OA-14: A SEAS ODYSSEY…

[arb]

Hello there,
A couple of 18cm Seas drivers having been suggested as possible replacement to SC165 in OA-14 (CA18RNX and ER18RNX), I wanted to start exploring this alternative route, and getting more information from the Carlsson community, for two principal reasons:
- I think that the probability of finding a new 6,5” driver capable of properly working in an OA-14 is getting close to zero, because of two facts: the 6,5” (16,5cm) format has become obsolete, at least in Europe (replaced by 17cm or 18cm (7”) frames), also, this size of mid-woofers is employed today in relatively small box volumes, the current crop of drivers having T-S parameters set accordingly.
- Seas has been assembling high-quality drivers in Norway for 60 years (the company is presently owned by its employees): reportedly, sample driver measurements are meeting factory specifications very consistently, unlike most Chinese brands. Price-wise also, Seas seem to currently offer better value than Peerless.
Consequently, I first took a look at the low frequency (LF) behavior of three Seas drivers, with the help of computer simulations, using Unibox (but my results should be very easy to check eventually, with Bästa or your program of choice...).
I. FINDINGS
Seas H1217-08 (CA18RLY)
http://www.seas.no/images/stories/prestige/pdfdatasheet/H1217_CA18RLY_Datasheet.pdf
Designed for large box volumes, this driver turned out a surprisingly good, perhaps even the best, performance in an OA-14 type box (35L, 33Hz). Poster 1:
https://docs.google.com/document/pub?id=1S2yHo7HtaUDEDB1tGEDmLsWjlicgEK083MY_5Xz3rQg
Not content with that, CA18RLY also operated extremely well in a 35L closed box (plugging the OA-14 vent), thereby providing very linear and clean bass (better than OA-12 actually): a trick eventually handy for difficult room situations. Poster 2:
https://docs.google.com/document/pub?id=1GfB97aADyMHZa46bSx8LH9NKsaAmtaKMXsxVYYIu8Vg
The superior built (metal-injected frame), combined with the lowest price of all working replacement drivers for OA-14, make it the driver of choice for value-seekers (aka cheapskates…).
Seas H1215-08 (CA18RNX)
http://www.seas.no/images/stories/prestige/pdfdatasheet/H1215_CA18RNX_Datasheet.pdf
Designed for smaller box volumes, this driver, already familiar to Carlsson fans in OA-50.LE and Larsen 4, kind of worked in an OA-14 type box, a bit like original SC165. Poster 3:
https://docs.google.com/document/pub?id=1d-_ZKEOZal2iJh4WPPdLWgS1QvfMoAw5FhKvwtU0b2k
But this was definitively not the highest and best use of the driver: Naqref repeatedly showed that CA18RNX could really shine when performing in a 20-22L vented box tuned to 28-30Hz, closely tracking the bass performance of the (arguably) reference for Carlsson two-way loudspeakers: Scan Speak 18W8545-00 in an OA-52 type box (20L, 30Hz). Poster 4:
https://docs.google.com/document/pub?id=1-ywKJkBN01I_oZsg0XK8LspxdtQNEHLKXhpNSVnSvL8
Seas H1456-08 (ER18RNX)
http://www.seas.no/images/stories/prestige/pdfdatasheet/H1456_ER18RNX_Datasheet.pdf
It is an upgraded CA18RNX: their T-S parameters ended being very close though, and the simulated LF performances almost identical (so, ER18RNX graphs won’t be reproduced here)... Checking the mid-high frequency behaviors should really be undertaken in order to distinguish them, but that was beyond the scope of this post...
With respect to distortion performances, John Krutke (Zaph Audio) posted graphs showing about 0,20-0,25% THD in the 200-5000Hz frequency band for all three Seas drivers, and F3 distortion at lower frequencies reaching about 0,5% for CA18RNX/ER18RNX and about 0,9% for CA18RLY. These were better distortion performances than those measured by Grafpro for both Peerless SC165 and B65oa-II (a result which would be consistent with the better construction of the Seas drivers).
http://www.zaphaudio.com/6.5test
Next post, I shall try sketching some alternative implementations...
ARB

[arb]

II. SOME POSSIBLE IMPLEMENTATIONS
Easy (and cheapest) route: Seas H1217-08 (CA18RLY)
The only step actually needed in this solution is mounting the CA18RLY driver onto OA-14. Eventually turning the vented box into a closed box merely requires plugging the tube port opening (preferably inside the box)… Further experimenting with partially filled tube port is easy to do and costs next to nothing…
All three Seas drivers share the same metal-injected frame, sized 17,6cm OD (Peerless: 16,5cm), which is reported to fit (barely…) onto the OA-14 plastic baffle board. Given also the different mounting hole patterns and mounting circle diameters (16,5cm, Peerless: 15,5cm), it entails either drilling new mounting holes into the baffle plastic, or fabricating an adapter-ring in order to secure the driver onto said baffle board (I, for one, welcome any input by folks who actually did mount these Seas drivers in an OA-14 enclosure…).
More comprehensive (read: difficult and expensive…) routes: Seas H1215-08 (CA18RNX) / H1456-08 (ER18RNX)
The ultimate level of performance will be squeezed out of these two drivers by totally re-engineering the loudspeaker internals (but preserving the stock (original) exterior appearance): so, it entails a new bass driver, new tweeters (both CT62 and T22 being currently available, but the better T22 performance should rightly complement the Seas…) and a new crossover (in an earlier forum discussion about L4, Naqref did indicate that ER18RNX actually required a more complex crossover than CA18RNX).
Imagination is really the only limit to the different ways of reducing the internal volume of an OA-14 box to 22L (filling it with some kind of granulated materials, for instance…) and to the various manners of tuning the port to 28Hz (for example, keeping the port length stock (37cm), and reducing the internal diameter to 5,5-5,6cm, would tune the 22L box to 28Hz…).
The stalling point for me right now is the crossover piece: is the Sonab crossover, as revised for T22, a good starting-point? Or should it be something else?
I do welcome input from Grafpro and Naqref, and others, on this essential piece of the puzzle…
Thanks to all!
ARB

[PappaBas]

I think the normal paper cone (CA18RNX) has a nicer roll-off than the Reed-cone (ER18RNX), note the peak on the ER at 4,5 kHz?

If you can make it fit it must be worth a try?

[arb]

I would tend to agree with you: CA18RNX is probably a safer bet.
But, Naqref is the one who suggested ER18RNX, and he certainly knows better than I do...
ARB

[paa]

I believe it was I who brought the ER18RNX driver up, and Naqref responded to that.
The ER18RNX has the new motor with a copper ring, giving less inductance and less distortion, closer to the Scan-Speak SD1 motor.
Of course it needs a new crossover, but it seems that others have found filter solutions for that driver that work very well in other loudspeakers, so why would that be impossible in this case?

[PappaBas]

Yes, the ER is technically somewhere in between the prestige and the excel line.
If the peak is problematic you can always notch it down?

[arb]

paa, you did mention ER18RNX first.
Naqref is obviously familiar with both CA18RNX and ER18RNX, and his input would be extremely valuable...
In the meantime, the following document does substantiate distortion improvements between CA18RNX and ER18RNX:
https://docs.google.com/document/pub?id=1LBMsx8ystfUwFAux6m8MfiQjhGJnUwis03ubUXEF8js
So, how do we get started for a crossover schematics?
ARB

[arb]

Regarding frequency response (FR) peaks and the need for a notch filter, comparing FR for Seas and Scan Speak drivers, the latter do also exhibit one, or even several, peaks: 18W8542-00 one peak (+5dB) at 1,8kHz, 18W/8545-00 three peaks (+4dB, +5dB and +6dB) respectively at 750Hz, 1,8kHz and 2,9kHz (Seas ER18RNX having one peak (+6dB) at 4,5kHz, by comparison):
https://docs.google.com/document/pub?id=1a6uatnTcmpcULDHpVIZmuTTTMeilxuNtyYaDzN7mwSs
(the graphs don't use exactly the same format and scale, since they were merely pasted from their respective original datasheets).
The crossovers for OA-51 and/or OA-52 had no such notch filters, did they?
ARB

[paa]

I guess the LE2 crossovers by Naqref are designed to deal with the 8545 peak(s) in some way, but the older designs are probably not.

But on the other hand, forum member avr7000 has tried the Scan-Speak Illuminator in an OA51 box, and did not experience any problem from it's rather pronounced peak, since it is not aimed at the listener, he said.
Also, the ER18RNX peak is situated at a higher frequency than the 8545 peak, more like that in the Illuminator.
Maybe this is non-problem all together?

[arb]

Yep..., don't attempts at linearizing FR end up frequently adding support to the old saying: "you win some, you lose some"?
I can't say that I am familiar with OA-51 or 52 and their crossovers.
Are they a better starting point for a ER18RNX/T22 crossover than the ones designed for OA-12/14 and OA-50 :
https://docs.google.com/document/pub?id=1c3tHY9dF3q1fte_8_Ne5P3coeeKcgXhvTjnySn23QDU
(drawings lifted from Grafpro's website)?
I really need a guiding hand for the crossover schematics...
Thanks!
ARB

[PappaBas]

According to Zaph-audio that has done a construction with this driver that peak does not seem to be a problem

Distortion seems good. The Scan-Speak Revelator is probably the best one but it costs about twice as much? My guess is that ER18RNX is pretty damn good value...