[mjosef]

Picked up a loaner from fellow ACer Shadowlight.
I am running it full range with my back up speakers, VMPS626R. Main speakers are active-biamped, and takes more effort to integrate a new component, so that will come later, over the weekend.
Right off it sound very clean, but on the lean side. I started off with it plugged right into the wall. I was impressed with the details and slam. However, I think I am hearing a kind of 'ringing' on the edges of some notes, and after a while(about 30minutes) the sound bordered on the irritating...I think I am "hearing" ultrasonic ringing, the only way I can describe it is visually, like seeing a 'flare' after the flame has passed.
I think I ended up a 'headache'(?) which I almost never get. So I thought, maybe I should plug it into my conditioner. And that seem to tame it some, there was less 'flaring' (though not completely gone, but reduced)and the notes had a fuller body, overall it was an improvement, smoother and more warm. But it does have a nice dynamic kick to mid bass notes, and good drive with the lower bass notes, which was my intent  for this amp.
After an hour and half I switched over the a Jolida102b for comparison, and yes, I didn't get that impression of  mid/upper notes 'ringing/flaring'. So I do wonder if Digital amps tend to have some kind of ultrasonic ringing, which can cause piano notes or percussive instruments to sound "blaring"?

I do have some suggestions for Shadowlight which I think would improve his build, some of which I see AmpDesigner333 has already noted. The power leads are too long and are routed past the RCA inputs, and same with the speaker leads and RCA input leads. With a different layout these connection can be 70% shorter. And of course a better grounding scheme.

Besides the phenomena I mentioned, this amp shows good promise, at least for my bass duties.
I also noted that it took about 40 minutes before it hit its stride, the bottom sounded fuller after that time period.
But I prefer the sound of music via the Jolida, which is slightly modified with paper in oil coupling caps.
Are there coupling caps, or any caps in the signal path of these amps?

I have an Exposure 2010 I hope to compare it to also, more later.
Thanks Shadowlight for the opportunity. 

[dewardh]

Mintzar,

I do not see a R9 or 10, On my boards it seems to be 60-61

On the boards I have the input resistors one would change are R9 and R11 as on Mintzar's, they're SMDs on the "bottom" of the board, quite close to the input connectors (and that orange input cap).  R60 and R61 are axiel lead resistors (220 Ohm) on the "top" of the board, and seem to be part of the power supply circuitry.  They are right next to the regulators, by the heat sink.  Don't mess with them unless specifically instructed by Tom ! ! !

The input resistors we're talking about are the ones labeled "R2" on the reference design schematic.  If you cannot circuit trace the board to positively identify those components (they will have different values depending on board model and gain) you should return your boards to Tom for any changes.

[wushuliu]

Just a few simple hints...

1. Make the power supply as large as possible (big electrolytic caps especially) and as high voltage as specs allow

What are the benefits to increasing the power supply that much?

*sigh*

Just when I thought I was out of Digi-key, they pull me back in

[wushuliu]

Picked up a loaner from fellow ACer Shadowlight.
However, I think I am hearing a kind of 'ringing' on the edges of some notes, and after a while(about 30minutes) the sound bordered on the irritating...I think I am "hearing" ultrasonic ringing, the only way I can describe it is visually, like seeing a 'flare' after the flame has passed.
I think I ended up a 'headache'(?) which I almost never get. So I thought, maybe I should plug it into my conditioner. And that seem to tame it some, there was less 'flaring' (though not completely gone, but reduced)and the notes had a fuller body, overall it was an improvement, smoother and more warm. But it does have a nice dynamic kick to mid bass notes, and good drive with the lower bass notes, which was my intent  for this amp.
After an hour and half I switched over the a Jolida102b for comparison, and yes, I didn't get that impression of  mid/upper notes 'ringing/flaring'. So I do wonder if Digital amps tend to have some kind of ultrasonic ringing, which can cause piano notes or percussive instruments to sound "blaring"

Very possible the ringing and blare could be coming from elsewhere. Or just needs more time to settle. Part of the appeal of the classdaudio amps (at the least the IR versions I have) is the smoother top end. With a 4 ohm speaker in particular as the frequency response has a slight roll-off...

[dewardh]

So I do wonder if Digital amps tend to have some kind of ultrasonic ringing, which can cause piano notes or percussive instruments to sound "blaring"?

If we're talking about the IRS2092 family of amps it is unlikely.  The switching frequency of 400kHz is beyond the hearing of any of us, and is filtered at the output anyway.  Apart from that I don't see any ringing on in-band square waves.  If the speaker you were listening with has a rising impedance at high frequencies (8 Ohm or above) there will be a slight rise in high frequency response, however, and that might contribute to fatigue with an already bright speaker or source.  That can be dealt with by adding a zobel to control speaker impedance.

Are there coupling caps, or any caps in the signal path of these amps?

There is a single input coupling cap of "questionable" character, and the output filter has shunt caps that might not appeal to the highest audiophile standards (though I personally doubt that the output caps are an issue . . .).  There are, of course, lots of caps in the comparator and timing circuits, and in the snubbers on the FET switches, but because of the way the signal is deconstructed for switching it is not clear whether those would be considered "in the signal path".

[pjchappy]

I am running it full range with my back up speakers, VMPS626R.

Not questioning what you heard.  However, did you make any adjustments to the 626?  Midrange and/or tweeter controls?


Paul

[mintzar]

A bigger power supply is ALWAYS a good thing. It will allow for more reserve power, which means larger dynamic peaks. That also means that the amplifiers themselves will be more stable, thus lower the noise. A well designed power supply will make a larger difference in a component than any other mod you will do. The power is what is used to CREATE the audio signal... better power = better audio signal.

The voltage at 100v is just fine for the voltage needed by the amp boards, but the capacitance could be higher. In a perfect world you would use small capacitors, but hundreds of them to get very high level of storage capacitance. By using multiple resistors you are lowering the series resistance. That means less noise and more extension, basically. You can also replace the caps on the power supply with Nichicon Muse (anywhere from $40-$80 per 100v capacitor)... compared to the cheap caps currently on the board you'll find a very nice upgrade.

The R9 and R11 resistors are SMDs on the bottom of the board near the inputs. VERY small. If you're not confident with your soldering abilities I would send it back to Tom to have him replace them.

[mjosef]

Not questioning what you heard.  However, did you make any adjustments to the 626?  Midrange and/or tweeter controls?


Paul

I had that dialed in previously using RS SPL meter and pink noise. I even backed off some more on the mid panel and tweeter, didn't help...what I was hearing came from the amp...confirmed when I switched amps. (long time VMPSer here, so really familiar with all the 'knobs'  )

[wushuliu]

A bigger power supply is ALWAYS a good thing. It will allow for more reserve power, which means larger dynamic peaks. That also means that the amplifiers themselves will be more stable, thus lower the noise. A well designed power supply will make a larger difference in a component than any other mod you will do. The power is what is used to CREATE the audio signal... better power = better audio signal.

The voltage at 100v is just fine for the voltage needed by the amp boards, but the capacitance could be higher. In a perfect world you would use small capacitors, but hundreds of them to get very high level of storage capacitance. By using multiple resistors you are lowering the series resistance. That means less noise and more extension, basically. You can also replace the caps on the power supply with Nichicon Muse (anywhere from $40-$80 per 100v capacitor)... compared to the cheap caps currently on the board you'll find a very nice upgrade.

The R9 and R11 resistors are SMDs on the bottom of the board near the inputs. VERY small. If you're not confident with your soldering abilities I would send it back to Tom to have him replace them.

excellent. thanks.

[pjchappy]

I had that dialed in previously using RS SPL meter and pink noise. I even backed off some more on the mid panel and tweeter, didn't help...what I was hearing came from the amp...confirmed when I switched amps. (long time VMPSer here, so really familiar with all the 'knobs'  )

Thanks for the clarification. 

[shadowlight]

I do have some suggestions for Shadowlight which I think would improve his build, some of which I see AmpDesigner333 has already noted. The power leads are too long and are routed past the RCA inputs, and same with the speaker leads and RCA input leads. With a different layout these connection can be 70% shorter. And of course a better grounding scheme.

I agree.  When I started to cut the hole for the IEC plug I was not paying attention to it and it ended up being on the wrong side, hence the wires going across the rca.  I plan to flip the back plate around and drill new holes.  As for grounding I am open to suggestion on how best to do that.

Thanks Shadowlight for the opportunity. 

You are welcome.

[mintzar]

Oops I need to clarify something in my post...

I said "by using multiple 'resistors' you are lowering the series resistance." I meant to use 'capacitors' not resistors...

[dewardh]

AmpDesigner and mintzar are correct about power supplies . . . it is hard to imagine a serious argument (other than cost) against larger caps in an amplifier power supply, especially for half-bridge Class D amps.  You may have to use a higher current bridge rectifier (or a slow start circuit) to keep from blowing diodes on startup, but that’s minor.  Good design practice would also include bypass caps and snubbers at least on the supply output, but snubbers especially must be “tuned” to the transformer/capacitors used, and the load, which combined with cost is probably why they are not present on the “general purpose” classDaudio supply boards.  It’s another opportunity for DIYers to improve an already good amp with relatively simple mods.

Industry practice is to use caps with a voltage rating about 30% over supply . . . so the 35 V supplies can use 50 V rating parts, the 50 V supplies can use 63 V parts, and 100 V parts are only needed for the 70 V supplies.  You can get higher capacitance in smaller cans for less money with the entirely satisfactory lower voltage rating parts . . . there’s no need for 100 V caps on a 35 V supply.  Do get the 3000 hour @ 105 degree parts, though . . . they last longer.  Low esr is also a plus . . .

There’s an interesting (three part, follow the links at the end of each page) article on power supply design at:

http://www.tnt-audio.com/clinica/ssps1_e.html

and on snubbers (although I think snubbing the supply output is even more important than snubbing the diodes) at:

http://www.hagtech.com/pdf/snubber.pdf

An example of a power supply with a snubbered output is at:

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

Note the use of 24,700 mfd (and paralleled caps) on each rail of this gainclone power supply.  Some might call it overkill, but I don’t. 

[wushuliu]

AmpDesigner and mintzar are correct about power supplies . . . it is hard to imagine a serious argument (other than cost) against larger caps in an amplifier power supply, especially for half-bridge Class D amps.  You may have to use a higher current bridge rectifier (or a slow start circuit) to keep from blowing diodes on startup, but that’s minor.  Good design practice would also include bypass caps and snubbers at least on the supply output, but snubbers especially must be “tuned” to the transformer/capacitors used, and the load, which combined with cost is probably why they are not present on the “general purpose” classDaudio supply boards.  It’s another opportunity for DIYers to improve an already good amp with relatively simple mods.

You think swapping out the 3300ufs w/ the 250w IR w/ 6800uf would be too much for the diodes?

[dewardh]

You think swapping out the 3300ufs w/ the 250w IR w/ 6800uf would be too much for the diodes?

Don't know.  The web site says that the diodes in the "heavy duty"  supply have a higher rating, but not by a lot.  I'm putting 30,000 mfd on a couple of my boards, so I'll probably change the bridge "just in case".  The inrush current can be quite high . . . the caps start out as a dead short, and almost the only thing limiting current is the (quite low) impedance of the transformer secondaries.  The bigger the caps the longer the inrush lasts, so the question really is "how many cycles can those diodes handle 10 times (or more) their continuous rating?".  25-35 Amp bridges are cheap . . . and cheap insurance.

[wushuliu]

Don't know.  The web site says that the diodes in the "heavy duty"  supply have a higher rating, but not by a lot.  I'm putting 30,000 mfd on a couple of my boards, so I'll probably change the bridge "just in case".  The inrush current can be quite high . . . the caps start out as a dead short, and almost the only thing limiting current is the (quite low) impedance of the transformer secondaries.  The bigger the caps the longer the inrush lasts, so the question really is "how many cycles can those diodes handle 10 times (or more) their continuous rating?".  25-35 Amp bridges are cheap . . . and cheap insurance.

Is there a particular bridge rectifier you recommend, since digikey and I are so tight...

[mintzar]

Michael Percy Audio sells the Vishay diodes and IXYs diodes, which are the ones I recommend. They are the fastest diodes on the market by far. The diodes need to be at least 2x the voltage of the TX. If you are using larger caps you should make it 4x to be safe. The problem is, that as the voltage gets higher the diode isn't as fast. So it's important to get a voltage rating that is just high enough, but not too high. The big thing with the diodes... and the caps for that matter... is what will fit on the board. You can always build a new PSU board, but it will be substantially easier to just replace the caps and bridge that are already there.

[tvyankee]

Hello,

Any new info on the custom case that was being built for these amps?

Thanks

[Mariusz]

jtwrace is working on some custom enclousures which are going to be available at reasonable price to anyone interested in dressing up there class D-amp.
Those interested can contact Jason ( jtwrace) for more info. 

P.S
Looking forward to some pictures Jason ...... ones they are finished.

Cheers
Mariusz

 

[jtwrace]

Hello,

Any new info on the custom case that was being built for these amps?

Thanks

Not really.  I'm just tweaking one thing that I learned from my first prototype.