Hypex NC400 Build

[benthe8track]

Other than the case I'm pretty out of my element on this one so looking for some input from you electical guys.
This came up in other spots and I'm still not sure if I will go bridged or not, in either case I'll cut/tap the holes for it just in case. I'm shooting for are more standard 17" wide than the tiny cases people typically use with class D amps. The red things are the NC400s, the SMPS1200s are self evident, as well as the soft start. The grey pcb up front is the driver board for the VU meters. The only thing I haven't drawn yet is the pushbutton I ordered (well I ordered a few to make sure the stainless matches).

Looking for some input for this layout before I get to far into the industrial design aspect of the case.



And an ISO render. Everything is to scale.

[5th element]

This kind of thing is quite important for making sure that noise and mains buzz/issues are kept to a minimum. A SMPS does help with this as it is free from traditional rectification noise and the otherwise induced transformer currents, still it's prudent to keep the layout as clean as possible to ensure satisfactory performance.

I know that this is going to overlap with some of your current overlap so bear with me on this.

First of all you want the mains inlet to be on one far side of the cabinet. You then need to place the softstart and SMPS units as close to the inlet as possible. My recommendation would be to place them down the left hand side of the cabinet as it is now, keeping the mains inlet in the same place.

Then place the input connectors as far to the other side of the cabinet as is possible.

The inputs work at a low signal level and are of a high impedance, this makes them the most susceptible part of the design to picking up any noise that you don't want. The mains inlet and the SMPSs are the noisiest parts of the circuit, radiate magnetic fields and work at high voltages, making them the most capable of coupling into place you don't want them to, so keeping the two as far a part is always a good idea.

Looking at the ncore modules, if you place the side with the input connectors to the north, the large connector to the left is where you input power, the connector to the right of it is the loudspeaker output and the connector to the right of that is the signal input.

The way you wish to lay this out is with the 4 ncores facing north with their connectors facing the back of the case. They should all be placed in one line, one closest to the back of the enclosure, then with each one further south of the other towards the front of the case.

The input connectors on the ncore boards should then all be closest to the right hand side of the case, the same as the input connectors. You then wish to take the cables you've got for connecting the ncores to the input connectors and run the wires directly to the right, to the right hand side of the case and then up along the side of the case to the input connectors.

Next you've got the power to route. Your power supplies, being down the left hand side of the case, are in a good position to connect to the power connectors of the ncore boards. Take the power wires directly to the left of the power connectors on the ncores and towards the power supplies. Make sure the power wires are twisted together, or braided. If you're using the ncore power cable then obviously just go with that.

Finally you've got the outputs of the ncores. The connectors for these should be placed away from the signal inputs and closer towards the power input end of the case. Loudspeakers are low impedance devices that require large voltages to operate, they do not couple easily with interference and the like, but the loudspeaker outputs can and will couple into the input circuitry, so again, you want to keep the loudspeaker outputs away from the inputs.

Run the loudspeaker output wires off to the left a little, from the connector, and then up to the outputs.

If you arrange things like this you don't have any wires crossing each other that could otherwise lead to problems and you also keep the sensitive areas of the circuitry away from the parts that could inject noise into them.

[BobEllis]

Those beautiful Wavecor Ardents of yours are just crying out for an amp like a Pass/First Watt F5Turbo or F5X if you want to go DIY. A couple of honking 100W+ class A monoblocks would be a great compliment. 8) :twisted:

5th has provided good guidance. I will just add keep all wires pairs twisted. Signal and return, positive power and negative, Mains neutral and live. Be mindful of where the current flows to minimize the effects of loop area. You don't want antennae radiating or picking up noise inside your enclosure.

[JonMarsh]

Good suggestions here; for example, a wider, more shallow enclosure might help with implementing some of Mat's inputs.

Now, as far as that first watt business goes, I'm more interested in the first 20 watts, and with THD+N that is mainly limited by the noise floor of the driving source, the nCore 400 doesn't give up much to anyone....





One might even argue that this is a First50 design...

And with the somewhat lower sensitivity due to the characteristics of the Wavecor SW223BD01, that's a good thing! :T

[benthe8track]

Thanks for the great feedback guys.
5th Element, is this what you were thinking? I've seen (especially on the machined cases) dividers added between the power supplies and amp modulus, think there is any actual advantage to that?



This still maintains the 16" wide floor I'm shooting for, will be posts in the corners that will fill it out a little bit. The depth is at 15" and I'm fairly limited by the size of the NCORs. The only thing I'm not really sure of are the mounting holes of the amp modulus in relation to the connectors. Can someone with the ncores kicking around take a picture of the top and bottom of them lined up with this image?



And some parts started to show up today. I'm a bit unsure of the blue VU meters:

[5th element]

Yes, that is pretty much what I was getting at, although I would put some distance between the ncores and the right hand side of the case.

Dividers can help, but usually the best method for reducing interference is simply good layout and putting space between the various parts of the design.

[mjb]

There are lots of nice NC400s layout examples on Google. Good luck with your project, it looks like fun!

[Johnloudb]

Nice project! How much did all the amp and power boards come to?

[benthe8track]

Nice project! How much did all the amp and power boards come to?

Thanks! If I go bridged it's about 2600, considering the case is being made on a favor. If I don't bridge it's about a grand cheaper. Still not 100%, I may just try without first but get the cables/case ready for future expansion.

Still waiting to hear back from the material guy so I can size everything in CAD properly. Hopefully I'll have some concepts worked out during my next days off.

[benthe8track]

Found a few minutes to put into this project this past week. I think I got the case concept roughed out now I'm just waiting to hear back from my sheetmetal guy about how he wants to preform the stainless origami. So the front will be waterjet cut, then the recesses on the VU meters chamfered. I'm looking for some sort of off the shelf tapped blocks so I can uses fasteners, I can weld it but that's not very professional. Right now the top, sides and back may all be one piece.



I had a thought, is there a cheap/easy pre-made board I can add that would facilitate some sort of LED clipping indicator? This goes back to the suggestion of only adding 2 more NC400s if I find I'm running out of jam.

[JonMarsh]

I had a thought, is there a cheap/easy pre-made board I can add that would facilitate some sort of LED clipping indicator? This goes back to the suggestion of only adding 2 more NC400s if I find I'm running out of jam.

That's trickier than you might think- a true clip indicator compares the input signal with the output signal and when there's a difference, provides a comparator logic output. Simplest circuit I could do in a production amplifier was based on a JFET input opamp, connected differentially to the inputs, and comparing the amplifier output signal with the inputs based on a gain attenuator. I've seen many clip indicators which are really just peak level indicators, and may be firing too early or too late. eBay is sometimes a good source for odds and ends like this, but trying quite a few search phrases, I'm coming up dry. Google will turn up lots of circuits for clipping indicators, which are actually just level circuits tied to the power supply rails, in most cases. The thing I liked about the one I came up with in the late 70's was that it would show ANY type of deviation of output from input, including protection circuit activity, or slew rate limiting or extreme HF output (OTOH, amp would safely do 100V/usec, which was a LOT for a late 70's bipolar design- was using the newest Motorola parts at that time).

[benthe8track]

That's trickier than you might think- a true clip indicator compares the input signal with the output signal and when there's a difference, provides a comparator logic output. Simplest circuit I could do in a production amplifier was based on a JFET input opamp, connected differentially to the inputs, and comparing the amplifier output signal with the inputs based on a gain attenuator. I've seen many clip indicators which are really just peak level indicators, and may be firing too early or too late. eBay is sometimes a good source for odds and ends like this, but trying quite a few search phrases, I'm coming up dry. Google will turn up lots of circuits for clipping indicators, which are actually just level circuits tied to the power supply rails, in most cases. The thing I liked about the one I came up with in the late 70's was that it would show ANY type of deviation of output from input, including protection circuit activity, or slew rate limiting or extreme HF output (OTOH, amp would safely do 100V/usec, which was a LOT for a late 70's bipolar design- was using the newest Motorola parts at that time).

So this guy wouldn't really do what I want it to do? What about this one? You wouldn't have a copy of your version somewhere would you? I've changed my criteria for one from "cheap" to "just available" haha.

[5th element]

The second one is far more sophisticated vs the first one and works on the principle that Jon was talking about in the previous post. It compares the output with the input and if they should be different, for whatever reason, it triggers the clip indicator.

[benthe8track]

The second one is far more sophisticated vs the first one and works on the principle that Jon was talking about in the previous post. It compares the output with the input and if they should be different, for whatever reason, it triggers the clip indicator.

Awesome thanks. Any recomendation for a place to build it for me on a PCB?

[JonMarsh]

Found a few minutes to put into this project this past week. I think I got the case concept roughed out now I'm just waiting to hear back from my sheetmetal guy about how he wants to preform the stainless origami. So the front will be waterjet cut, then the recesses on the VU meters chamfered. I'm looking for some sort of off the shelf tapped blocks so I can uses fasteners, I can weld it but that's not very professional. Right now the top, sides and back may all be one piece.



I had a thought, is there a cheap/easy pre-made board I can add that would facilitate some sort of LED clipping indicator? This goes back to the suggestion of only adding 2 more NC400s if I find I'm running out of jam.

This looks very nice Ben, but it's clear some people have WAY too much free time on their hands! I don't know when I could find time to do the sketch up drawing in this detail right now!

Somewhere I know I have the documentation for that amp, but finding it would be a b*tch- likely stashed away somewhere clever in one of my three storage units.... I'd check out the resources Matt proposed.

[benthe8track]

This looks very nice Ben, but it's clear some people have WAY too much free time on their hands! I don't know when I could find time to do the sketch up drawing in this detail right now!

Somewhere I know I have the documentation for that amp, but finding it would be a b*tch- likely stashed away somewhere clever in one of my three storage units.... I'd check out the resources Matt proposed.

Haha well Hypex was kind enough to provide me with the .STEP file for the power supplies, hence whey the NC400s are just cylinders. I work 7/7 but I don't have THAT much time.

[5th element]

Awesome thanks. Any recomendation for a place to build it for me on a PCB?

Unfortunately not It will also require a very small modification (resistor divider connected to the output of the ncore) to enable it to work with the ncore.

Do you have a scope?

[benthe8track]

Unfortunately not It will also require a very small modification (resistor divider connected to the output of the ncore) to enable it to work with the ncore.

Do you have a scope?

I don't have a scope but I'll see if the E&I guys have one at work or try to beg borrow or steal one in town.

[5th element]

If you can get your hands on a scope then it is relatively easy to see if your amplifier clips at the maximum volumes you will be interested in listening at. Putting in a clip indicator seems a tad overkill imo. They are useful things for people in sound reinforcement and discos/night clubs etc, where the venues and requirements can change, but at home, a scope is all you really need.

[benthe8track]

I'm just finalzing the mounting hole locations for my chassis guy and had a thought: how hard would it be to add a 12V trigger to the softstart module? Anything off the shelf I can use for this?

EDIT: So if I'm understanding this right the trigger sends a constant 12v, so I could just use a relay to close the switched input on the softstart?

[5th element]

It depends if the 12V trigger is powerful enough to actually drive the relay, or whether it is simply a very low current 'signal' present, flag, type of interface.

[benthe8track]

It depends if the 12V trigger is powerful enough to actually drive the relay, or whether it is simply a very low current 'signal' present, flag, type of interface.

Looks like 15mA max is typical.
Maybe a relay isn't well suited. I need to translate that 12v signal to the latch input on the soft start. The front panel switch will also attach to that input.

http://www.hypex.nl/docs/softstart_datasheet.pdf

[JonMarsh]

I'm enjoying seeing you work through these issues- I'm HUGELY frustrated that I have so little time to work on my own that I can almost foresee being able to buy a NAD M22 before I can find time to assemble my nCore mono blocks!

[benthe8track]

I'm enjoying seeing you work through these issues- I'm HUGELY frustrated that I have so little time to work on my own that I can almost foresee being able to buy a NAD M22 before I can find time to assemble my nCore mono blocks!

Haha well this electronics stuff can be frustrating--designing the enclosure is fun at least. It was interesting, I read that the M22 is based on the NC400 and not the NC1200 that we initially thought. That gave me more reassurance of the value of this project.

[5th element]

Ben, it sounds like what you need is a nominally, always on, circuit. This would use a small transformer to power some logic circuitry that would accept the 12V trigger voltage and close a pair of relays, for soft start, once triggered.

The SMPS1200 already has a trigger function built in, but it's polarity is annoyingly reversed. That is pin 1 on J5 accepts a DC voltage and if present it will shut the unit down into standby mode, but if the voltage is removed the SMPS will soft-start start up. The standby power draw is less than 1 watt so isn't much of a concern.

[JonMarsh]

Haha well this electronics stuff can be frustrating--designing the enclosure is fun at least. It was interesting, I read that the M22 is based on the NC400 and not the NC1200 that we initially thought. That gave me more reassurance of the value of this project.

That's interesting- where did you hear about that? I came across this thread on Audiocircle which mentions it, and also has a lot of trash talking by people who haven't seen or heard one or held it in their hands, but are willing to pass summary judgements (OK, pretty normal for the internet, I guess....)

http://www.audiocircle.com/index.php?topic=128950.0

That's an interesting idea, too- compare a standard NCORE 400 build against the M22. OTOH, my build won't be particularly standard, with balanced input transformers and what not... the problem as always, is finding the time, and at the moment, setting up someplace to work on it!

[BobEllis]

Rod Elliot's sequencer circuit http://sound.westhost.com/project144.htm might offer the basics you need to trigger your amp. Tie the Collector of the BC549 in the "12V Trigger" section to the power control pin of the SMPS1200 and a 1-2K resistor to an always on power supply of appropriate voltage to shut it down. Apply 12V to the circuit's trigger and the SMPS power pin goes low. The 1K resistor to the always on power supply limits the current. Hope this helps.

[JonMarsh]

BTW, Ben, if you haven't read this PDF

http://www.linearaudio.nl/linearaudio.nl/images/pdf/Volume_1_BP.pdf

and this interview article,

http://www.soundstageultra.com/index.php/features-menu/general-interest-interviews-menu/455-searching-for-the-extreme-bruno-putzeys-of-mola-mola-hypex-and-grimm-audio-part-one

I think you want to....

[benthe8track]

That's interesting- where did you hear about that? I came across this thread on Audiocircle which mentions it, and also has a lot of trash talking by people who haven't seen or heard one or held it in their hands, but are willing to pass summary judgements (OK, pretty normal for the internet, I guess....)

http://www.audiocircle.com/index.php?topic=128950.0

That's an interesting idea, too- compare a standard NCORE 400 build against the M22. OTOH, my build won't be particularly standard, with balanced input transformers and what not... the problem as always, is finding the time, and at the moment, setting up someplace to work on it!

Found it here: http://www.digitalaudioreview.net/20...he-rest-of-us/

Rod Elliot's sequencer circuit http://sound.westhost.com/project144.htm might offer the basics you need to trigger your amp. Tie the Collector of the BC549 in the "12V Trigger" section to the power control pin of the SMPS1200 and a 1-2K resistor to an always on power supply of appropriate voltage to shut it down. Apply 12V to the circuit's trigger and the SMPS power pin goes low. The 1K resistor to the always on power supply limits the current. Hope this helps.

Thanks for the input Bob, the only thing is I want to retain the softstart.

I'll have to double check the actual current the 12V triggers put out but would this relay do the job?

http://www.digikey.ca/product-detail/en/TXS2SA-12V/255-1394-5-ND/385899

BTW, Ben, if you haven't read this PDF

http://www.linearaudio.nl/linearaudio.nl/images/pdf/Volume_1_BP.pdf

and this interview article,

http://www.soundstageultra.com/index.php/features-menu/general-interest-interviews-menu/455-searching-for-the-extreme-bruno-putzeys-of-mola-mola-hypex-and-grimm-audio-part-one

I think you want to....

Thanks Jon, I've read the first link but not the 2nd one.

[BobEllis]

Looking at the data sheet, I would expect any trigger output to be able to drive the <6 mA the 12V version needs to switch. You still need a power supply to turn the hypex PSU off as 5th showed you. I have all the parts needed to build a transistor trigger in my junk bin so I thought that might be easier. (and click free)

The way I understand the SMPS1200, the soft start and power control are integrated, so you can't use just one. It also seems like using the PSU's signal to shut down the amp modules when the PSU shuts down would be a good idea.

[JonMarsh]

Found it here: http://www.digitalaudioreview.net/20...he-rest-of-us/

Thanks Ben- snagged it for the references in Evernote!

Sounds like they've customized it in some interesting ways, between this article and what's on their website.

[benthe8track]

Looking at the data sheet, I would expect any trigger output to be able to drive the <6 mA the 12V version needs to switch. You still need a power supply to turn the hypex PSU off as 5th showed you. I have all the parts needed to build a transistor trigger in my junk bin so I thought that might be easier. (and click free)

The way I understand the SMPS1200, the soft start and power control are integrated, so you can't use just one. It also seems like using the PSU's signal to shut down the amp modules when the PSU shuts down would be a good idea.

Sorry you'll have to excuse me I'm not picking up what you're putting down. If I use this guy below, can't I just use a relay on it's switched input? That shouldn't require any extra power should it? The 12v closes the relay which starts the soft start which powers up the 2 SMPS1200s.

Image not available

[BobEllis]

The Hypex power supply has a soft start capability built in - no need for that one. http://www.hypex.nl/docs/SMPS1200_datasheet.pdf It also has a signal line that you connect to the amps that will shut them down when you shut down the power supply. That prevents speaker thumps.

[benthe8track]

I'm back to work tomorrow but my stuff should show up from Hypex. Because our dollar is in such terrible shape I scaled it back somewhat. Went with a single SMPS1200 and two NC400s. The case is drilled for expansion of another SMPS and two more NC400s if I decide to go that route down the road.

My friend making the case for me back home says it will be ready by the time I'm back in town next Thursday.
It will look a lot like this:

[benthe8track]

Laser cut and CNC formed:

[---k---]

Pretty.
I wish I knew local people who could do cnc and metal work like that for a reasonable price.

[benthe8track]

It will be a lot prettier with the plastic off haha.
Laser cutting can be surprisingly cheap, it's often more cost effective than buying the raw sheet, at least in my experience.

[5th element]

It will be a lot prettier with the plastic off haha.
Laser cutting can be surprisingly cheap, it's often more cost effective than buying the raw sheet, at least in my experience.

This is my experience too.

[JonMarsh]

Laser cut and CNC formed:

I'm jealous! :W :B

I'm also so far behind in stuff that getting to my build seems like an ever receding goal... do you know, I haven't even found the time since Christmas to finish setting up the new/replacement drive for the music server and getting that on the air? Or getting the new Rubidium clock and analog power supply setup with the Mutec that's been sitting here for almost a year?

And my Cambridge amp was getting flaky before Christmas- doesn't like to turn-on with a load connected, something with the protection circuits. Works after that. I may have to go rob a bank (First National Bank of Jonmarsh) and get that M22, the way things are going...

And.... (cue the Starwars Imperial theme... ) we have been invited to respond to a new project RFQ at THAT customer... by April 10... it's not that long I've been out of the gulag, and we don't have enough minions on hand locally (even with the IR acquisition) that I may escape that fate this time... let's hope I'm just being overly dramatic... and this will pass me by, leaving me unscathed, and working a normal job a while longer...

Yeah, whistling past the graveyard I am....

[---k---]

Maybe you should buy a Moto X (can I say I HATE my work issued iPhone) with a case to hang it on your belt. Add a Moto 360 watch and a pair of Google Glasses. Wear it all to work for a few days and see what happens. And, if you have to, start all conversations casually with a list of things you can do on Android that you can't do on other platforms.

[JonMarsh]

Just to stir the pot a little more, build the sense of anticipation, here's a review of another nCore design, just out on Stereophile... online finally.

http://www.stereophile.com/content/theta-digital-prometheus-monoblock-power-amplifier

They're mono blocks, and they're purty, and purty expensive, too, at $6K each. But they're not crazy expensive... well, by some standards. Go price some of the new Boulder amplifiers!


Output impedance is low, so frequency response is pretty good, including minimal response interaction with the simulated loudspeaker load.




1 kHz square response is quite clean for a CLASS D amp- too bad we can't get CD players that look this good.




1 kHz distortion versus output power is in the ball park for what I might expect for an nCore design, though not quite as good as the published nCore 400 specs. Certainly you could call this a "First 10" amplifier as regards the quality of the first 10 watts, which appears to be limited only by noise performance.




Distortion versus load impedance and frequency are shown here at 20VRMS, moderately high output power: obviously, not so happy with 2 ohm loads, but it's actually lower than many other tested class D amps at 8 ohms.



That rising harmonic distortion at the top end might raise concerns about high frequency intermodulation distortion...



But that is actually pretty good as power amps go. Let's face it, you wouldn't want to drive this level into your tweeters- 250W at clipping mixed 19 and 20 kHz would surely send them to the dead speaker heaven in the sky...

The reviewers subjective impressions comparing with his reference amp and another more expensive new Hypex based class D amp:

Comparisons
While auditioning the Theta Prometheus monoblocks, I also listened to my Mark Levinson No.334 ($5900 in 1999) and the MBL Corona C15 monoblocks ($25,000/pair). A single Prometheus or Corona C15 tips the scale at less than half the weight of the No.334. The No.334 has softer-, warmer-sounding bass, and delivers 200W less at clipping into 4 ohms than either class-D amp. This might be a problem with less sensitive speakers and bigger rooms. More important, the Levinson, which I've thoroughly enjoyed for 15 years, seemed compressed, shut down, and dark, with a two-dimensional soundstage. In comparison, both class-D amps were faster, more open, more transparent, and produced a greater sense of three-dimensionality. Compared to the Prometheus, the MBL seemed leaner in the midrange and upper bass when driving the Revel Salon2s. The Prometheus had a much more successful integration with the Revel Rhythm2 subwoofer than the other two amps, being able to produce a full, driving midrange, and to reveal layers of three-dimensional detail. Overall, for me, the Theta Prometheus delivered the most emotionally involving sound.

Well, this got me thinking again... (a dangerous proposition, as you all well know...)

After reviewing project commitments and the flaky protection circuit behavior of my Cambridge Audio 840 amplifier (which still sounds great, just doesn't want to turn on with a speaker load connected), and seeing how long it will be until I work down the to-do list to my own nCore DIY project, I (watch out, here come the internet cliches) broke down and pulled the trigger on an NAD M22, as I'd really like to get the audio system fully functional again in the near future, not just by next Christmas... (this includes wrapping up the data prep on the new drive for the music server, and setting up the Mutec MC3 and the new Rubidium oscillator and linear power supply for THAT....




Most places are quoting 1-2 weeks before shipping; Crutchfield just reported "Low Stock". We'll see how that turns out... stay tuned.

Funny- before hitting post, I got a call from Crutchfield, to confirm the order, and that they have it in stock and it will ship to be here by Friday!

BTW, this is the only place I've seen interior shots- (captured from video, though)








Definitely not a stock nCore module- but I didn't expect that, based on the write ups I've read about what NAD brought to the table for this project. Note the copper bus bar on the top side of the PCB. Pretty clean looking layout all in all, and lots of small low impedance caps... (big caps have more inductance- not a good thing in the power supply for a Class D amp).

[benthe8track]

Very cool! Looking forward to hearing your impressions of the NAD.
I've been trying to find a reasonable price for the NAD M12, I'm ready to buy but I'm not paying MSRP.
It's too bad about your Cambridge, it was on my short list but I decided I wanted analog inputs to simplify things a bit.

[JonMarsh]

This does have balanced analog inputs, 200K input impedance shunted by just 100pF; I think the new Cambridge preamp I got around XMAS time should drive that just fine!

The new version of the Cambridge amp looks pretty good, too,





But I'd promised GF to look into the Class D to "save energy" (40 watts idle power versus 180W; output efficiency around 90%, instead of 40% for an AB amplifier). The 840W I've got is a regular space heater, which is OK in the winter, but not so great in the summer time down here in the more southern parts of the world!

Yes, that's the special 10A IEC AC connector- I'm surprised that isn't on more larger amplifiers.

It should be here Friday- I'll pop the top and do a nice picture of the innards to share...


For contrast, the back of the M22:



It weighs about half of the Cambridge, coming in at just under 20 lb.

[---k---]

Looking forward to hearing your review.

[JonMarsh]

Looking forward to hearing your review.

Hah! Me, I'm just looking forward to hearing it! I'll set it up first with the existing re-clocking setup, so I'm not changing other stuff in the system. I realize I have a lot of the recordings that the guy doing the review on Stereophile used- maybe I should track down the rest! :W