11 minutes ago, Ittaku said:

No, I wasn't asking that. You were the one saying it should be +6. Where's the +6?

In the picture you posted.   LR2 drivers are at -6dB .... and the sum = 0 .....  ie where the two drivers are equal, their sum is +6dB.

 

 

Just now, davewantsmoore said:

In the picture you posted.   LR2 drivers are at -6dB .... and the sum = 0 .....  ie where the two drivers are equal, their sum is +6dB.

Sorry I thought you meant there should be a further +6dB peak on top so there should be a summed peak at the crossover frequency?

Quick LR2 overlay.... 

 

 

 

 

Just now, Ittaku said:

Sorry I thought you meant

Heh.  Yeah, I just realised too.   It's easy to say the wrong thing, and mean something you didn't, over the net.   

 

Just now, Ittaku said:

there should be a further +6dB peak on top so there should be a summed peak at the crossover frequency?

Heh, no.  ?

 

However, for your original chart with the pink line (power response) .... it should NOT be flat.    It should look like this for a properly summing filter (red).

 

 

Roger on that. We were simply talking over each other lol. Anyway it's been enlightening for me because I didn't know a resistor would have such an effect. I'll have to redesign the midrange crossover values entirely if the resistor stays, which is likely.

Here's what it looks like without the resistor... you can see what I had in mind when I originally designed the crossover, though clearly the midrange is too loud.

 

Okay so making an attenuation circuit for the midranges with 2 resistors instead of a single one in series either before or after the crossover gives a much better phase response at the crossover points.

 

 

Now to get myself some duelund resistors of appropriate values and try it out.

That's much better looking summing.   

 

Of course, now your crossover points have moved .... and while you can make the on-axis response the same as before, your off-axis responses will be different.

15 minutes ago, davewantsmoore said:

That's much better looking summing.   

 

Of course, now your crossover points have moved .... and while you can make the on-axis response the same as before, your off-axis responses will be different.

Yes I'm aware. It's a simple change though, and to try and keep the crossover points the same, control the bandpass gain in addition to the intrinsic combined higher sensitivity, and bring the phases together, would require a lot more changes than just a couple of resistors (and I'm not even sure I can yet.)

Edited July 31, 20187 yr by Ittaku

Re-re-re-rebuild. Before ordering any more resistors I kept playing till I came up with this one. I'm much happier with this when trying to meet all the requirements.

 

These guys finally arrived. Geez they're so big... compare them to the mills non-inductive I also use, and a pen for size.

 

Soldering them in shortly. Waiting on new caps from Miflex and Mundorf too, to bring it up to the next level

Edited August 8, 20187 yr by Ittaku

1 hour ago, Ittaku said:

These guys finally arrived. Geez they're so big... compare them to the mills non-inductive I also use, and a pen for size.

 

Soldering them in shortly. Waiting on new caps from Miflex and Mundorf too, to bring it up to the next level

 

  They will sound better than Mills (I have used both) - a 1ohm res in series with my ribbons, so make them a 3ohm instead of a 2ohm load.  (My current ribbon amp can cope with 2 ohms ... so I no longer need them.)

 

Andy

 

Let me introduce the most humble member of my hifi family, Mr Burn-in. The speaker cables connected to him cost 200x as much as he does, but he does his job well and saves me using 1000W and valve hours.

Edited August 8, 20187 yr by Ittaku

Tpa3116 is a pretty decent amp tho

Tpa3116 is a pretty decent amp tho

For 30 dollars it's quite insane.

4 minutes ago, Ittaku said:

9 minutes ago, Happy said:

Tpa3116 is a pretty decent amp tho

For 30 dollars it's quite insane.

Add a tube buffer and a decent PSU you’re already playing like an audiophile

On 08/08/2018 at 1:54 PM, andyr said:

  They will sound better than Mills (I have used both) - a 1ohm res in series with my ribbons, so make them a 3ohm instead of a 2ohm load.  (My current ribbon amp can cope with 2 ohms ... so I no longer need them.)

I'm still using the Mills by the way, that's what I meant when I said I use both. The Mills go in parallel since they're not in the signal path and have far less effect on the sound (regardless of what some might say.) After a day of burning in they're sounding lovely; they've taken a kind of abrasive edge off the midrange and made it sweeter yet there seems to be a bolder sounding fundamental such as on a piano note. Impressive.

 

According to the DHL I should be getting the first of my Miflex caps tomorrow. Unfortunately the bigger ones aren't available since the 8.2uF are not stock items anywhere and I could only put them on backorder, but the 0.68uF bypass ones should make an audible difference so I'll put them in without waiting for the big ones (these are for the tweeter.) The 8.2s are monsters and weigh in at 1.75kg; can't weight [sic] till I get them

Edited August 9, 20187 yr by Ittaku

22 hours ago, Ittaku said:

1.75kg

Woah.   

 

 

22 hours ago, Ittaku said:

have far less effect on the sound (regardless of what some might say.)

Assuming you mean me .... if you look carefully at what I said ....  I was not trying to say that your report of "it had less audible effect" is wrong ....  only that the popular generalisation that it "doesn't matter" is in contrast to circuit analysis which shows it does have an effect. 

 

 

You tried to argue against these facts of  how the current is distributed through the circuit and/or how voltages and currents are relevant in the speaker loop.   You are/were wrong about those things .... but that doesn't mean your report of "it had less audible effect" is in flawed .... only that there must be some interesting explanation for it.

These guys arrived today. Pen in picture for scale again. Might be a while before I get a chance to give them a try. Still waiting on backorder for the big ones.

59 minutes ago, Ittaku said:

These guys arrived today. Pen in picture for scale again. Might be a while before I get a chance to give them a try. Still waiting on backorder for the big ones.

They both measure 0.67uF and are only 1nF apart... whilst my capacitance gauge may not be an ideal oscilloscope based setup for absolute values, it's usually good at picking differences and these are virtually identical. Here's hoping the big ones are that close since that's what will really matter at 8.2uF.

Forgot to mention, I also bought these guys. They were meant to be a temporising measure to go in the woofer's crossover (in parallel) until I decided what to put in there definitively. They are a standard aluminium metallised polypropylene capacitor, but they are immersed in oil and sealed in a "self extinguishing potting compound" whatever TF that means, but they're solid AF. I eventually decided to buy the metallised PP EVO Mundorf in oil capacitors for this job but now I'm not convinced they're going to be any better than these. Mundorf may have a reputation as a producer of audiophile passive electronic components, but Miflex have actually done nothing but make capacitors for the last 60 years and this is part of their audio range with almost the same approach to manufacturing. The only difference is that the EVO oil range are effectively multiple parallel capacitors, being a roll of capacitors, but they're not as solidly reinforced as these. The advantage of parallel capacitors is lower inductance and ESR, but that's totally irrelevant in a woofer crossover -  especially in the parallel component! I'll have both to audition within the next week, but I expect I will not hear a snot of difference between them. I'm inclined to trust a manufacturer that has been making nothing but capacitors for 60 years more...

 

 

After testing them, their values are within 0.2uF of each other, so quite well matched between them.

19 hours ago, Ittaku said:

These guys arrived today. Pen in picture for scale again. Might be a while before I get a chance to give them a try. Still waiting on backorder for the big ones.

Had them burning in for over 12 hours with Mr Burn in in a random location in the crossover before soldering them in this morning. They've replaced an equivalent sized Multicap RTX which is a polystyrene and tin foil cap. A much larger Multicap PPFXS polypropylene and tin foil still does the bulk of the grunt work in parallel with these so any change to the sound needs to consider that they're not the only cap there, though much larger than a little "bypass cap". The biggest difference to the sound was a dramatic deepening of the sound stage. My setup tends to have quite a forward sound stage so this was immediately obvious for its change. Additionally there appears to be slight extension to the top end as well. I've not noticed any timbral changes to the treble per se, though, but being a reactive component with a lot of dielectric, I expect a lot more burn in to be required for them to fully open up. I don't have much past experience with PIO caps so I don't really know how long to expect, though it should be a lot less than stubborn teflon caps. It's still interesting that they stipulate the dielectric is both paper and polypropylene in oil. There's definitely nothing worse about the sound, which at this early stage is very promising. I can't wait to get the big mofos in there as well and be done with Multicaps entirely. Alas there's no indication of how long the backorder wait will be. And now for another 10 hours or so of listening; Beethoven is calling

Narrowing down on the design that fits all my criteria best and sounds the best now. Will probably build a more substantial version of this soon in pine for now. Will consider final timber once I confirm this one works well.

On 11/08/2018 at 11:39 AM, Ittaku said:

The biggest difference to the sound was a dramatic deepening of the sound stage. My setup tends to have quite a forward sound stage so this was immediately obvious for its change.

 

Interesting, Con.  But just which of your new caps brought about this "dramatic deepening of the sound stage"?

 

The brown ones ... or the black ones?

 

And where did you buy these 'Miflex' caps - I've never heard of them! 

 

Andy