It was in reference to an analogy of waves in a pond. Due to the difference in size of the waves in relation to the plate, the DML plate is almost instantly full of complex wave patterns, not like a stone in a pond that have clear ripples that almost die out before hitting the edge many seconds later.
Regarding size of DML plates, I find there is a sweet spot where they still sound tight and where the low mid isn't too uneven in the response. I guess using multiple plates could be an idea to get tight mid and treble if you still want DML for the sub region for some reason.
Not sure what problem in Tectonic graphs you refer to?
Possibly the effect of stronger transients due to larger primary driver area might be more pronouced at high levels with a quad setup, but should be noticable in home listening as well I would think. Perhaps not worth it if a single exciter gives a better reponse, but not so sure that is generally true.
Steve!
Incredible sound Mate!
crisp! Im going to try it. I've found your materials locally.
For my room I'll go with a build similar to yours.
for travel in the guitar bag I also ordered some bamboo cutting boards, Carbon fiber 6x12 0.5 1mm 2mm and 3mm
I'd like to try to get to about 100-200hz if possible in as small package as possible
to fit in my gig bag 8x10 is about as large as I can do.
Can of worms.
Incredible sound Mate!
crisp! Im going to try it. I've found your materials locally.
For my room I'll go with a build similar to yours.
for travel in the guitar bag I also ordered some bamboo cutting boards, Carbon fiber 6x12 0.5 1mm 2mm and 3mm
I'd like to try to get to about 100-200hz if possible in as small package as possible
to fit in my gig bag 8x10 is about as large as I can do.
Can of worms.
I have thought long and hard about answering this post , but have come to the same conclusion as audiofrenzy .
there are too many incorrectly this and incorrectly that , incorrectly.
If we cannot agree that two exciters will interfere with each other and that the sound from the exciter area will always be louder ,which are pretty basic principles.
Conversation seems pretty pointless.
you have your fixed ideas , I will leave you to them.
Steve.
Frubdoo.
Is it an electric or an acoustic guitar you are using this panel for ?
The weights or mounting points will help fill in the response between 100hz and 200hz.
The spine will also protect the exciter, preventing it from ripping off if accidentally dropped.
A blob of blu-tack in the centre of the coil area will also help damp the high frequency peaks, so that you can adjust the sound a little to your preference.
Be warned that wood does warp over time , I'm not sure what to coat the panel in to stop this ?
Typical Ben used perspex but this material has a rising response and is quite harsh at the top end.
Maybe others have ideas for more suitable materials ?
The problem is that the panel must be very robust to handle the amount of knocking about it is going to get.
Steve.
Is it an electric or an acoustic guitar you are using this panel for ?
The weights or mounting points will help fill in the response between 100hz and 200hz.
The spine will also protect the exciter, preventing it from ripping off if accidentally dropped.
A blob of blu-tack in the centre of the coil area will also help damp the high frequency peaks, so that you can adjust the sound a little to your preference.
Be warned that wood does warp over time , I'm not sure what to coat the panel in to stop this ?
Typical Ben used perspex but this material has a rising response and is quite harsh at the top end.
Maybe others have ideas for more suitable materials ?
The problem is that the panel must be very robust to handle the amount of knocking about it is going to get.
Steve.
Also what I have noticed is that when a second exciter is placed on the panel one would expect to get twice the volume spl level output but that is not the case which means the two exciters are interfering with each other and cancelling out certain frequencies.
Comb filtering is the result of two or more identical signals interacting at a slight delay. When two exciters are placed side by side one exciters wave will reach the opposite side faster then the other exciter due to one being closer to the opposite edge which will make a delay between the two exciters causing comb filtering. The further apart the exciters are from each other the more the delay in the signal which in turn will mean greater combing effect.
Comb filtering is the result of two or more identical signals interacting at a slight delay. When two exciters are placed side by side one exciters wave will reach the opposite side faster then the other exciter due to one being closer to the opposite edge which will make a delay between the two exciters causing comb filtering. The further apart the exciters are from each other the more the delay in the signal which in turn will mean greater combing effect.
Audiofrenzy.
Yes I noticed this problem of lack of db with two exciters many many years ago, when I purchased 40 exciters thinking I would have them in a row down the panel similar to a podium panel.
I ended up using one exciter only on each panel😁
Using two exciters in push pull mode, front of panel and back seems to give more power handling and the exciters don't seem to heatup as much( more testing needed) but there is a big problem above 2.5k as the exciter body itself is now interfering with the sound ,not the comb filtering ?
Steve.
Yes I noticed this problem of lack of db with two exciters many many years ago, when I purchased 40 exciters thinking I would have them in a row down the panel similar to a podium panel.
I ended up using one exciter only on each panel😁
Using two exciters in push pull mode, front of panel and back seems to give more power handling and the exciters don't seem to heatup as much( more testing needed) but there is a big problem above 2.5k as the exciter body itself is now interfering with the sound ,not the comb filtering ?
Steve.
With all the time you've spent on the exciter area researching different things I was wondering if this was going to come up, if that was a major issue or not.
As far as the multiple exciter and the pistonic first wave area and so on discussion, just a basic knowledge of how speaker motors work and a little common sense is all that's needed.
Toddincabo.
I noticed the exciter response problems from the back of the panel, from day one ,many years ago.
but this is to test if it is viable for a lower midrange woofer panel .
which it seems it might be, depending on panel type and exciter power.
I usually overlap the response of my panels and my TLs speakers in the 100hz and 300hz area to minimise room response suck outs.
It might be nice to do this with a panel and get rid of the box sound all together, maybe use the TLs below 40hz if the panels cannot go any lower?
Steve.
I noticed the exciter response problems from the back of the panel, from day one ,many years ago.
but this is to test if it is viable for a lower midrange woofer panel .
which it seems it might be, depending on panel type and exciter power.
I usually overlap the response of my panels and my TLs speakers in the 100hz and 300hz area to minimise room response suck outs.
It might be nice to do this with a panel and get rid of the box sound all together, maybe use the TLs below 40hz if the panels cannot go any lower?
Steve.
Amplitude is logaritmic, not linear, so your expectation is incorrect. It would be really magical if DML could get double the volume for double the elements, but unfortunately they are also bound by the laws of physics. So it will be similar to having double the speakers, where you should expect +6dB if you also double the amplification.
You are using knowledge about regular speakers and trying to apply those to DML, but they are very different in nature in that DML does not have a phase coherent signal to start with. It is emitting different frequencies at multiple places on the plate at the same time with different relative phase. You will notice that when you put several speakers together and in different positions, you do not get the same comb filtering issues as when working with regular speakers. And the same should be true when using multiple exciters on one plate.
So far I have not seen any proper measurements that show this issue you speculate about, but would be very curious to see that. Had you done any measuring?
I dont believe DML's are very different in nature. I also believe DML's have comb filtering issues its just not as severe as conventional cone drivers but it is still there.
No, I have not done any measurements, have you?
Even Dr. Jose Bertagni says the waves of two or more exciters will interfere with each other. Bertagni also explains how to reduce this interference effect. Now maybe its something other then comb filtering that causes this interference between two exciters but there is a interference effect between 2 or more exciters when used on the same panel.
Last edited:
ahhhh, I totally whiffed on that one. It would be nice to get rid of the box sound if someone can figure it out so that it still thumps hard like a conventional enclosed sub.
Last edited:
Of course if they are very different is a matter of definition. They certainly are different, especially when it comes to issues with phase and issues arising from that like the comb filtering you describe.
Altough we view regular piston speakers as the normal way to reproduce sound it is very unnatural. Natural sound for our ears and brains are phase incohererent signals. When we do use piston speakers to create strongly phase coherent signals, our ears and brain becomes confused as soon as we introduce several sources and/or reflections. This easily becomes disturbing as our brain struggles to make sense of the unusual information.
If you use a DML speaker, there can still be reflections, and the principle of phase is still in play, but the difference is that our brain will mask the information that is below a certain threshold. A DML signal will be full of "conflicting" phases and resulting comb filtering, like any natural sound. But it is so many that when we interpret the whole signal we cannot really make them out, just as we no longer can hear the individual reflections of the walls in a room as easily with DML.
I do think there are possible configurations where they could work a bit with/against each other in some frequencies, but in cluster configuration I have not noticed that. As I mentioned I have not done identical plates with single or multiple exciters, but have done measurements on different plates with both one and four exciters, and have seen Tectonic measurements, and cannot say I have seen or heard any of the issues you describe. I suspect something happens with the FR as well as transients due to large exciter area, but my subjective impression has been that it is beneficial if anything.
I'm not saying you are wrong and that multiple exciters are worth it or perfect for hifi use, just that if there are issues with a cluster configuration I can say from experience that they are very minor indeed. Neither me or several experienced audio professionals that heard them have identified any such issues, and nothing has stuck out in FR graphs at least.
Bertagni is a long time ago, and I would argue that Tectonic has a bit more and up to date knowledge to base their designs on. Did Bertagni use cluster configuration or spread out exciters? Didn't he try to use a crossover with separate exciters for treble?
If the amplifier can deliver, it will give double the power in 4 ohm, yes. But that is not a doubling in SPL. Of course SPL is usually measured in dB, so a doubling would be insane, for example going from 100 to 200 dB. If we are talking about doubling perceived volume, that is a bit subjective, but usually that is put around 10db. Only doubling power will give you 3dB increase which we perceive as a very small but noticable increase. Doubling elements as well as power will give you 6dB increase. A 6dB is a decent increase in perceived volume, but nowhere near double.
My whole point is that there is some type of interference between 2 or more exciters on the same panel whether its comb filtering or something else.
I shouldnt have said twice the spl/db but it doesnt sound like its much louder then a single exciter. Heck my panel with 4 exciters is not much louder then a single exciter.
Now when I use a single exciter on separate panels even with a smaller surface area they are much louder then when using multiple exciters on the same panel with a larger surface area.
But anyway I dont really care I just do me. LOL
I sure get more SPL out of my four exciter plates than I did with one, so you must be doing something wrong or misunderstand something. With the exciters connected in series-parallel, which I guess is the only sensible option, you still have the same impedence as a single exciter, so will get the same total power from the amplifier. But they will be able to handle 4x the power.
I can say for sure that I haven't noticed a decrease in sensitivity, and is just really amazed how my plates could keep up with 8 massive double 18 subs using Linea Research amps, just using some measly Sanway amps that should be giving 250w per plate (4000w total) according to the chinese rated watts. If what you said was true, I would have been able to do that event with 1000w of power, which I really wish could be true. So again, I would really be excited to see measurements supporting your claims.
Not everyone has to care about efficiency or power density at all, but for those that do care it is good if we are clear on what is speculation and what is based in fact or experience.
I can say for sure that I haven't noticed a decrease in sensitivity, and is just really amazed how my plates could keep up with 8 massive double 18 subs using Linea Research amps, just using some measly Sanway amps that should be giving 250w per plate (4000w total) according to the chinese rated watts. If what you said was true, I would have been able to do that event with 1000w of power, which I really wish could be true. So again, I would really be excited to see measurements supporting your claims.
Not everyone has to care about efficiency or power density at all, but for those that do care it is good if we are clear on what is speculation and what is based in fact or experience.
Last edited:
Nice link. The part starting at about 9:10 is the best part, as it shows what comb filtering would look like with white noise, or with a sine sweep in REW. That is, like the figure below, with sharp nulls at regular intervals.
I've done a fair amount of measurments with single exciters and with pairs of exciters, but I've never seen a frequency response that looks anything like that, with either single or multiple exciters.
If two exciters really cause comb filtering, such a response should appear when two exciters are used, then disappear when either is removed, right? Seems easy enough to test:
1. mount a single exciter, set the mic at a reasonable distance from the panel (1 meter or the like), and capture a frequency response.
2. mount a second exciter, and capture a second frequency response.
3. See if the second measurement looks like the figure below, while the first does not.
Right?
Whoever is willing, please try it. I'd especially be interested to see measurements that shows comb filtering with 2 exciters, as I've never noticed it myself. That said, I never really looked for it before, so I could possibly have missed it, though it seems it would have been obvious.
In the past, I think Steve has said that if he brings the mic close to the panel and moves it around, he sees "the waves move around", and cited that as evidence of comb filtering. But actually, that is only evidence that the panel is modal, not evidence of comb filtering. When micing close to the panel, there will be peaks in the frequency response at modal frequencies which have antinodes closest to the mic. As you move the mic around the panel, you move the mic closer to other antinodes (corresponding to different modal frequencies) and farther from the original ones, hence the peaks in the frequency response move to new frequencies. Hence, such "waves" are to be expected for any DML, regardless of how many exciters it has, and not evidence of comb filtering.
Eric