Yes, if one gets parts for free... eight output tubes @ $100 each to begin with
I certainly don't mean to sound argumentative, and i agree that bolserst definitely has been the most informative person so far on this thread. I'm just having a little trouble understanding where the loss is happening. If the speakers are truely efficeint and the amp is putting out the power why are they so insensitive? I understand that many of the ideas I'm talking about have been tried and I'm sure I'm missing something, but I don't think people are understanding the synergy of the combination that should counteract may of the limitations mentioned. Of course I realize there are many problems to work out and experiments to try but that is what I live for. I have already designed and buillt many speakers that some very knowledgeable people told me wouldn't work and that would have terrible shortcommings that just weren't there. In fact the most frustrating thing about this site is all the people that tell me things I've already done are immpossible. I never discount knowledge from anyone but when people give me answers that simply tell me it won't work quit talking about it and just build what everyone else is building, I don't put much credience in it. I'm certainly not saying that you have done that here, I just mean that you shouldn't be offended if I question it whenever someone says "it won't work".
Back on topic, It just seems to me that the electrostatic should be the ultimate in efficiency due to the ultra lightweight diaphram. I keep hearing these figures of 80-90db @ 1W and thinking I have built many a horn that hits 110db on 1 Watt. So I'm trying to figure out what is keeping the electrostatic from hitting this. It seems to me that if the current ELS stadard design is a purly capacitive load, jumping the stators with a coil of the proper resistance and inductance should counteract the 6db per octive high pass slope and also reduce capacitance and if I can create a transformerless output stage I wouldn't need more that 10 Watts to creat earsplitting levels. I understand that it wouldn't be cheap and it would take some work to make it safe but from what I theorize it just might be totally worth it. Also I understand that I won't necissarily achieve a lot of gain through the horn loading but if I can eliminate some bass cancelation and treble beaming (I live alone so a small sweet spot is not a huge issue for me). I realize that a perfect speaker is probably not possible but if it isn't the constant goal serious improvment would cease.
I'm sure posting 4 times in a row is bad form but I always forget how much I have to say(talking to me in person is much worse I assure you). Just thought I'd throw this out there. If I do try to design a DD amp shooting for 10-20W and whatever voltage you folks think would be best what output tubes would you recomend? I wonder if a completely direct coupled (absolutely no caps in the signal path) amp could be done?
My attempt will be with some HVfets Stacked up.
I already have I design started that I have used on 200v
But Until I get some sort income going (it has been over a year now) that project is on hold.
I have all of the shematics of those earlier amp designs and they can be found on Roger Sanders websight Innersound.
Here are some of my designs,
http://www.diyaudio.com/forums/headphones/189984-all-discrete-class-headphone-amp.html#post2596223
jer
Last edited:
Electrostatic forces are much weaker than magnetic forces. And, as I mentioned previously, there is a maximum electrostatic force that can be generated per unit area before ionization in the gap occurs. Baxandall estimated this value to be roughly 50N/m^2. He also made a comparison with an 8” dynamic driver which could generate about 2000N/m^2. So, right there with all other things being equal the ESL would have 30dB lower peak output. Fortunately, the ESL doesn’t have to deal with the mass of a cone or voice coil, so it gains a lot of that back. In general though, ESLs need to be larger(more area) to generate the same peak SPL as dynamic drivers.
Moving from peak output capability, to the question of sensitivity. If the matter of transformer parasitics wasn’t an issue you could pretty much make an ESL’s sensitivity(SPLout for voltage in) whatever you wanted. In fact, with high enough step up ratio you could reach the maximum SPL capability before ionization with just 2.83Vrms input. In practice, transformers designed with adequate bandwidth have a limit on the possible step-up ratio. This pretty much defines the sensitivity for transformer based ESLs. As far as DD amplifiers, seems you could set the gain to whatever you wanted. The maximum SPL capability would still be the same though. Exceeding the voltage gradient(stator-to-diaphragm) and the air in the gap starts to conduct and output goes down.
See comments above on maximum output capability and sensitivity.
The crux of the issue is that ESLs have a brick wall limit on how much force they can generate per unit area that is about 40 times smaller than dynamic drivers of similar size.
Absolutely. This is on of several standard approaches outlined by Walker and Baxandall back in the 50s & 60s.
You may not need much power, but you will still be limited in the peak SPL you can create by the ionization issue I discussed above. For earsplitting levels, you will need a large area ESL. If you would like to pursue your horn experiments, you might pick up a used Acoustat. That way you only have to deal with the horns and not ESL construction issues.
From my experience, horn loading an ESL does not gain you any efficiency at mid and lower frequencies other than what you would expect from the horn acting like baffle extensions that help with phase cancellation. At high frequencies, where the radiation impedance has shifted from acting like a lump of mass to resistive, you can get some gain from horn loading an ESL. But, since the native response of an ESL is rising with frequency you won’t be needing the efficiency boost in the top octaves unless you can somehow get the mids and lows boosted even more for a flat overall response. With this in mind, using segmention instead of a horn to increase dispersion in the top octaves makes sense. Segmentation also reduces the capacitance of the impedance with rising frequency which is a nice thing for the amplifiers or transformers driving them.
Just FYI, JonasKarud posted a paper on a horn loaded ESL tweeter a while back that showed modest sensitivity increases were possible. But, the gains were all at frequencies well above where air to starts acting like a mass load on the diaphragm.
http://www.diyaudio.com/forums/plan...learn-how-build-better-esl-3.html#post2035796
It ain/t the watts, it's the volts.
Work backwards: 1:100 transformer, transistor amp pumping 100 watts into 8 ohms,,, what voltage is being delivered to the panels?
My DD amp was essentially push-pull transmitter tubes driving a large group of wire-wound resistors. Hot stuff. Oh yeah, there was an ESL panel connected across the resistors but that's a minor part of the load.
Yes, There is a very nice and large library in Ann Arbor it has been a long time since I was there ,but, It was well stocked in the electronics section.
jer
I know but I was just thinking in terms of how loud I want it. I guess I just start with what I know and work from there. I've running around like a chicken with it's head cut off lately replying whereever I get internet. When I get a chance to sit down and relax perhaps I can think this through proper like.
bolserst: I missed your big long post above somehow before. I guess that explains everything pretty well. I guess the point I was missing was that electrostatic forces were so much weaker than magnetic. And I guess I really don't understand exactly whatr is meant by the air ionizing. Is this something that is affected by the size of the gap? Perhaps instead of electrostatic I should be looking at some kind of ribbon design using aluminum foil. I am a lover of horns mostly for the shear dynamics so I don't know if electrostaics could really satisfy me, maybe I would like them for a bedroom system for sleepy time music. I'm still studing this though, I don't give up that easy. Someday I will make a breakthrough and revolutionize the audio industry, it's my goal.
When air ionizes, it turns into a conductor. Think lightning, or corona if current is limited.
The onset of ionization in air is determined by the voltage gradient or field strength in the gap. It is calculated as (voltage/gap width). If it exceeds about 3kV-4kV/mm the air ionizes and starts to conduct.
Ribbons do have advantages over ESLs as far as available driving force. But, you will run into the same issues when trying to horn load ribbons since their moving mass is also dominated by airload at mid and lower frequencies just like ESLs. You can get sensitivity gains when horn loading tweeters though, as I previously mentioned.
ESLs can be very loud and dynamic, it just takes more area than many people want in their living rooms.
You and me both!
Ribbons are great but they require massive amount of magnets and get quite costly because of them.
I love my appogee's but I can get the same qaulity if not better with my DIY ESL's as far as normal listening levels in a lighter and smaller package.
But my goal is to just try it and see how a liitle panel would perform, as the only good that can come out of it is to widen the disperssion pattern and cut down on the beaming effect with out resulting too sectioning the stator.
Granted my little panel can get screaming loud but it is no match to a good comperssion driver as far as maximum SPL is concerned.
But it does approach that of pain although I have not measured it yet.
Any one whom has built a set and has the proper drive system can attest to this as the more surface area the more Spl.
The exteremly low distortions that the driver creates is a hard thing to beat with any other driver.
Until you have actualy heard one it is hard to explain as there are alot of very good convenitional drivers availible.
My best advice is to search out a stereo shop that has a set of MartinLogans and give them a good listen and then switch to a boxed speaker and you will see what a world of difference there is.
I did this as I was building my first panels and when I got them done I was floored!
But having a good listen showed me what to expect.
I spent no more than $50 building them.
Since then I have I have spent another $50 to $100 setting up the proper drive system although not perfect and the difference was 10 fold!
jer
Ribbons are great but they require massive amount of magnets and get quite costly because of them.
I love my appogee's but I can get the same qaulity if not better with my DIY ESL's as far as normal listening levels in a lighter and smaller package.
But my goal is to just try it and see how a liitle panel would perform, as the only good that can come out of it is to widen the disperssion pattern and cut down on the beaming effect with out resulting too sectioning the stator.
Granted my little panel can get screaming loud but it is no match to a good comperssion driver as far as maximum SPL is concerned.
But it does approach that of pain although I have not measured it yet.
Any one whom has built a set and has the proper drive system can attest to this as the more surface area the more Spl.
The exteremly low distortions that the driver creates is a hard thing to beat with any other driver.
Until you have actualy heard one it is hard to explain as there are alot of very good convenitional drivers availible.
My best advice is to search out a stereo shop that has a set of MartinLogans and give them a good listen and then switch to a boxed speaker and you will see what a world of difference there is.
I did this as I was building my first panels and when I got them done I was floored!
But having a good listen showed me what to expect.
I spent no more than $50 building them.
Since then I have I have spent another $50 to $100 setting up the proper drive system although not perfect and the difference was 10 fold!
jer
So, why, I wonder, do people not mount speakers(such as ELSs, and planars) in a wall so that they radiate into both rooms bordering the wall with no negative feedback from the otherside. This would eliminate the width requirements, cancelation effects, and cabinet resonances (I guess you would still have to deal with wall resonance but I think it would be simpler.) and space issues. Just a thought.
I've heard Martin Logans. I used to work at a place that sold them(Although the idiots had them hooked up to a Technics or some such midfi receiver) and I also heard them in a shop in FL connected to a $15K Krell set up(which was breathtaking). And I think I've heard every other general type of speakers as well. I have heard nothing that beat the virtually uncompressed dynamics of horns, or the detail and emediacy of ESLs and I wanted to find a way to have both. Perhaps I shall have to sacrafice the coherance of a single driver and combine a ELS panel with a horn subwoofer. I just don't know if the ESLs would keep up with the horn dynamicly. I don't know that they wouldn't I just don't know. How big would an ESL need to be to achieve 110db down to 150Hz in say a 16'x25'x8' room, would you think? Mating 7 of those to a tapped horn subwoofer and some homegrown tubeamps, could be the ultimate Home Theater.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.