high efficiency DIY ESLs or DIY Ribbons

[thadman]

Ive been researching DIY ESLs and DIY Ribbons for quite some time and have been putting them off because of the difficulty of the construction and their main benefits were negligeble to me...

Im on an extremely low budget, by most people on this forums means and would love to complete a line source. I have the resources to build a DIY ribbon (access to machine shop and neodymium magnets), but have heard that they usually end up sounding poorly and being extremely inefficient. I also have the resources to build a DIY ESL.

Im looking for something that will match the performance of the bohlender graebener RD75, without sacrificing the top octave if possible.

The only issue is...I want to have extremely high sensitivity (match sensitivity of the woofer array) and a crossover point of 300hz so I can cross directly to a stack of RS390HFs (4 of them to be exact). Id love to run it dipole and take advantage of the 6dB+ boost from the combination of the front and rear waves to drive down distortion on the bottom octave (which occurs at 1 octave below the fundamental frequency of the baffle). If I can get the midwoofers out of the equation...Ill save a lot of money ($350 for amplification + $500-600 in drivers).

I realize its very difficult to get high efficiency from a single panel, but what if I made several small panels with extremely small CTC spacing? I could make 10-12 8" ribbons/8" ESLs per side.

[JonMarsh]

This is one of those situations where I know about and have tested something someone is developing which pretty much does what you're talking about, except for going as low- and being inexpensive. NDA's prevent me from saying anything else.

I'm afraid it's like switch mode power supplies- pick any two- low cost, high density, efficient, inexpensive. Three of four is very unlikely, and four of four is out of the question.

[thadman]

I guess...what I want is dynamics not necessarily efficiency. If I can get dynamics out of a large ESL/ribbon then Ill be plenty happy.

Do you think I should pursue building a DIY ESL/ribbon or should I just cut straight to the chase and buy some RD75s? The only way Id be able to afford them would be by omitting a woofer array and crossing them at 250/300hz...

Would the RD75 still possess dynamics and (near immeasurable distortion) that are comparable to a woofer array of 16 RS125s at this low of frequencies if driven hard enough (behringer ep2500 comes to mind)?

I have a DCX2496 for crossover duties...

The 4 RS390HF + RD75 is one option...the other options would be 4 RS390HF + 16 RS125-4 + 64 ND16FA/16 BG Neo3s

[ThomasW]

Jon and I experimented quite a bit with my spare set of RD75. They sounded quite a bit better when used with 500Hz-650Hz crossover points.

[thadman]

How loud can the RD75s get without sounding strained? Do the dynamics match that of a dynamic woofer line array?

How *much* better does it sound at 600hz vs 300hz? Were you measuring them open baffle, dipole, or monopole? If you were measuring them dipole/OB what size baffle were you using?

If I installed it on a 24" (560hz) baffle, wouldnt I get a boost around 280-560hz alleviating the excursion requirements in that passband?

Hmm...that 600hz xover point presents some issues. I could go with 16 RS270S-8s but that wouldnt come close to matching the low end displacement of 8 RS390HFs.

[JonMarsh]

The RS50 and RS75 are MIDRANGE drivers- anyone representing them as anything else is either assuming low playback levels and limited high frequency performance, or lacks integrity. Just look at the radiating area, OK? 'nough said.

It's about distortion, IMO. There's a very significant difference in how "at ease" they are with a 600 Hz versus 250-300 Hz Crossover.

On a 25" dipole baffle, you'll get about a 3 dB bump in SPL at 250-300 Hz max. If someone wants to runn them that low, my recommendation would be a waveguide- but at 250 Hz we're talking about something that's going to be around a 24' radius- total width about four feet. Sorry, it's just plain old physics.

RS270 barely makes it to 600 Hz before energy storage issues are problems raise their head- but 8 per side would give you about double the LF output of an Orion system, for example (that's +3 dB). That doesn't sound like a lot, but it would probably sound fairly good if tuned carefully. Remember that the front cavity which creates the resonance that requires the notch filter at 6 kHz also limits HF dispersion- they narrow down dramatically above 6 kHz.

This is what ThomasW is running, and the best sounding setup with RG75's that I've heard- most folks who've heard it don't have complaints.

They're run between 600 Hz and 6 kHz- lower on the top end would be better, but the Pansonic leaf tweeters Tom has won't handle lower. Something like the big Founteks would be another matter- that's why I've got 16 of those sitting in my storage room. The black panels are Acoustat 1+1. They make a nice electrostatic upper bass and lower midrange. IB woofer below 100 Hz.





This system does not sound strained at levels like 106-108 dB.

[ThomasW]

This is the test baffle. The small line beside the RD75 is a pair of Neo8"s



This lead to more advanced prototypes. The aluminium lines of Fostex planar tweeters was originally part of the system seen in the pic Jon posted. They were 'retired' when multiple Technics H400 leaf tweeters were purchased.

[Dennis H]

Forum member Paul W is in the "upgraditis" mood again from recent postings. Maybe he'd sell his Wings with a DIY true ribbon. I know you can't buy anything with a long ribbon like that for less than 5 figures.

[warnerwh]

Would it work to put a pair of RD 50's flanking a row of midwoofers if you angled the RD 50's at small angle of say 15 or 20 degrees?

A few years ago wanting to build myself a killer set of speakers I did some research on the RD series drivers and they will do what I want but I don't have the knowledge to design a speaker myself. It seems like if you used two RD 50's per side you could lower the crossover to maybe 400 and get more sensitivity. This wouldn't cost all that much more than a pair of RD 75's.

Remember I'm very much a novice but the idea I had a few years ago is still stuck in my head. This after hearing two different DIY speakers using the RD 75's. In both cases I'd want to do far better than those guys did.

[ThomasW]

Would it work to put a pair of RD 50's flanking a row of midwoofers if you angled the RD 50's at small angle of say 15 or 20 degrees?

Recipe for problems with comb filtering. Getting off axis with the RD planars is problematic since they beam at higher frequencies.

It seems like if you used two RD 50's per side you could lower the crossover to maybe 400 and get more sensitivity. This wouldn't cost all that much more than a pair of RD 75's.

Unfortunately it's just won't help that situation very much.

[thadman]

You mentioned that the RS270s start to have energy storage issues at 600hz...how high could I take the RS390HFs before running into energy storage issues?

[Paul W]

As they say, been there, done that!

Since you are on a tight budget, large ESL panels may be best. As Dennis said, ribbons capable of 600Hz (without flogging them) require more money than build time. ESLs tend to be lower cost, but higher labor content. ESL cost rises very slowly with radiating area, ribbon (motor) cost rises exponentially. The LF drivers mentioned will benefit from a HF radiator that plays low so cutting corners with a low-cost ribbon will compromise SQ.

To give you an idea of what can be done at low cost, I once built segmented ESLs using reverse connected tube output transformers, surplus HV supplies pulled from copiers, Home Depot style light grids, and about a zillion feet of 30 gauge wire-wrap wire. Build 'em big, segment for HF pattern control, and you're there!

I see you're in the right area on DIYaudio. There is also a Dutch site (ESL Circuit?) with excellent information from several builders, unfortunately I've lost the link.
Paul

[thadman]

Could I build ESLs that wouldnt have directional problems at higher frequencies and mid frequencies? What about dynamics? Can they produce 110dB+ without sounding strained? How about the crossover, could they mate well to a tower of RS390HFs?

[Paul W]

Could I build ESLs that wouldnt have directional problems at higher frequencies and mid frequencies?

Pattern control is where wire stators come in. You can "segment" the panels (by electrically grouping the wires) for any reasonable pattern.

What about dynamics? Can they produce 110dB+ without sounding strained?

This is a budget project? :W Some really large arrays can supposedly play that loud, but haven't heard them personally. High dynamic range is generally a weakness for ESLs.

How about the crossover, could they mate well to a tower of RS390HFs?

If large enough to play 100-150Hz or so, yes. Think big.

[thadman]

How big is big enough? Could a 12" x 7' ESL generate the distortion levels and dynamics found with a comparable woofer + tweeter array or RD75?

[Paul W]

Soundlab ESL

Thad,
When you want 110db from an ESL, you want something on this scale; where radiating area and baffle size are complementary.

The fortunate thing about ESLs is that cost is only indirectly related to size so they are a viable alternative for great performance with a relatively small budget. Power supply doesn't change at all, and transformer cost does not increase 1:1. Time consuming, yes. Super performance, yes. Huge, yes (unacceptable size is the only thing stopping me from building a pair).

12" x 7' might beat RD75s, but it sure won't deliver a clean 110db. Go listen to the new Martin Logans that are roughly that size.
Paul

[thadman]

are diy planar magnetics possible? I assume thats what type of loudspeaker the RD75 is.

[ThomasW]

A DIY version of any speaker is of course technically possible, how realistic it is, is another question entirely.

The RD series are a push/pull magnetic planar. This means multiple lines of magnets on both sides of a piece of mylar. There's flattened copper wire bonded onto the mylar. They're made from pretty thick steel (1 RD75=35 lbs) with formed channels to hold the magnets, and they're held together with rivets. There's a warning label not to attempt self-service due to the fact that the magnets on opposite sides of the drivers repel each other.

A better choice might be to buy a pair of Magnepan MMG's ($550+shipping) or (1.6 $1500 +tax). These provide both a midrange and a tweeter. Place a line array of cone woofers beside either to take the midbass work load off the panels.

[thadman]

That doesnt sound any harder than a diy ribbon, do you have any pictures detailing the internals of a planar magnetic or something on that order?

Are they capable of dynamics similar to ribbons?

I'd really like to use 1 driver to cover the 300-20khz range, because it will allow me to use the RS390HFs as dipole midbass/bass and alleviate my budget.

[ThomasW]

That doesnt sound any harder than a diy ribbon, do you have any pictures detailing the internals of a planar magnetic or something on that order?

Nope, what I posted is what I know

I'd really like to use 1 driver to cover the 300-20khz range, because it will allow me to use the RS390HFs as dipole midbass/bass and alleviate my budget.

Everyone would like that driver, unfortunately it doesn't exist.

[thadman]

Whats the upper limit of the RS390HFs response (while maintaining fidelity)?

[ThomasW]

Whats the upper limit of the RS390HFs response (while maintaining fidelity)?

Beats me we haven't independently tested them

http://www.partsexpress.com/pdf/295-468g.pdf

[thadman]

I've already read all parts express has to offer...non-linear distortion seems to rise pretty steadily above 1200hz...and gets crazy around 1700hz. Crossing over at 400hz 48dB/octave would eliminate the 3rd harmonic of that rising peak, but that doesn't account for linear distortion/energy storage. I'm aware that linear distortion is usually a reflection of non-linear distortion...so it appears they would probably be fine but I'm a little wary of using a 15 up to 300hz...

Do you think (asking for your educated opinion) these would sound muddy in comparison to other midbass/lower midrange drivers of much smaller size (ie RS225s, XLS8s, etc)

[ThomasW]

The design illustrated by the pictures I posted, represent my best guess as to how this type of speaker should be built

[gvinson]

A quick side-note on the 600hz crossover....

For the longest time I had my line-source planars crossed at 300hz to monopole bass. At the same time, I had seen upteen references by Thomas W that the ideal crossover was 600hz for the RD75s, all of which I eagerly resisted.

After spending the last year of free time working to integrate dipole bass into my planars....I now concur. 600hz is likely perfect for the RD75. I have mine set at 520hz. Mine are not the RD75, but an earlier version (the RD57). Apparently the math plays out the same.

Thadman, perhaps I am wrong, but low-distortion and big dynamics is the whole bowl of soup for large planars. Good luck with your design. :T