for the driver and in this case it is a midrange IMD signal and how that means that no more than 0.5 Watt max is dissipated by the voice coil. if for example there was 3dB less sensitivity in one driver, then you would have to bump that up by two and maybe even four times and that becomes no way to have an equal playing field (and potential driver damage too). Also, what is actually being measured here is not directly the distortion of the driver, rather this is the distorted current produced by the amplifier via a 50 mOhm current sense resistor due to the nonlinear impedance of the driver.
I already knew that this nonlinear impedance could be measured, I then came across Earle Geddes in a video who said he knew how to measure nonlinear aspects of drivers electrically and with no acoustic measuring devices. I thought he was talking about using a current source and measured as a voltage directly across the speaker terminals. But subsequently realised that it would work and yet not work, why? Because the available current sources are actually rather poor. In fact they are bad. Now I believe that Earle Geddes must be referring to use a voltage source and a very good one. Like the Purifi amplifier is a SOA voltage source. Now the nonlinearity is measured as current and not voltage. And it works and of course you can use different kinds of test signals and explore the nature of distortion. I came up with my choice and got the above results.
Sorry, I just typed the above just thinking out loud. But I think this is a really interesting way to measure driver nonlinearity purely electrically.
But now it cannot be patterned because it looks like I have given it away (for free?).
The curious YouTube video is here:
Last edited:
In my experience the Satori MW16P-8 has a slightly chest cold sort of midrange character .You notice it on some male vocals and violin.It is not unpleasant but compared to say a good soft dome midrange it is there.I guess every cone material has some tonal character and that is the one I associate with that driver.
I have used MW19P-8 and not noticed that. I am not even sure what a chest cold sounds like... sorry, I couldn't resist that glib remark. 😉
Actually, the clarity I get is remarkable and that is what listeners have told me. Very open sounding. I can't imagine MW16P-8 sounding less than that?
Actually, the clarity I get is remarkable and that is what listeners have told me. Very open sounding. I can't imagine MW16P-8 sounding less than that?
The MW16P does sound good but I believe it does have that character.It is very subtle but it is there. Every cone mid/woofer I know of has some type of tonal character.Some of those are annoying [like the B&W kevlar cones or Seas Magnesium] but I would say the Satori is pretty benign.
It does have dip 1000-1500Hz.Jeff Bagby commented on that.Better a dip than a spike apparently.
It does have dip 1000-1500Hz.Jeff Bagby commented on that.Better a dip than a spike apparently.
Last edited: