If you print it upside down you would only need a brim or a small amount of support underneath until the walls have formed? Also using high infill ratio with a thin wall just makes the printer do a lot of very small moves that do little for strength. Add more wall loops or perimeters depending on your slicers name for them if you want it to be more solid (if the wall is only a few mm you could make it completely solid). Raising the infill beyond 50% or less if the pattern is gyroid, wastes plastic and print time unless the part is something like a big flat disc.
I did a combination of Scan Speak Ellipticor 18WE with Bliesma T34B with waveguide and it did not sound good. The sound was too obtrusive, far from the neutrality of the Ellipticor tweeter. It seems to me that the T34A is more pleasant than the T34B.
@fluid I tried slicing it in both orientations and there was minimal difference between print time as well as material usage. The conclusion I came to was having a tight shape of the first layer giving me the security of a very flat and dimensionally accurate sealing flange. That was more of a priority than the back portion being as perfect. I find that taller, thicker cylindrical prints can warp towards the top and cause a slight layer shift if things aren't that optimal as the print progresses in height.
I ended up printing my latest waveguide in separate parts, to make use of the flatness and print quality in that orientation so it makes sense. Open frame printers can more susceptible to atmospheric changes impacting the print than enclosed ones.
@mV8 I'd tend to agree regarding the Be vs alu T34 versions. I do however believe the difference in sound is more due to the dome profile/shape than material being more of an influence with on axis sensitivity gain making the primary breakup frequency be more "in your face". I found that its important to use a series notch on the T34B, otherwise it sounds to aggressive.
Crossing the tweeter a bit higher than usual helps the edgy character on the T34B. By a bit higher I mean above 2.5k, which for my purposes is lower than usual. I'm crossing around 3.2 - 4k from the M74As on my build, which isn't symmetrical based on the spread between both LP and HP. On text book LR2 slopes the two will cross at -6 dB, summing flat. Elliptical filters will offset this further, but the goal is all around reduced energy at 3.5 k
Crossing the tweeter a bit higher than usual helps the edgy character on the T34B. By a bit higher I mean above 2.5k, which for my purposes is lower than usual. I'm crossing around 3.2 - 4k from the M74As on my build, which isn't symmetrical based on the spread between both LP and HP. On text book LR2 slopes the two will cross at -6 dB, summing flat. Elliptical filters will offset this further, but the goal is all around reduced energy at 3.5 k
@tubelectron Is that a Beyma CP22 tweeter in the smaller speaker? What woofer are you using in the right speaker? Beyma?
I'm glad you were inspired to build that design with the MD60N. Its a good mid, but I wish they would have dampened the chamber a little better.
I'm glad you were inspired to build that design with the MD60N. Its a good mid, but I wish they would have dampened the chamber a little better.
In essence any driver can be modified. It's not clear to me that the MD60N can be modded in a non-destructive manner.
The D7608 can clearly be modd-ed by adding a real enclosure and non-flat faceplate, which would extend the low end and mid-end. As is, it's IEC268-5 rated power handling suggests it is capable of LR4 500Hz stock, so it will be interesting to see how much more you can improve it.
Simulation-
D7608 (stock unit) SPL with rated 100 hr RMS Noise test with 2nd order Butterworth high pass with 80W (25V applied) (IEC 18.4)
So clearly it won't be destroyed with 500Hz HP filter, but could do with some optimization to improve it's performance below 600Hz...
The D7608 can clearly be modd-ed by adding a real enclosure and non-flat faceplate, which would extend the low end and mid-end. As is, it's IEC268-5 rated power handling suggests it is capable of LR4 500Hz stock, so it will be interesting to see how much more you can improve it.
Simulation-
D7608 (stock unit) SPL with rated 100 hr RMS Noise test with 2nd order Butterworth high pass with 80W (25V applied) (IEC 18.4)
Reference:
https://www.scan-speak.dk/datasheet/pdf/d7608-920000.pdf
https://www.scan-speak.dk/datasheet/tech/Scan-Speak_Technote02_Powertest.pdf
Distortion (dBr) with ~94dB (1KHz:https://www.scan-speak.dk/datasheet/pdf/d7608-920000.pdf
https://www.scan-speak.dk/datasheet/tech/Scan-Speak_Technote02_Powertest.pdf
So clearly it won't be destroyed with 500Hz HP filter, but could do with some optimization to improve it's performance below 600Hz...
Last edited:
I do not know whether this difference in sound is due solely to the radiation characteristics. In my opinion it is not, but I could be wrong. The SBA beryllium tweeter has a similar sound, with it being definitely more directional.
T34B with waveguide is often used with 7-8 inch midrange or midbass speakers with rather low crossover, i.e. 2 kHz and below. In such configurations, this tweeter sounds simply awful. Raising the crossover cut helps a bit, but I don't know if it makes sense when you can choose a smaller Bliesma T25B for the job.
I've tested many variations of filtering. From LR2 filtering to LR4 and elliptic filters. It was better with the elliptic filter, but there is still a clearly audible signature of this tweeter. I haven't tried yet to attenuate the speaker's diaphragm breakup with a series filter, maybe this will improve something.
@tktran303 The D7608 is definitely capable of a 500 hz 4th order HP and will easily cope with 105+ dB levels from a thermal standpoint. I've already tested this myself and the results were more than promising. This big dome will dish out decent SPLs once it has a proper air space to work in.
The factory specs show a highly elevated Qts that can easily be lowered to under 0.7 with proper chambering and dampening. Again, I already figured that out via my own research, otherwise I wouldn't be bothering with it.
My standards are rather high when it comes to SPL capability. I wouldn't bother with anything that can't hang with at least 105 dB system SPL. Dynamic range capability is very important to me, especially on a 3 way system. Why would someone bother building a 3 way speaker to begin with if it can't deliver decent dynamic range?!
Please forgive me for sounding like a broken record, but the evident lack of lower mid SPL on the factory supplied FR curve is simply due to a missing rear chamber. This certainly isn't doing this driver any favors in the marketing department. This is also mainly what produces the elevated 3rd order distortion below 1k. Once the diaphragm has a proper linear spring to work with, it makes things sound much cleaner in the lower mids and even improves SPL capability in this area.
A sealed volume of air is by far the most linear behaving suspension compared to any other mechanical means of restorative force to the diaphragm. This is why high compliance acoustic suspension woofers can produce significantly more accurate sounding bass compared to more common low compliance woofers. The same principle holds true for the D7608.
The factory specs show a highly elevated Qts that can easily be lowered to under 0.7 with proper chambering and dampening. Again, I already figured that out via my own research, otherwise I wouldn't be bothering with it.
My standards are rather high when it comes to SPL capability. I wouldn't bother with anything that can't hang with at least 105 dB system SPL. Dynamic range capability is very important to me, especially on a 3 way system. Why would someone bother building a 3 way speaker to begin with if it can't deliver decent dynamic range?!
Please forgive me for sounding like a broken record, but the evident lack of lower mid SPL on the factory supplied FR curve is simply due to a missing rear chamber. This certainly isn't doing this driver any favors in the marketing department. This is also mainly what produces the elevated 3rd order distortion below 1k. Once the diaphragm has a proper linear spring to work with, it makes things sound much cleaner in the lower mids and even improves SPL capability in this area.
A sealed volume of air is by far the most linear behaving suspension compared to any other mechanical means of restorative force to the diaphragm. This is why high compliance acoustic suspension woofers can produce significantly more accurate sounding bass compared to more common low compliance woofers. The same principle holds true for the D7608.
For the 375L speaker, at left :
- Beyma CP21/F
- Beyma 8M60N
- Beyma 12B100R in Bass-Reflex configuration.
For the 475L speaker, at right :
- Beyma T2030
- SATORI MD60N
- Beyma 8M60N
- two Beyma 12BR70 in Sealed Isobaric configuration.
Yes. At the request of @profiguy, I tried to open it in order to check what was the chamber damping.
But nothing wanted to move, so I did not insisted... Sorry ! 😕
T
@tubelectron 12BR70 and 12BR100 are nice drivers for larger boxes. The CP21 and CP22 are some of the last ring radiator tweeters available. Both are good, but the CP22 is a little better
@profiguy This is such an informative thread - thank you!
I have a few thoughts/questions prompted by your comments on the D7608 - hope you don't mind.
Earlier in the thread (see here: https://www.diyaudio.com/community/threads/the-dome-midrange-thread.407105/post-7549708) you were suggesting a filled chamber of about 1 litre, with extra felt added on top of the factory felt, to increase flow restriction. Your recent work has led you to a smaller chamber, about 0.35 litres. Are you still advocating extra felt?
Second, I'm struggling to get my head around the idea that extra felt, restricting airflow, can lower the Qtc, to 1 or less, as you and others have mentioned. I'm not trying to be difficult - just to understand. According to the factory data, the D7608 has high mechanical Q (7.75) and unusually high electrical Q (2.22), giving a total Q of 1.73. My (possibly incorrect) assumption would be that mechanical measures, such as airflow restriction, could affect the Qms directly, rather than the Qes. But to get down to a total Q around 1 or less would require a massive reduction in the Qms, from 7.75 to perhaps 1.5 or less. It seems implausible that added felt could bring it down that much. No?
Also, I'm struggling to get my head round what's going on when you put a mid dome in a small chamber. When you put a woofer in a small box, you get a higher Qtc than if you put it in a large box. Isn't it going to be the same for any driver?
To understand better, I decided to try modelling the D7608 in a sealed chamber, in Hornresp. In a large chamber (more than a couple of litres), excursion peaks significantly around the 300 Hz driver resonance, reaching 0.4 mm, the linear limit, with just 2.83 volts. That certainly backs up your view that this is a tricky driver to implement!
A small chamber will bring the excursion down - but at the same time it will also raise the Fc, and the Qtc. Interestingly, a good compromise seems to be a chamber of around 0.3 to 0.4 litres. It can cut peak excursion by almost half, and it raises Fs to a bit over 400 Hz. (If you reduce the chamber volume further, you can get a greater reduction in excusion, but you push resonance and Q higher, and that gives a response peak closer to where you might want to cross.) Further reduction of excursion would have be done in the crossover, which is quite tricky, I suspect, given the Q of the resonance. Stuffing the chamber completely helps a bit, but only makes a very minor difference to Qtc, which is over 2. (This is with the chamber fully filled with Hornresp's default box lining, which has Fr of 40,000. Acording to Hornresp help, that's in rayls/cm. I don't know what specific material and density that corresponds to, but I'm guessing something quite dense?)
Does all of that make sense? I'm really not trying to argue with you - just wanting to understand what's going on. I have never heard the D7608, but you've got me on the verge of buying some to try.
@ianbo What you're asking is exactly how I looked at the concept and challenges of tackling the higher Qts and low xmax of this SS mid.
The practical world is (sometimes) fortunately not that predictable when it applies to interfacing a driver with a suitable enclosure. Luckily the SS engineers did a good job of designing the D7608 motor and giving it more practical xmax than it theoretically has doing the basic math. I've found the xmax is in reality closer to double of what its claimed in the SS specs. Secondly the Qts responds very well to critically dampening the airflow at the back end of the chamber. The high Qms is actually a good thing, as it allows a higher degree of low level detail retrieval. Its also very uncommon for a cloth dome to have this high of a Qms.
Lowering Qts using airflow restriction does hurt sensitivity, but its not in the main passband we use the driver in. Its just there to extend the area under the curve around Fs so the driver doesn't run into overexcursion. It also enables for a cleaner rolloff slope at the chosen HP point.
I honestly didn't think any of these things I mentioned were practically possible, but my experimentation proved me wrong. This is why I always advocate thinking outside the box and not letting spec sheets govern your overall attitude towards the suitability of any driver. This has happened to me several times with other drivers ie. Silver Flute W20RC38 and W17RC38, Audax TW034, HM170CO, PR170MO, Eminence Beta 8A, etc.
I also advocate paying more attention to impedance curves, as these can often tell you more about a driver than the FR or TSPs. Also, its beneficial to use your ears when evaluating a driver for use in various applications. Sometimes a driver may be unsuitable for one application, but near perfect for another. Crossover application is another area where most designers fall into a rutt. Sometimes all that's needed is a different filter topology or target frequency. I think the biggest shock I got from the actual performance performance of a driver is in fact the D7608. It looks and feels like a toy in your hands but it actually packs a punch in the real world.
The larger chamber was initially thought to be the only way to get better lower mid response out of the D7608. Luckily having attempted to use a smaller chamber was an eye opener. It also changed the way I approach other dome and compression drivers.
Sometimes bigger isn't better with rear chambers on domes and even cones. I learned that the type of chamber chosen has a massive (often unexpected) effect on the FR and other driver parameters. This is why I'm not a spec sheet jockey like so many guys who stubbornly pound their fist on the table, stating that measurments in general are the holy grail yard stick to consider when assessing the performance and suitability of a driver.
One of the biggest practical lessons I've learned from this analogy is when I first tried the Silver Flute W20RC38. It was claimed by many (Zaph in particular) this driver suffered from excessive distortion across the board. When I started to listen to the driver more attentively, I was shocked at what I heard. While it wasn't a world beater, it performed much better than expected after researching the specs and reviews. I'm not saying Zaph wasn't being truthful in his assessments, but as I've often observed with 3rd party distortion measurements, these can greatly vary as a result of how and under which circumstances they were performed. Even the type of amplifier used to drive the speaker can skew things, not to mention THD measurements can be misinterpreted in quite a few ways. Even the way a driver is clamped or orientated can change TSP measurements.
There are some statements which were made (sometimes by highly respected "experts") stating that specific set-in-stone rules apply to how people perceive and even prefer various types of distortion. That is a highly debatable topic which causes alot of arguments and brow beating, especially when people disagree with the "experts". It's especially a potential wild card when psychology is brought into the mix. That can be regarded as a highly subjective POV, depending on who you ask.
So as I finally apologize for my long rant, I just wanted to point out how easy it is to get caught up in the little details of spec sheets and measurements, having them misleed us into dismissing and a potentially suitable driver for certain applications. Sometimes things aren't the way they appear on paper.
The practical world is (sometimes) fortunately not that predictable when it applies to interfacing a driver with a suitable enclosure. Luckily the SS engineers did a good job of designing the D7608 motor and giving it more practical xmax than it theoretically has doing the basic math. I've found the xmax is in reality closer to double of what its claimed in the SS specs. Secondly the Qts responds very well to critically dampening the airflow at the back end of the chamber. The high Qms is actually a good thing, as it allows a higher degree of low level detail retrieval. Its also very uncommon for a cloth dome to have this high of a Qms.
Lowering Qts using airflow restriction does hurt sensitivity, but its not in the main passband we use the driver in. Its just there to extend the area under the curve around Fs so the driver doesn't run into overexcursion. It also enables for a cleaner rolloff slope at the chosen HP point.
I honestly didn't think any of these things I mentioned were practically possible, but my experimentation proved me wrong. This is why I always advocate thinking outside the box and not letting spec sheets govern your overall attitude towards the suitability of any driver. This has happened to me several times with other drivers ie. Silver Flute W20RC38 and W17RC38, Audax TW034, HM170CO, PR170MO, Eminence Beta 8A, etc.
I also advocate paying more attention to impedance curves, as these can often tell you more about a driver than the FR or TSPs. Also, its beneficial to use your ears when evaluating a driver for use in various applications. Sometimes a driver may be unsuitable for one application, but near perfect for another. Crossover application is another area where most designers fall into a rutt. Sometimes all that's needed is a different filter topology or target frequency. I think the biggest shock I got from the actual performance performance of a driver is in fact the D7608. It looks and feels like a toy in your hands but it actually packs a punch in the real world.
The larger chamber was initially thought to be the only way to get better lower mid response out of the D7608. Luckily having attempted to use a smaller chamber was an eye opener. It also changed the way I approach other dome and compression drivers.
Sometimes bigger isn't better with rear chambers on domes and even cones. I learned that the type of chamber chosen has a massive (often unexpected) effect on the FR and other driver parameters. This is why I'm not a spec sheet jockey like so many guys who stubbornly pound their fist on the table, stating that measurments in general are the holy grail yard stick to consider when assessing the performance and suitability of a driver.
One of the biggest practical lessons I've learned from this analogy is when I first tried the Silver Flute W20RC38. It was claimed by many (Zaph in particular) this driver suffered from excessive distortion across the board. When I started to listen to the driver more attentively, I was shocked at what I heard. While it wasn't a world beater, it performed much better than expected after researching the specs and reviews. I'm not saying Zaph wasn't being truthful in his assessments, but as I've often observed with 3rd party distortion measurements, these can greatly vary as a result of how and under which circumstances they were performed. Even the type of amplifier used to drive the speaker can skew things, not to mention THD measurements can be misinterpreted in quite a few ways. Even the way a driver is clamped or orientated can change TSP measurements.
There are some statements which were made (sometimes by highly respected "experts") stating that specific set-in-stone rules apply to how people perceive and even prefer various types of distortion. That is a highly debatable topic which causes alot of arguments and brow beating, especially when people disagree with the "experts". It's especially a potential wild card when psychology is brought into the mix. That can be regarded as a highly subjective POV, depending on who you ask.
So as I finally apologize for my long rant, I just wanted to point out how easy it is to get caught up in the little details of spec sheets and measurements, having them misleed us into dismissing and a potentially suitable driver for certain applications. Sometimes things aren't the way they appear on paper.
Last edited:
Thank you! I for one enjoy your long rants posts, which are always thought provoking. The only thing stopping me ordering up some D7608s is the fact I have too much other stuff to work on. Might do it anyway. 😆
Edit: In a chamber of 0.35 litres are you still advising extra felt on top of the factory stuff?
Edit: In a chamber of 0.35 litres are you still advising extra felt on top of the factory stuff?
Last edited:
@ianbo Yes, I advise extra dampening media in the revised chamber recommendation. It can be adjusted to needs or taste, depending on how the driver is crossed and how much overlap and heavy lifting in the lower mids it needs to do.
I can also report the durability of the D7608 after having pushed it hard and low for extended intervals. I stopped at 80W 50% duty cycle bursts of 800 hz because I didn't have the desire to waste a good (and expensive) driver for the sake of proving a point I already made. This driver has an aluminum VC former, which is why it can cope with the thermal abuse. Its sort of like what Dynaudio did with their dynamic burst signal testing, proving low power compression and distortion. A 50% duty cycle test is a more realistic scenario IMO.
I can also report the durability of the D7608 after having pushed it hard and low for extended intervals. I stopped at 80W 50% duty cycle bursts of 800 hz because I didn't have the desire to waste a good (and expensive) driver for the sake of proving a point I already made. This driver has an aluminum VC former, which is why it can cope with the thermal abuse. Its sort of like what Dynaudio did with their dynamic burst signal testing, proving low power compression and distortion. A 50% duty cycle test is a more realistic scenario IMO.
I've been thinking about aperiodic enclosures, and how they aim to flatten the impedance of a driver, using acoustic resistance. I guess that's a way of thinking about what you're advocating - an acoustic resistor at the rear surface of the motor housing, leading into a small sealed chamber. No?
Anyway back in post #258 you said:
Did you manage to put together something more on this? I can only find a few brief comments, including this in post #263:
Elsewhere you mention adding felt.
Anyway, I'd be really grateful for any further info on how to add flow resistance to bring down the Q.
Anyway back in post #258 you said:
Did you manage to put together something more on this? I can only find a few brief comments, including this in post #263:
Elsewhere you mention adding felt.
Anyway, I'd be really grateful for any further info on how to add flow resistance to bring down the Q.
Thanks. I've read that thread too, and there's just a couple of brief comments there. Here's a bit from post #8, (March '24): "I recommend removing the OE felt if you have a lighter density F15 type wool felt to replace it with. The mid needs some air flow restriction to reduce the secondary 1k peak as much as possible. Use progressive density wool inside the chamber, getting denser as you get to the back end and sides."
But his thinking has changed a bit since then, as far as I can tell. Here's post #32 (Oct '24) from that thread: "I've been finalizing a chamber design to have a CAD file (and STL file) designed for the D7608 with the help of a generous fellow member on here. This will permit the use of the original felt material already on the dome. Some very light stuffing with fiberglass, Rockwool or sheeps wool is all the dampening that is required."
(I thought, when I read that, that it meant no extra felt was needed with the smaller chamber, if you retain the original felt, just loose stuffing. But I guess I misinterpreted that.)
I'm just trying to gently push @profiguy to give us a bit more detail on his recommended damping scheme, because (a) he's warned us that the D7608 will sound poor if the damping is not done right, and (b) he's praised its potential when the damping is done right.
He's been very generous already, sharing his thoughts and experience on this driver, and no-one has any right to push him for more. But he did indicate there's more to come, and I'm asking nicely ... 😀
But his thinking has changed a bit since then, as far as I can tell. Here's post #32 (Oct '24) from that thread: "I've been finalizing a chamber design to have a CAD file (and STL file) designed for the D7608 with the help of a generous fellow member on here. This will permit the use of the original felt material already on the dome. Some very light stuffing with fiberglass, Rockwool or sheeps wool is all the dampening that is required."
(I thought, when I read that, that it meant no extra felt was needed with the smaller chamber, if you retain the original felt, just loose stuffing. But I guess I misinterpreted that.)
I'm just trying to gently push @profiguy to give us a bit more detail on his recommended damping scheme, because (a) he's warned us that the D7608 will sound poor if the damping is not done right, and (b) he's praised its potential when the damping is done right.
He's been very generous already, sharing his thoughts and experience on this driver, and no-one has any right to push him for more. But he did indicate there's more to come, and I'm asking nicely ... 😀
@ianbo Yes, there's more to come soon. I've been swamped with all sorts of stuff.
The D7608 doesn't take well to incorrect dampening. Its a fussy driver that way and needs the right back volume. I'm very surprised SS didn't bother to give this mid a proper enclosure, since it falls on its face without it. No shortcuts to be had with this mid. Luckily it's pretty durable.
The D7608 doesn't take well to incorrect dampening. Its a fussy driver that way and needs the right back volume. I'm very surprised SS didn't bother to give this mid a proper enclosure, since it falls on its face without it. No shortcuts to be had with this mid. Luckily it's pretty durable.