What's this? Amplifier Project(s)!

[cjd]

Spun off from the Dayton RS 3-way tower thread.

I have a couple chip-amp projects in the works including a X (or super symmetry) chip-amp (based on a proposal by Nelson Pass and much leg-work by Terry Aben in researching component values). Just PCB's really. And power supply.

The simpler projet is not really that much different than the offerings at chipamp.com except that I'll be leaving room for a bit more experimentation on the PCB and may be optimizing the layout some (everyone argues about what's optimized, so I could just as easily be full of it.) It would be stable to 2ohm. The goal is something that can make simple and compact monoblocks, so a 5 or 7 channel setup would be no problem at all. Somethign that could be built into a speaker cabinet, or tossed into a larger box to make a traditional multi-channel amp. Cheap simple flexibility.

The X-amp is where I start to run into trouble on the power supply. I'm not sure what I want to do here. I'd really like to get regulation in there but I honestly do not know how to manage it with the required amperage covered.

There have been some folks that went SMPS and really liked the results. No clue how to manage this myself at all.

If this thread dies a quick death, so be it. If anyone here can provide some direction and/or help, hints, whatnot, that woudl be great.

DIYAudio just doesn't really cater to my specific limited range of knowledge well. Complete idiots get help (somewhere in there I think I got beyond that, though I'm close), and there's lots of stuff out there for folks that know a good bit more than I. Me, I barely know what I'm doing.

C

[Dotay]

I can't offer any insight or knowledge but I can say that I would be interested in potentially building and amp much like what you are describing, so any info you can scrape together in this thread will not only benefit you but me as well. :T

What are you expecting the total cost of this amplifier to be if you went the big multi-channel box approach?

[Brian Bunge]

These things don't need much heat sinking do they? If not, an amp that can be attached to a 1/8" thick plate and bolted onto the back of a speaker cabinet would be pretty cool. I too, can tell I'm already way over my head, but I'd really like to learn.

[cjd]

I'd go with a 1/4" thick plate personally. I have some on a 1/8" 2" L and it gets hotter than I like (though never trouble so far on 4ohm speakers despite being a LM3875 which doesn't really like 4ohm).

But yes, quite do-able.

I don't have much by way of updated stuff here at work so I'll have to hold off posting more details.

As for the cost of the basic monoblocks, not sure yet. You can get away going unregulated if you use a larger transformer, but they'll want ~220VA or more per block in that case. Regulating may allow for a smaller transformer, but adds the cost of regulation. Cost may be in the ballpark of $100/channel, perhaps? Really hard to say at this point. For the (theoretic) sound quality, not too bad on cost but not significant savings over something like Outlaw's gear. If you add a buffer cost will be up but you could then run it straight off a multi-channel sound-card such as the M-Audio 410 for those of us with a HTPC.

The X-amp will cost more without question.

The real issue for me is what kind of sound quality is possible at the price-point. Good, but I'm not wholly convinced of the economy yet. I *think* I prefer the amp stage in my TA-2A to my LM3875 amp, but the Stasis topology it uses is Nelson Pass designed. . .

C

[Paul Ebert]

Can you either describe the X amp to me or point me to discussion on diyaudio? I just started considering building a chip amp. I NEED a cheap project to keep my sanity (I think I'm serious about that :E ) and I'm vaguely concerned about the electrolytic caps in my ancient Stax DA-80 amp.

So, I'm looking for a exceptional sounding chip amp to build. Anything that NP has had his hand in is definitely of interest.

Thanks.

Paul Ebert

[cjd]

Well, Nelson Pass's SuperSymmetry (X) topology is his own thing and he describes it in a passle of places. His proposal was such:

Image not available

Terry took it and made it work, but he went with single LM3886 chips per side which makes it really only work well with 8ohm speakers, 2ohm being tough unless you drop supply voltage a fair amount.

I decided to toss the LM4780 at the circuit and parallel the on-die chips to allow it to be quite stable at 4ohm and reasonable to throw something with lower impedance at it on a higher supply voltage. Perhaps my downfall, as the 3886 isn't *too* bad on a 4ohm if you use an 18-0-18 250VA-ish trafo per channel. Except, of course, that I often find my speakers dipping a bit lower. So, play it safe, or something.

So, a schematic of the basic circuit, and JUST the basic circuit.

Image not available

And an in-progress PCB that has room for some additional voodoo elements that I haven't put into the schematic at this point (comes in a bit shy of five inches wide ~4.65 at the moment I think?)

Image not available

What do you call cheap?

The original thread can be found here.

C

[Brian Bunge]

Chris,

Can you get 7 of them in a rack mountable enclosure?

[cjd]

Yes. I can.

Can you figure out how? Power supply would be tricky in this case, as you really don't want (I think) to stack toroidal transformers, and you'd HAVE to with this as they may take 500VA each (regulation may allow smaller transformers per).

C

[Paul Ebert]

What do you call cheap?
C

Oh, say $400 or less. After December the financial constraint will, good Lord willing, go away, but I'm not sure I can wait that long. I've been considering building Mauro Penasa's 'audiophile' LM3886 amp (see
http://www.diyaudio.com/forums/showt...threadid=54571 )

Got a suggestion?

Thanks for the info.

Paul Ebert

[JonMarsh]

Chris- it looks fun. Pass's super symmetry topology is pretty cool, and developing an offshoot for chip amps is interesting. With any appreciable current, you'll find PCB layout and design to be pretty important, because of the way high current traces can interact with low level signals. Don't be surprised if you have to try a few layouts before you get one that reaches the circuits real potential. Testing with a basic IM analyzer will show this up... ought to be able to do that on a PC with a two tone test signal (NCH tone) and one of the audio spectrum analyzer software with a good audio card. (No, I don't do it that way- old Luddite that I am, I use an HP8903 Audio Analyzer).

~Jon

[cjd]

Paul:

Mauro's topology is most interesting! Really slick stuff, and it breaks a LOT more from the classic chip-amp. A particularly adventerous soul might actually use Mauro's work and combine it with Nelson's work. It should be very much do-able, since each side of the SuSy amp is not really anything significantly different from the datasheet model.

Jon: That is my fear. Because that would be the point at which I would get almost completely lost for a while and need assistance until I once again understood what I was doing. Or maybe the point at which I would give up on a PCB and wire point-to-point. I still consider that, as it would save the expensive of a PCB. Headache vs. cost and potential headache. Hmmm. You see my dilemma.

C

[Dotay]

Any updates on this?

I've been thinking about building an amp eventually so I'm trying to do more reading on the subject. Is there any place I can read about the differences between the different amp technologies (chip amp vs. solid state vs. whatever)?

[cjd]

First things first.

I *will* be getting to this project, and possibly soon. Two weeks left on this gig and I'm out, with nothing lined up yet. So I'm looking at some potential free time.

But, it will for sure get rolling as the really cold months sneak in because I simply can not work in the garage (my workshop) at that point.

Chip-amps are solid state by the way. They're just putting full circuits onto a chip instead of discreet.

diyAudio has a bunch of info, but it can get thick with opinion and disputed "fact" and all sorts of things. But they cover many realms.

I think I will be starting with a simple parallel circuit, do some testing, and then double up plus the "X" circuit, more testing. I may end up moving from the LM4780 to pairs of LM3886 if thermal issues point me that direction. Won't know.

Also still not sure what I'll be doing power-supply wise when it comes to more than two chips per supply.

C

[Dotay]

diyAudio has a bunch of info, but it can get thick with opinion and disputed "fact" and all sorts of things. But they cover many realms.

Yeah, I've tried to read through some of the threads but most of it goes so far over my head I really don't feel like I'm learning anything, and especially when people are always claiming different things it's hard to find the good designs. Then the seemingly good threads have 50+ pages to them so finding info can be quite tedious. Is there a more noob friendly place to learn about DIY amps or is diyaudio pretty much it?

[Saurav]

There are a couple of personal websites that cover the basics fairly decently. Nothing at the level of what Chris is trying to design, but if you're starting with a simple schematic and want some help with figuring it out, there are some websites out there that can be helpful.

Try this one for chipamps, it's written with beginners in mind (IMO). So it glosses over some details and the experts would probably disagree with some points in there, but for the most part, it's a decent introduction.

http://myweb.tiscali.co.uk/nuukspot/decdun/gaincloneindex.html

[jdybnis]

Try reading some of the oldest threads. At least the beginnings of the ones with 300+ posts. They are more informative than the recent stuff. Many of the posters have been building gainclones for years. There isn't much left to say about such a simple circuit. At this point they're arguing over how many angels fit on the head of a pin.

[Marzen]

Super Symmetry Amp

I'd be interested in a 5 channel version with buffered ins. I'm afraid I wouldn't be much help in the design stage, hi-fi specs & amp topology aren't really my bag. I could help with: pcb routing as I have an obsolete version of Protel at home somewhere, general power questions, emission testing, cnc rack case fabrication & black anodizing for case & heatsinks. I'm not terribly squeezed for time at work right now, this is typically a slow quarter.

...Or maybe the point at which I would give up on a PCB and wire point-to-point. I still consider that, as it would save the expensive of a PCB. Headache vs. cost and potential headache. Hmmm. You see my dilemma.
C

No need for P2P wiring if you'd rather have a pcb. I did a control board last year where I got 5 experimental pcbs (max size each 8" x 11") for $88. I might be able to sweet talk someone at work into doing the cad work for the Juki surface mount placement, wave solder for thru hole I can stay late & do myself.
Some initial questions:
1) How many pcb layers were you thinking of?
2) Any shielding requirements?
3) Any smt parts or all thru hole? (I have most of the parts in smt stock)
4) Can you use a separate outboard pwr supply?

Anyway, it sounds like an interesting project, as you said - cold weather is comin', more indoor pursuits soon.

[cjd]

I'm working with two-layer layouts. I probably need to make the switch to Eagle at some point here, but ExpressPCB is just so simple. From that company a pair of prototype boards is ~$98 shipped so not bad at all. Of course, your pricing is much better! 8x11 would fit six of these as the design currently exists I think.

Parts, this is designed for all through-hole. I think that for most DIYers doing surface-mount stuff by hand is too much a headache. Besides, most of the stuff you could surface mount isn't the expensive stuff anyhow.

Shielding requirements: I don't know. On the PCB's there is some need to route well, but I'm not aware of any need for specific shielding. In cases, many people prefer to shield the amp modules from the power supply section.

This is designed for a separate power supply, whether it be all in one case or not.

I have, however, had an inspiration looking at someone elses project though, with regards to my approach. I think I may make a more flexible "simple" ~100-120W with a single LM4780 on it (already planned) and leave room for the additional resistors needed for the SuSy topology. THis gives me one board that is used for MANY situations that could even be "stacked" in the SUSY configuration for more power. The trick: For the "mirror" side, the board simply gets flipped (and at least the chip gets mounted "under" the board).

That leaves me without a clear idea what to do in the power supply still. Power consumption I think is recommended to be ~5A per LM4780 (or pair of LM3886). The LM338 is, of course, just fine in this case as a simple adjustable regulator - however, it alone won't support two chips, and the schemes I've found for paralleling use obsolete chips. I'm not sure if this could be balanced as simply as the actual amp chips being used (0.1ohm resistor on the output) and no servo mechanism, or not. For that matter, one coudl power each chip separately leaving the "X" circuitry unpowered which can probably leech somewhere... but ... see, I just don't know.

Unregulated I think I know where I want to go. But I think both options should be tried.

C

[Marzen]

... At this point they're arguing over how many angels fit on the head of a pin.

:rofl:

Ok, I just got back from Nelson's amp page. I read the notes on the super symmetry topology. Nice, neat, & concise notes. I actually work a lot with differential opamp ckts & I like the simplistic 'less is better' design approach.
I followed the link to DIYaudio & was instantly struck with a sense of 'this is out of control'! I get the same feeling with engineers from NASA & Sandista Nat'l Labs. I'll have to go back & wade thru the muck to find the meat & potatoes. This is where the production side of my brain kicks in with some basic, realistic design considerations:
1) Can I do it better AND/OR cheaper than existing commercial products?
2) Compared to an existing commercial product of similar specs, what's my cost to replicate end product. (Personally, saving 30% is no bargain to me. Sometimes it's impossible to overcome their price break on components)
3) Is there a completed design with published specs that I can copy, or at least use the majority of the proven circuit for my implementation?

CJD, how does the x amp meet with the above. You seem to have done a lot of foot work already which begs the question: Why this one?
Here's my requirments:
1) 5 or 6 channel power amp w/ better specs than your standard HT rcvr.
2) Lower cost than a comparable brand name unit (> 30% savings)
3) Low design time

Disclaimer: No offense to die hard diy'ers, I need to show a profit somewhere. Ditto for big budget engineers with extensive development timelines.

[Marzen]

I'm working with two-layer layouts. I probably need to make the switch to Eagle at some point here, but ExpressPCB is just so simple. From that company a pair of prototype boards is ~$98 shipped so not bad at all. Of course, your pricing is much better! 8x11 would fit six of these as the design currently exists I think.

Those were done in Taiwan, shipped over on the proverbial slow boat from China (3 months). Add ~ $25 for air.

Parts, this is designed for all through-hole. I think that for most DIYers doing surface-mount stuff by hand is too much a headache. Besides, most of the stuff you could surface mount isn't the expensive stuff anyhow.

Ugh, I'm gonna sound REALLY self centered here when I ask 'who cares'? I SHOULD ask - do you want to release a thru hole pcb cad design for this into the public domain, or do you just need one for yourself? If the later, I can put the parts on at work, and really keep the size down with a 2 sided pcb.

Shielding requirements: I don't know. On the PCB's there is some need to route well, but I'm not aware of any need for specific shielding. In cases, many people prefer to shield the amp modules from the power supply section.

I asked about shielding 'cause you mentioned others taking the SMPS approach (using PWM controllers?). Did the others mention anything in regard to having RFI probs as a result? I thought they were noisy.
Trace routes are already well established for power ckts, no real problem there (Jeez, I wish auto rout actually WORKED). I also wonder if the 3886 is in my Protel library.

This is designed for a separate power supply, whether it be all in one case or not.

I like the idea of a separate outboard pwr supply with the chip amps on one larger pcb in its own case. YMMV

I have, however, had an inspiration looking at someone elses project though, with regards to my approach. I think I may make a more flexible "simple" ~100-120W with a single LM4780 on it (already planned) and leave room for the additional resistors needed for the SuSy topology. THis gives me one board that is used for MANY situations that could even be "stacked" in the SUSY configuration for more power. The trick: For the "mirror" side, the board simply gets flipped (and at least the chip gets mounted "under" the board).

Could you post a link for this one?

That leaves me without a clear idea what to do in the power supply still. Power consumption I think is recommended to be ~5A per LM4780 (or pair of LM3886). The LM338 is, of course, just fine in this case as a simple adjustable regulator - however, it alone won't support two chips, and the schemes I've found for paralleling use obsolete chips. I'm not sure if this could be balanced as simply as the actual amp chips being used (0.1ohm resistor on the output) and no servo mechanism, or not. For that matter, one coudl power each chip separately leaving the "X" circuitry unpowered which can probably leech somewhere... but ... see, I just don't know.
Unregulated I think I know where I want to go. But I think both options should be tried.
C

The X-amp is where I start to run into trouble on the power supply. I'm not sure what I want to do here. I'd really like to get regulation in there but I honestly do not know how to manage it with the required amperage covered. There have been some folks that went SMPS and really liked the results. No clue how to manage this myself at all.

Lemme pull down the data sheets from digikey & have a look. You should know I'm not a 'ground up' kinda guy. I'm more like 'hey, let's "borrow" from good existing designs and incorporate it without causing a patent infringment'. Anyway, post what info/links you have and let's see if someone else has already done the dirty work for us! Jeez, I need a beer after this. I usually just look at pictures & check the weather on Sat morning :E

[cjd]

Alright. I'll try to answer your questions here.

The basic chip-amp is almost *too* simple. A few resistors, some caps, the chip, and some power. That's about as much as I can handle from a design stand-point. Good for DIY efforts.

Cost, a lot of that is actually in the power supply and sundry components (cables, connectors, etc.)

I am not sure you could be more cost effective than something like the UCD modules at this point. But they're not SuperSymmetry either. For that matter, I'm not sure it would be cost effective compared to stuff like the big Outlaw units. I really waver on whether I want to continue down this road or not.

SuperSymmetry gives some common mode rejection above and beyond what you get anywhere else - so really, it's a "how cheap can I get SuSy, or can I do without it just fine."

Compared to a monolithic supersymmetry amp, reports came back that the chip-amp based version lacked a little bit - my read was that it had higher contrast, lost subtlety. Some tweaks have happened since then, but I don't know to what result.

For me, the "fun" of doing the work, putting it together, is what makes it worthwhile. Not necessarily cost savings - obviously, if I were to consider my labor I doubt very much it would ever really be the way to go.

I'll ponder surface-mount. It does have the advantage of not needing through-holes, which means fewer troubles routing.

A simple paralleled chip-amp (non-balanced input, ~120W@4ohm) would be something like this. You can, by the way, simply parallel till you're blue.



1.5 inches by two inches as pictured. Probably would get a bit bigger to allow for larger caps at the chip. 3 critical resistors per "channel", mute, in>g, then zobel and output ballast resistors. The circuit is pretty much straight out of the datasheets.

C

[JoshK]

You'll excuse my total ignorance. I am familiar with the chip amp threads over at diyaudio and have a couple friends who have built some, but I haven't read through them, nor do I understand what SuSy really is or what it brings to the table, but I have heard of it and I know it is all inspired by the XA series amps.

I have been intrigued lately with the thought of building a pair down the road. Then I saw the discussion of the Aleph J and the JFET topology. I like the *idea* that it is made to emulate the linearity of a triode. In my relatively short audiophile history (I am only 30) I've owned an ATI 1505, Cary V12i, Gamut D100 (actually just borrowed for a really long time), Spectron Musician II, Perreaux R200i and a host of other SS amps for my HT system that I tried out in my main rig. I have UcD amps now. I have also compared some really uber expensive hybrid Blue Circle that was clearly on the top of the aforementioned list.

I think from my limited experience there is some things that the tubes do that the other technologies don't, and not just euphonically. Whether or not a good discrete class A no feedback SS amp could do this too I don't know, I haven't heard one. But from what I have heard there is a sense of layer removed from the music and a more connectedness. Of course there is a slew of things I didn't like about some of the tube amps I have owned and heard and ultimately made me sell them. My guess is that linearity, or at least from reading other's thoughts, has to do with the missing ingredient.

Has anyone analyzed the harmonic distortion spectrum of these chip amps and examined what harmonic products they produce, what IM is like and how linear they are? I ultimately don't care what the technology is, tube, ss, chip, switching, so long at the end product has all the goods.

I think it would be intriguing to hear an Aleph J amp and see how that one stacks up, but I am afriad that 25wpc isn't going to be enough for the Modula MTMs to rock out.

Ultimately I am building the Arvo Pärts to be fully actively amped, or at least biamped. The UcD's hardly have any competition in the bass, especially for the price so they will be the woofer amps no matter what, but I'd be curious for mid and tweeter amps to play around with the chip amps, class a amps, etc.

[Marzen]

Yeah, the Hypex UcD modules were the start & stop of my search for diy amps. Everything you want in diy except the cost ($150 x 5 ch = $750 +). I'll check Global Sources for cheaper modules direct from Taiwan.
I spent an hour back at diyaudio on the chip amp section (avoiding the pass labs forum :roll: ). I saw some designs that scored well on test measurements although I didn't see a total cost on them however. Nor have I been able to find any comparisons to commercial amps. I'm looking for more of a Heathkit approach. God, there are some big toroids on that site, are El's dead?
I get the 'fun' aspect, at least for speaker design. Power amps aren't so interesting to me. Fab'ing a cool case and assembly would be nice 'tho. I'll PM you my work number, call me during the week as I've been on the pc too much lately. There's always a better tool for cat skinning out there...

Cheers man!

[cjd]

The 47labs is rather famous for their chip-amp (the GainCard and they have another now too - Suguraki or something).

6moons has some reviews of some chip-amps (47labs and otherwise)

They have potential. Lots of potential, really.

I know for a fact that a 5 channel@~120W could be done for less than $400 quite comfortably. If you go the chipamp.com route, their dual mono kits are $55. Add adequate transformers, wire, power switch, plugs, RCA jacks, terminal jacks, and you're set.

With the chip-amps many folks go with very little capacitance and no regulation - so the toroids get very big rather fast since they're basically playing BOTH roles. As to El's, I think they're just difficult to come by in appropriate sizes and all that.

I don't care for the circuit design part. PCB layout is fun, because it's a puzzle. Populating a PCB is nothing - a minor headache, no more. Case building, fun again.

Image not available

I now have feet on the top part as well.

C

[capslock]

No idea why your shoes need oil, but how about:

Meine Schuhe brauchen Öl

or

Meine Schuhe wollen geölt werden

or

Meine Schuhe müssen geölt werden

[Dennis H]

Well, since they are wooden shoes, it should really be in Nederlandse taal. Where's TacoD when we need him?

[Marzen]

No idea why your shoes need oil...

Hahaha! Because they SQUEAK? My wife says I have about a 4 yr old childs grasp of German. Growing up in Mississippi, my English probly ain't none better <sic>. The avatar is Erwin Veltkamp. He kinda looks like me, and I built my first HERMS system thanks to his web page (i love the passed out naked blond on the picnic table). Great sense of humor & very detailed guy. You can find him here at http://www.barleys.nl/?hi
I hope he wouldn't take offense in my using his pic. I have a lot of respect for the man.

[Marzen]

I sent your X amp schematic with the LM4780 over to the asst. dean of engineering at a local school. He looked at it and said something like, 'what kind of industrial production equipment do you keep in your living room? This is a little over the top for noise cancellation, not to mention the large transformer you'll need.' But he did recommend a regulated pwr supply on it. However he didn't answer my question of whether a pi filter + zener would suffice after the bridge rectifier as 'enough' regulation.
I was told he's really into xo design so I thought I'd shoot Jon's modified CE filter & response charts over to him on Wednesday. He does like to use xo examples in the filter portion of the course.

[cjd]

Hehe...

Over the top? Or just barely sufficient.

I must apologize for being rather not here on this at the moment - my days are full of last minute project work and talking with recruiters trying to line up my next job.

The realization that *nothing* appealed to me right now has pretty much made up my mind to take the one offered (where it's just the barest of bare portions of my skillset - indeed, it's like being a carpenter and getting a job where all I'm doing is hammering nails all day)... good company, some potential for the future,...

*sigh*

Regulation, I agree on by the way. I can point towards some of the things folks have tried/used on chip-amps.

And that circuit is deceptively simple once you break it down. A handful of components for the SuperSymmetry circuit, and then essentially a quad of basic chip-amp circuits that add a mess but not much by way of complexity.

C

[mikec]

Not quite as rewarding as building your own amp from the ground up, but for not much dough, you can assemble a UcD-based amp. The general consensus is that these modules are in a whole different league (sonically speaking). Got UcD?

[Dotay]

Not quite as rewarding as building your own amp from the ground up, but for not much dough, you can assemble a UcD-based amp. The general consensus is that these modules are in a whole different league (sonically speaking). Got UcD?

Where can I get more information on these? Price?

[mikec]

Where can I get more information on these? Price?

Their web site is http://www.hypex.nl.. Adire Audio is the distributor in the US. Also, see the diyAudio Class-D form for additional indormation.

[cjd]

I don't really know how they compare sound-wise, but since they're single-ended input (from everythign I can see) they're more costly per watt than an equivalent chip-amp setup. About twice as much, give or take, though they spec a little lower on transformer requirements.

A little more compact, to be sure. And, of course, nothin wrong with having lots of options out there! Really a very slick option.

C

[JoshK]

UcD's are optimized for the XLR input, fwiw. In fact, care needs to be taken when using a single ended input to avoid grounding issues.

I have built 6 channels of UcD 400's, similar to Mike. In four of my channels, I am running them off of larger SLA packs (with parrallel PSU caps of course).

[cjd]

UcD's are optimized for the XLR input, fwiw. In fact, care needs to be taken when using a single ended input to avoid grounding issues.

I have built 6 channels of UcD 400's, similar to Mike. In four of my channels, I am running them off of larger SLA packs (with parrallel PSU caps of course).

That's good to know. I dug a little further and saw nothing stating this, except that it is quite obvious to me when looking at the board diagram. Go figure. I only looked at Adire's site before.

Anyone know if the power ratings are RMS ratings? (Assuming they are, but assumptions are very dangerous. . .)

Now, if the price were a little better. . .

C

[Dennis H]

Anyone know if the power ratings are RMS ratings?

A small quibble, RMS applies to voltage and current but not to power. Power is power. I assume you mean continuous power and yes the UcD can run continuously at rated power as long as you provide adequate heat sinking and ventilation. However the power ratings are at 4 ohms, not the more usual 8 ohms.

[cjd]

Thank you Dennis...

I'm still learning stuff when it comes to amps, but you did answer my question. And then a bit.

One last question: Are the UCD modules 2ohm stable?

Marzen: for what it's worth, on the "WHY" question: because an improvement is still an improvement, whether you can hear it or not. How many tweaks do we have in the audio world where no one questions whether it works, just whether it is needed? :P

C

[mikec]

One last question: Are the UCD modules 2ohm stable?
C

They do claim to bench test them into 1 ohm loads prior to shipping. However, I recently ran into difficulty with my UcD400 trying to drive two Dayton Reference subwoofers in parallel. The problem is likely my power supply design and not the fault of the module. I have used my modules with highly capacitive loads and they are completely stable.

[TacoD]

Well, since they are wooden shoes, it should really be in Nederlandse taal. Where's TacoD when we need him?

The translation in Dutch is (how to oil my woodenshoes);
"Hoe zet ik mijn klompen in de olie?" or
"Hoe olie ik mijn klompen?".

Maybe you want to ask "Do I have to oil/ treat my woodenshoes?"
In Dutch: "Moet ik mijn klompen ook behandelen?"

klompen = woodenshoes.
hoe = how to
olie = oil

But I only know of woodenshoes which make that kind of grinding noise, that's the typical woodenshoe sound :B (and one of the reasons to wear them ).

[Yves Smolders]

The UcD400 are stable at low impendances, however, you should note that while they are stable, they do not "double" power from 4 into 2 ohms.

The UcD are current limited at 20A (or 21A?) and such, using their +/-60V rails will output less than 400W as the impendance drops under 4 ohms.

There is always the UcD700 @ 700W/4R with 30A peak power

As for PS, 500VA transformers should be plenty for 2 modules UcD400, 800VA would be overkill. For caps, at DIYaudio a calculation was once made by Bruno Putzeys, that 8250uF per rail per channel into 8R is plenty, double that for 4R - Hypex seems to go for 10.000uf per rail/channel in their PS designs.

In this configuration they could even handle high power sine waves continuously into the specced output power.

As for balanced inputs, i'm running them unbalanced, simply by connecting the - (cold) lead to the ground as close to the source as possible - this is recommended by Hypex. Noise is extremely low.

[cjd]

Well... been a while since any breeze has wandered around here, but I've not been idle.

I just shelled out a ton of cash for parts - most will be useful regardless (transformers, etc.) but the PCB's are prototypes of course, and if things don't work who knows, I could blow up some resistors or something!

Should be fun.

C

[GMorris]

I have a Mauro Penasa My_Ref RevC amp powered by a Twisted Pear kookabura pre-amp and am very happy with the results. Smooth and detailed. The RevC has had quite a bit of thought put into it...check out Mauro's homepage but you will need to translate it ( I used altavista but this page is helpful for various translation duties http://translation.langenberg.com/ ). Mauro's web page: http://www.webalice.it/mauro.penasa/ Go to "plans audio" then "RELAZIONE_MY_REF " for a fairly thorough description of design process. I also have tried out the RevA which was not very happy (overheated/oscilated) with my higher capacitance braided Cat 5 speaker cables. The RevC does not have this problem though.

Very long thread on the My_Ref design: http://www.diyaudio.com/forums/showt...threadid=54571

kookaburra thread: http://www.diyaudio.com/forums/showt...threadid=64215

The Twisted Pear site http://www.twistedpearaudio.com/ now has a very interesting Nelson Pass influenced "x" pre-amp http://www.diyaudio.com/forums/showt...threadid=67448
and a logarithmic stepped attenuator with 128 steps controlled via a micro controller circuit and a potentiometer which is not in the signal path.

Diycable has UCD-180 modules for $85 now http://www.diycable.com/main/product...roducts_id=576
which makes a 5 or 7 channel surround amp very possible.....

[cjd]

Yeah, I gave up doing my own version of the BOSOZ and just orderd the Twisted Pair kit, plus the Joshua Tree which looks promising.

I have considered Mauro's layout, but ultimately need/want more power than that has - plus I do want to be able to run balanced cables and put the amps at the speakers with the pre-amp conveniently close to my chair. So, I'm messing with the SuperSymmetry chip-amp concepts. It've made it modular though, so I'll start off with a single 3886 board, make sure that works (well, power supply first, then single amp board), then test my bridge board, and then test the 4780 board. Lots of flexibility here - in theory my bridge/SUSY board should be usable with Mauro's kit. Component values may need to be worked on though.

C

[GMorris]

I can solder and follow diagrams but actual design work is well beyond my means. Good luck with your endeavors though. Russ from TP has mentioned that he is working on a higher powered amp based on Mauro's design.

[cjd]

Got a whole mess of parts today. Some of those components are teeny! eek. What HAVE I gotten myself into, I wonder!

Just waiting on PCB's at this point! Aaaargh.

C