Got Scanspeak Revelator 15W/8530K00 and Scanspeak Tweeter D2904/980000 - What to build?

[prasanth.nath]

Greetings.

I have an ellam 98 mk II http://www.troelsgravesen.dk/Ellam98-mkII.htm designed by Troels Gravesen of Denmark.
I made the terrible mistake of buying a JBL 305 mk II recently, because of the one-man-industry-disruptor we all know about.

To my utter horror, the ellam is thoroughly unlistenable, after listening to the 305. [Of course, I am sure the fault is mine, not the designer's]

I am not technically competent to explain why the 305 is so much enjoyable, compared to Ellam.

I was wondering how best to utilise the drivers I have. I think there are 2 paths.
1. I could try using minidsp and configure alternate cross-overs on ellam.
Also experiment and make it an active kit using the AIYIMAs of the world to drive it.
2. Or I could try out a different design itself using the same mid-woofer/tweeter.

Would the community be able to offer some suggestions around option 2 please?

Note: I have some experience using REW to measure drivers (using minidsp U-mic) and trying out cross overs in dsp.

 

[preload]

I've built a couple of designs from Troels. If it's "unlistenable," you may want to double-check everything, including polarity, crossover values, etc. Did you use the exact same crossover components/brands he spec'd, or did you substitute? Do you have measurements? I've worked with the 8530k before and a bunch of different Scanspeak tweeters before. They all sound pretty good.

 

[prasanth.nath]

Exact components.
Measurements, I presume you are talking about individual drivers? Unfortunately No. This was built by someone else for me, and he chose to make it such that is is not openable. If I have to measure the drivers, I have to break open the kit .

If you are talking about the combined response - can do that.

 

[restorer-john]

Sell the drivers, crossovers and cabinets. Cut your losses, move on and put the disaster behind you and buy another JBL 305 mk2, so you have a pair.

I made the terrible mistake of buying a JBL 305 mk II recently

 

[prasanth.nath]

. If there were a good buyer's market here (in India), I could buy an array of JBL 305 mk IIs.

 

[Rick Sykora]

A quick REW response measurement would certainly help. If much trouble, the most common problem is the tweeter polarity. This design requires it to be reversed from normal. This is often overlooked. If it is wrong, there will be a notch in the frequency response as shown in the green trace below.

 

[Jim Matthews]

I have also migrated to the JBL camp.

There's something about having a dedicated internal amp that gets the most from a given driver. Couple that with a cunning waveguide and you arrive in the same place as many of us.

FWIW - the waveguide based designs from TG approach this problem but retain the same traditional passive crossover.

Might I second the notion of selling off the drivers (and pricey caps) to fund a 705 or 708 purchase. As good as they are, JBL refined the next generation.

 

[moonlight rainbow dream]

I am not technically competent to explain why the 305 is so much enjoyable, compared to Ellam.

This is why we measure. Even quick and dirty in-room or nearfield measurement of each would help enormously in understanding what's going on.

 

[AnalogSteph]

A quick REW response measurement would certainly help.

Defo.

OP, have a look around for anything resembling a microphone that you might already have. You want something that is decently flat in the 1-10 kHz range. I wouldn't trust a webcam mic too much, and headset mics can be all over the place as well, but if you got some USB condenser job for video conferencing during the pandemic that would definitely be worth a try.

I had a look at amazon.in, and it looks like a good 3000 rupees will get you a Maono AU-A04 USB condenser mic set, which I know sounds quite good even though it's not exactly a measurement mic (frequency response takes a nosedive somewhere around 100 Hz). There have been some reports indicating that the electronics may have changed for the worse, so I would have a look at what the input level setting does and at which point it makes no sense to reduce it any further to avoid clipping (the one I bought last year can be left at full gain all the time, and actually I had to install Voicemeeter to provide some digital gain).

 

[prasanth.nath]

I have a UMIK (minidsp). Will post soon. Brushing up theory from https://www.audiosciencereview.com/...om-measurement-tutorial-for-dummies-part-1.4/ now

 

[preload]

I have a UMIK (minidsp). Will post soon. Brushing up theory from https://www.audiosciencereview.com/...om-measurement-tutorial-for-dummies-part-1.4/ now

Yes, that's perfect! You can use that to take a FR measurement using REW. Please post it.

Also, are you absolutely sure that the crossover components are exactly the ones specified by Troels? And why wouldn't you be able to open up the speaker to look inside? Aren't the drivers and the terminal cup mounted with screws?

 

[prasanth.nath]

Yes, that's perfect! You can use that to take a FR measurement using REW. Please post it.
Also, are you absolutely sure that the crossover components are exactly the ones specified by Troels? And why wouldn't you be able to open up the speaker to look inside? Aren't the drivers and the terminal cup mounted with screws?

Yes they are mounted with screws and I can remove the driver. My hands can just get inside if I remove driver. But can't see anything clearly. The cross over kit is fixed permanently to the box. The components are exact. I know 'coz I bought them from Jantzen audio directly.

 

[preload]

Yes they are mounted with screws and I can remove the driver. My hands can just get inside if I remove driver. But can't see anything clearly. The cross over kit is fixed permanently to the box. The components are exact. I know 'coz I bought them from Jantzen audio directly.

Ok, well in that case, time to measure!

 

[Draki]

The drivers are one of the finest in their respective category. If they sound as you describe them, something is wrong.
Have the exact same set of drivers, with 2K2 LR4 (acoustic) xover, sharp-edged 28x20 cm baffle (I would make them differently today, but that was then...some 15 years ago). The measurement is recent, 0-180 deg, and not half bad. Far-field only, so valid from 250 Hz up. Besides the decent numbers, subjectively they are sublime. I am really puzzled by your description.

 

[prasanth.nath]

Here are the measurements.

Mic placed approx near woofer axis, about 1 meter away. Speaker at usual listening positions - wall mounted shelf.
1. Ellam 200 to 20k
2. For comparison, jbl 305.

Chain.
Mac air --> D10 --> NAD 320BEE --> ellam 98 mk II
For jbl: D10 --> jbl

Attaching the raw files also.

Note: mic is not placed on a dedicated stand. Instead, it was placed on a small table. A drum-set is close by => plenty of reflecting surfaces. We should look at the graphs as an approximation.

Suggestions welcome.

Edit: happy to make more elaborate measurements if that helps.

 

[Draki]

Just glancing at the measurements, the tweeter is padded too much. It is down by 5-6 dB relative to the bassmid. A wrong value resistor was used, which opens the issue about the values of the other components used...

 

[prasanth.nath]

Thank you (all) for your comments.

Attached is the list of components that were used.
But, since this was assembled by someone else, it is possible he made a mistake. I guess I will have to attempt a surgery (most probably destructive).
May I ask: if I have to break it open - should I try to exactly replicate the same ellam front end (including stepped baffle), or should I do something different?

Edit: I just realized this does not really explain what the Resistor values were.

 

[voodooless]

It lists 16 resistors without any values. I don’t think Troels would like you to post the whole crossover.. Can you identify the L-pad on the tweeter: those are two resistors in series from + to -, with tweeter connection in the middle:

 

[Draki]

The stepped baffle is there for a reason: TG used LR2 topology and the tweeter set back is a part of the whole design. So yes, keep the baffle stepped (if you are keeping the exact same design).

 

[Rick Sykora]

While Lpad resistors may be wrong, unless Troels layout was used, could be simple miswiring. I had a comparable issue happen when I accidentally wired the tweeter across the entire Lpad.

Would check to see if tweeter negative wire runs to connection between R3 and R4. If connection goes to the other side of R3, may be the cause.