Newform & digital amps

[Andrew Pratt]

I'm curious what your thoughts are on this concept and if there might be an AudioWorx version of some degree possible. It seems simple enough to build with a RD ribbon on top of a bass bin of some description. Could an RD75 be used on top of a pair of M8a's in a bass bin using a Behringer as the XO? Also if the RD's aren't available anymore what about an array of smaller ribbons?

http://www.newformresearch.com/digital-amp-packages.htm

[ThomasW]

1) the Newform 'ribbons' (actually they're not true ribbons) need a 1kHz XO point. That's smack in the middle of the midrange ..... oops, bad idea.

2)cheap digital amps are just that...... and good digital amps are as expensive as good analog amps, so why bother?

3) the entire RD product line is available from our friends at PE. I received a pair of RD75s yesterday that I ordered for a friend.

4)one can certainly put any of the larger RD series planars on top of a bass bin. Look at the B&G RD Series Gallery pages to see many examples of that

[Bruce]

Quoted by ThomasW
2)cheap digital amps are just that...... and good digital amps are as expensive as good analog amps, so why bother?

What do you think about the methodology of keeping the signal in a digital PCM format until the last second before converting it to analog with a PWM digital amp/DAC process, like that used by the TI PurePath (equibit) technology?

This means I am not talking about any of the digital amps that throw in an additional A/D and D/A cycle after the traditional DAC (standalone DAC or DAC integral to a player).

I know inexpensive digital receivers like the Panasonic (based on the equbit technology) are not up to the standards of say a TACT amp (using the same technology), but it seems audiophile user responses are very positive for the Panasonic.

[JonMarsh]

Merry Xmas, Andrew!

As Thomas points out, you can add bass bins to either Newforms or RD drivers- but the point which isn't discussed often enough, IMO, is should you? Becuse, in this case, you're combining point source drivers with line source drivers.

Why is this a problem? Well, line source arrays don't follow the same square law fall off in SPL with distance as point source speakers- they fall off at about half the rate. Further complicating the matter is that whether an array acts as a line soruce is dependent on the wavelength in relation to the size of the array. This can be ameliorated by floor to ceiling arrays, which give you a virtual array height which is higher, due to the reflection from the planar boundary.

What happens in practice is that the mids and highs don't fall off as fast as the low mids and bass. Even line source speakers like the RD's have minimum working distances so that the FR balance is as intended- 10 feet for the RD50, 15 feet for the RD75. In the lower range of the theorectical bandwidth of the RDs (below 500 Hz) you can get marked variations in the response depending on baffle, distance to user, and boundary interface issues. This is part of why they work a bit more consistently and with greater ease when you cross them over higher than the 250 Hz some folks try to get away with- I prefer about 600 Hz.

In practice you have to tune the system for a range of listening distances- and to get the power response close to balanced, my opinion is you're better off with dipole midbass/low mid and diople RD- and a line source also on the midbass. You need that just to keep up with the dynamic range, considering the losses in a dipole. This is why my calculated match to an RD50 is 8 Extremis 6 per side. Less robust midwoofers wouldn't really cut it.

Even so, if I had the money and could deal with the size and weight (i.e., not having to move it), I'd prefer to have an RD75 with 12 Extremis and 12 JP2 per side. But I'm not really the gonzo neuvo kind of guy some people on this forum think I am, so I'm just going with the RD50, and will be tuning for an integration distance of 10-15 feet for the listening window.

My version will use a super tweeter ribbon array, but you could use BG's or Newforms without supper tweeters, and just the line array of Extremis. That's assuming the Extremis are clean to beyond 1 kHz, which would be necessary for the Newform. So, I suppose that would be the AudioWorx solution...

The Newform 345 ribbons are 45" tall, and should be usable in this configuration- note that they are monopole, not dipole. Cost is a bit over $1K per pair. The shorter ribbons will be a line source at higher frequencies, and start transistioning to a point source at lower frequencies.

If you have a chance to listen to one of Jim Griffin's line source projects, (like the Linus Array) which have been popular with the DIY community, you can get a quick lesson in how the tonality changes with varying distance. The Linus, IMO, has other issues, due to the integration of the planar tweeters used and the midwoofers, but that's part of the problem- getting the drivers that will work well together in that range. Jim Griffin's array paper is interesting, too, though if you're really interested, I'd go back to the work published by JBL.

JBL Paper

[ThomasW]

Bruce,

I'll defer to Jon on the variations of digital amps. He's the one with the published AES paper and was extensively involved in the design and development of the Crown K series of amps.

His opinion in a nutshell is, 'just say no to Class D for anything involving a fullrange audio signals'.

I heard the $30,000/pr Jeff Rowland tiny ICE based monoblocks at the RMAF, they sounded pretty good... :T

[Saurav]

His opinion in a nutshell is, 'just say no to Class D for anything involving a fullrange audio signals'.

So is it the bass that they don't handle well or the treble? My guess would be treble, because switching noise and other issues should show up closer to that end, but that's a guess.

I've often wondered about tri-amping my speakers, but don't really want to build another tube amp. And I agree with you, a cheap digital amp like the Sonic Impact wouldn't be of the same quality as the rest of my system. I'm not sure what an easy solution is for something that would be used only above say 2kHz. And for me, I don't need too much power either (and music doesn't have that much energy up there anyway).

[ThomasW]

For the Crown K series Jon thinks they're ok up to 1kHz, 'cheap' stuff should be limited to no higher than 320Hz and actually lower is better. IOW they're ok for subwoofers

[Andrew Pratt]

I wasn't as worried about the amp side given alternative amp choices available it was more a question on how they used the Behringer as the XO with what appeared to be a fairly simple speaker. Could the RD50 be used with your line array of Extremis and the Behringer as the XO using analog amps? My room is set up so that I'm 10 feet back from my speakers. There's an additional 5 feet behind the main speakers but only 1.5 on either side of them as the room width is only 12' wide. The M8a's I have now sound great in the midrange and image well but they lack that "air" in the upper end that I'd like to get back. Its likely mostly due to my room acoustics as even the Maggie 1.4's sounded a little muffled in this room.

[ThomasW]

Yes you can use it as just a simple (if you can figure out how to program it) crossover.

Note that the RD series drivers like all broadband planars have a fairly soft top end. This is also true of the Newform ribbons, the Neo8's etc. That's why Jon's adding a megabuck line array of the Fountek ribbons to his RD50's. And that's why I have 16 leaf tweeters/side mated to my RD75s.

Now some people are happy using EQ to boost the highest frequencies of the RD series planars. If that doesn't work the options are $$$$$

[Bruce]

My question was focused on the ability of the Behringer to do the XO while still in the PCM digital domain, but I failed to include that part.

Jon, does that mean you give the thumbs down to the TI based TACT amps?

[JonMarsh]

Hey guys,

Remember, I'm the major league luddite in this bunch at HT Guide; though I've worked on class D amps, and still do (our company is looking at some new possible business relationships) my favorite amps are non loop feedback solid state or vacuum tube.

I'm concerned about amplifiers/systems also which rely on SPDIF input and do volume control in the digital domain; at low volume settings, if attenuation is done in the digital domain, it's reducing the number of bits representing the data at the output. This is a problem with the Digital EQ and crossovers, too, and why folks are often looking for post control with analog volume controls...

If someone told me I had to switch over to digital tomorrow, I'd go with the ICEPOWER stuff. However, you can get pretty good performance with the UCD modules available from Adire, or the LC Audio Zap pulse modules.

Keep in mind, class D or digital is no panacea; you still have the issues with power supply quality, grounding, input signal noise rejection, etc, that comes with a conventional analog amp- but you also have to take some care because of EMI issues and the interaction of the output filter in a class D amp with the load impedance at high frequencies- and most of the modules I've seen have pretty primitive setups for output filtering.

See below- This is an impressively simple chasis internals for a 100 Watt/channel amp, but with those small filter caps, do you think you're going to get Krell like control of low bass?

Image not available


The K series from Crown at least have the power supplies to deliver some guts in the low end.

~Jon

[Davey]

Jon,

I'm not sure what the story is on those smallish capacitors in that amplifier. However, I see some folks going overboard also.

Here's another example of a nicely done multichannel amp using the Hypex modules: (it's not mine.)

Image not available

I'm talking to more and more people with ears I trust that are very impressed with these Hypex modules for full-range use. Hmmmm.

These are not digital of course. Sorry for moving the thread off-topic.

Cheers,

Davey.

[Bruce]

I'm still reading and trying to understand and I'm not even close to really understanding digital amps (my experience is in chips for digital data networking products).

I thought the TI chips were also of the no feedback variety.

[Andrew Pratt]

I know Rotel will be releasing a series of Digital amps in the coming year that they say sound very very good. I'm not sure what chips they are using but knowing Rotel their power supplies won't be the weak link

[JonMarsh]

Those are "digital" as far as I'm concerned; they are switching amplifiers. What varies in the type of power stage topology and control scheme; PCM, PWM, DSD, etc. These are the UCD modules which are available through Adire from Hypex in NL.

Now THAT's a pretty clean implementation for a multi-channel amp.

If I had spare time and money, it would be fun to put together a little rack mount jobbie using a Par Metal case, Plitron transformer, my own supply filter bank, and some of these... maybe sometime in 2005. Maybe not! Too many things going on.

Any one interested in switching amps ought to go back and read some of the original papers by Brian Attwood, Gerald Stanley, and the papers by Karsten Nielsen.

~Jon

[JonMarsh]

Last comment on this topic- then I need to get something useful done today!

6-8 years ago I spent a lot of time and effort working on analog solid state output stages which could operate safely into a variety of loads, including capacitve, without the customery decoupling inductor for stability. Why? Because in my experience the 2-6 uH output inductor used in virtually all amplifiers has a deletrious impact on the sonic qualities; it increases the output impedance at high frequencies, and will resonate with the load impedance at some frequency, often lower than expected in the case of capacitive loads (ESLs), and in general reduces detail and naturalness.

The results of that work were incorporated in a very limited production studio amplifier, and also used by Ayre Acoustics in the V series amplifiers. Eliminating this output inductor results in a smoother, less hard, but more detailed (in a natural way) sound.

Now, Class D amps have to have an output filter, and the inductor value is much higher than this 2-6 uH. And just like speaker crossovers, the behavior of this filter is very subject in most cases to the impedance of the load;- going from a 4 ohm load to a 16 ohm load at high frequencies wreaks havoc with the behavior of simple output filters.

In my 1990 paper on a Class D design, one of the things covered was the design effort I spent developing a cauer-elliptic filter which with the use of additional zoble components was relatively immune to loudspeaker impedance swings. Unfortunately, there's no free lunch, and extra components means extra cost and size. I don't see anyone in the current market place using such techniques.

For reference:

"A Class D Amplifier Using MOSFETs with Reduce Minority Carrier Lifetime"

by Jon Hancock

Presented at 89th AES Conference, Sept 1990; published in AES Journal September 1991.

Anyone that can handle a 3 MB email attachment, I'll mail you a copy if you'd like. There were a lot of other relatively "heroic" design efforts made for this project which were NOT disclosed in the paper, though discussed partially in an earlier paper at PCIM in 1988, including an ultra high linearity triangle generator, with ramp height programmed by the supply line sensing, giving the circuit line feed forward correction without feedback.

~Jon

[Davey]

Jon,

I'm not positive, but I believe those Hypex modules already have a Zobel on the output side of the inductor.
If not you would recommend adding one for folks who are using these to bi-amp to compensate rising inductance of the drivers? A smallish capacitor of 0.47uF or so and a resistor?

Cheers,

Davey.

[Davey]

Sorry Jon. It seems Bruno has pretty much answered this question already.

http://www.diyaudio.com/forums/showthread.php?postid=496165#post496165

Davey.

[JonMarsh]

Well, if it works as advertised, so much the better. Putting a high Q resonance within a feedback loop can be a bit tricky, though.

[Andrew Pratt]

Anyone that can handle a 3 MB email attachment

Handling the attatchement would be easy compared to comprehending what you wrote

[Jerry Pease]

very informative thread, thanks for the info guys.

[JoshK]

Or you could do like I do and run the UcD400 modules off of 8 12V7ah SLA batteries and a bank of smoothing caps. :lol:

[ThomasW]

If you want to start a new thread about DIY battery power please do so. I think you'll get some very interesting input from our resident power application engineer.

[Paul Ebert]

If you want to start a new thread about DIY battery power please do so. I think you'll get some very interesting input from our resident power application engineer.

Yes, Josh, please do!

I'd like to hear about your implementation and what those more experienced / knowledgable than I would have to say. I thought about doing this same setup a while back. Queried about it on another forum and got no response.

Also, (while I'm off thread topic) is that a DIY Teres in your avatar?

Paul Ebert

[JoshK]

Yes, Josh, please do!
...Also, (while I'm off thread topic) is that a DIY Teres in your avatar?

Paul Ebert

'Nother thread started.

Yes that is my DIY Teres. The top-o-line Teres was being previewed and had finally seen the picture when I was building my Teres' base so I decided to use that as a direction. I still haven't gotten a chance to finish it, meaning looks wise, but it runs and sounds pretty darn good.

[Yves Smolders]

I know Rotel will be releasing a series of Digital amps in the coming year that they say sound very very good. I'm not sure what chips they are using but knowing Rotel their power supplies won't be the weak link

The Rotel 7 channel digital amp uses Icepower modules - they have their own power supply. There ain't much Rotel in the Rotel here :W

Still the ICEpower modules are respectable performers.

As for UcD, well, one with a well designed PS is not a subwoofer amp... That time is long gone, wake up & smell the roses, Class-D is here and it's good, really good.

Still a bit rough with the cables & all but it's waiting for another enclosure:

With a possible 2 x 400W into 4 ohms, with 20A peak current per channel, it can handle about anything but the toughest loads. The PS used here is 500VA, 42V secondaries, 40.000uF of british slit-foil caps.

in my 1990 paper on a Class D design, one of the things covered was the design effort I spent developing a cauer-elliptic filter which with the use of additional zoble components was relatively immune to loudspeaker impedance swings. Unfortunately, there's no free lunch, and extra components means extra cost and size. I don't see anyone in the current market place using such techniques.

Well read up on the UcD design. Flat frequency response, independent from the load. Added on top of this, flat THD curve across the frequency curve. As much THD into 1 khz than into 19khz.

Allow me to quote someone from the Planar Speaker Asylum running maggies with UcD... yes... Maggies....

Why:

Stable to less than 1 ohm- very load independant
Effortless power
Smoother sounding than many SS amps
Very low distortion
...
I am currently using a UcD 400 (400w per channel @4 ohm) based amp I built with a oversized power supply and never find it lacking- I am actually looking for larger Maggies to match the amp.

[ThomasW]

Yves,

When you're a twiced published member of the AES, have published internationally more than 40 technical whitepapers on power semi-conductor applications, have been invited by Crown to help design the Crown K series of amps, have 30+ yrs experience designing digital and analog power amps, and are generally considered an authority in the power semi-conductor field, I'm sure Jon will be pleased to follow your lead. But until such time, it might be wise to refrain from telling Jon (or any of the rest of us) what we need to do.

Thanks,
ThomasW

[Yves Smolders]

Yves,

Here's friendly suggestion.....you're obviously totally ignorant of the technical credentials of the man you're lecturing. As a result, you might want to follow the advise of this old saying.....

"It's better to remain silent and be thought a fool, than open one's mouth and remove all doubt."

Wow there,

I'm not trying to lecture anyone, maybe you misunderstood the tone of my message (as my native tongue is not english)

This post was just being informative about the UcD design, Hypex UcD in particular.

JonMarsh is quoting his 1990 paper, and in the mean time, Bruno Putzeys's UcD has solved one of class-d's problems like JonMarsh already pointed out a possible solution for. Where exactly am I lecturing him? I'm just stating what the UcD is today, i'm not lecturing him anything, no instead, i'm even saying that new class-d designs are avoiding the very defects he points out.

Frankly, I'm a bit amazed, and even hurt, that after this first post on a very interesting forum, you basically tell me i'm dumb and should keep it shut - then proceed to change your post entirely since I got your first post's quote by e-mail and the tone seems even worse. I don't feel welcome at all, and i'm surely not waiting 10 years before I make a post.

Once again, there is no "lecturing" or "follow my lead" intended to anyone here, I was simply reading up on class-d / UcD related threads on the net and trying to add my little effort in letting the world know about this nice class-d design.

If the "wake up and smell the roses" part is bothering you, well, it's not pointed at you, or JonMarsh, but to the many out there that have listened to a class-d subwoofer design from 10 years back, and ignore this class of amplification on general principle. For me it's a relatively cheap way to audio bliss.

If I really came over as being someone lecturing all of you, then mea maxima culpa, that was not intended.

I will refrain from posting anything else on this forum, as I don't like the elitist version of "shut up" that I was told here, goodbye.

[Dennis H]

Yves, I think Thomas got up on the wrong side of the bed today.

FWIW, Jon said he bought some UcD modules to try out in his spare time.... we all know how much time he has to spare so I'm sure he'll get to them right away, maybe even this decade.

[ThomasW]

Yves, I think Thomas got up on the wrong side of the bed today

Actually it's a industrial strength migraine day (in a cluster of 5 straight days of migraines).

Yves, I apologize for the rude and confrontational nature of the post...the migraine meds make me WAY too testy..... ops:

Unfortunately I get tired of reading about the digital amps being the 'new paradigm'. I recently spent 3 days at the RMAF listening to several dozen of them, and am still not impressed. Obviously they sound orders of magnitude better that the designs of a decade ago. But I certainly don't want any hooked up to my Maggies, my ESL's or anything else.

Dennis is correct, Jon's had UcD modules for quite sometime. He knows want they can and can't do.

I'll bow out here, take more meds and go crawl under a rock until my eyes stop strobing between black and white patterns, and I can post in a more civil manner.....

[Yves Smolders]

Thomas, sorry, wasn't aware that this wasn't exactly your day (and migraines can be killers...)

I'll try to watch my words carefully as well, to avoid misunderstandings.

Jon seems to be a designer "God" - I'm sticking around to see what he thinks of the UcD's.

And yes, there might be a large shift towards class-d in 2006, look at Icepower designs, the Nuforces and also flying moles... but the "classic" designs will stay also - it's always been like that in audio! LP is still around, tubes are still around - if it were all hype it wouldn't last.

I'm not stepping in boat of the panny lovers - $299 receivers can be good but won't be $15.000 krell killers - however, if they perform at $299 whereas you had to pay $1299 the year before, then you already have a revolutionary shift in price/quality.

For me, the $2/watt that the UcD's cost (more or less) outperform everything at the same to double the price, *in my ears* - and that's why I chose them... Before hearing them, I was settled on a Rotel 1095, and was having budget problems... and yes, I also chose thema little bit because I had to put them together, feels more like something of yourself in the music.

Audio is still about making yourself happy with the best you can get: For one person it's portability, don't take this guys iPod away from him - for the other it's old trusted LP's while staring in the warm glow of vacuum tubes, another guy wants a 7.1 channel orchestra in his living room - or you could be the kid that wants 130+ db SPL in his car - if it makes you happy, it's right for you.

[Cowanrg]

Thomas,

hey there! i just ordered a pair of UCD400AD's a couple of days ago, before i saw the discussion on it here. ill let you all know how it turns out. ill be using them with my new BG speakers My radia 520i's should be showing up here soon too.

[ThomasW]

Hey Rob,

What about all those Pass Lab clone chassis sitting in the closet? ...:wink:

[Cowanrg]

well, im going to try out these magic UCD modules and see what it think. if they do it for me sonically, im going to use THOSE in the pass labs cases.

im just not sure i want to move forward with the aleph2's anymore. they are very demanding. they create an immense amount of heat, and im only left with 100 watts. its a LOT of work (3 years) for an amp thats less power than what im using now. sure, its class A, but i listen LOUD and have a large room and power hungry speakers.

so, lets hope the UCD's live up to the hype. But, in any event, the cases will be complete this year...

i built a prototype of the meter the other day. and it looks pretty good, check it out:

http://www.cowanrg.mesanetworks.net/projects/meters.htm

[mikec]

well, im going to try out these magic UCD modules and see what it think.

Do your modules use the AD8620 input buffer?

[ThomasW]

Meters are going to look sweet, nice job ..... :T

However, everyone will be disappointed if the chassis don't generate enough heat to fry an egg..... :wink:

[goskers]

I just finished reading the JBL paper linked by Jon. I can't say what this really has to do with nearfield arrays. It does give the basics but focuses on venue setup with different arrays which will produce certain power responses for both near and farfield.

It still seems that Dr Griffin's paper is the best I have seen for generalized information regarding nearfield array design.

[Cowanrg]

MikeC,

i didnt realize they sold them without the input buffer (yes, i got the AD versions),

Thomas,

haha, i know... im thinking about putting dummy loads on the heatsinks to warm them up, so people can touch them and say "daaaaang, you could fry an egg on these" :roll:

[JoshK]

I like meters! :T

Your's however look much cooler than mine.

[Cowanrg]

yep, meters are nifty. they add a nice touch to an amp design. plus, i like the "old-school" feel of the meter mixed with the high-tech milled metal. not my design of course, but i like it.

[JoshK]

Yeah I like the old school touch too. In my case, my meters serve the function of telling me when the battery packs need to be recharged, but they look cool as hell IMO too.

[Yves Smolders]

Cowanrg,

I'm looking forward to your UcD results! It's always nice to see someone with "good ears", and even more important, a lot of audio experience check out new technologies.

Enjoy the "ride"!

[Cowanrg]

according to jan-peter, they shipped on the 28th (even though i didnt place the order until the 29th), but in any event, they shipped in the past few days. SO, i should get them in a week or two.

im all read to test them out, i have high hopes for them, but im also not married to them either. if they just dont sound as good as they should, they are getting sold.

[Lex]

Yves, I think Thomas got up on the wrong side of the bed today.

Uh, I was wondering if he went to bed there for a bit, lol.

Unfortunately I get tired of reading about the digital amps being the 'new paradigm'. I recently spent 3 days at the RMAF listening to several dozen of them, and am still not impressed. Obviously they sound orders of magnitude better that the designs of a decade ago. But I certainly don't want any hooked up to my Maggies, my ESL's or anything else.

Amen to that statement. I've spent zero time with them, and am already tired of hearing about them, frankly... Is that fair? Well, I'm so vested in quality solid state amps, I really don't care if it's fair or not. With a Levinson 336, Proceed HPA-2, AMP5, and a Parasound 1500A on hand, one might say I am "amped up" on something.

Lex