Class-d soon in stereo!!

The two boardsHave had some time to finish a main board more, so soon I’ll have the oppotunity to try the amp out on my stereo set-up.

So soon the truth will come out ….. does this designhave any potential as a full rage amlifier?

I have done some modifications as well. Main point being that I have switched the output MosFets from the old IRF640 to the much newer IRFB5620. These are MosFets dedicated for class-d amplifiers and are around 3 times as fast as the IRF640! They are also capable of a bit more current.

I’m preparing an old Yamaha surround amp for receiving new guts … have taken out all the old stuff, have placed a new and bigger transformer and will add the two channels plus a power supply board.

The transformer are capable of either 2x33V or 2x38V at idle which will be some 2x30V / 2x35V AC  at nominal load.

I bought a bunch of these transformers a while ago, for 4 USD a piece ….. they are not used but was part of a clerance sale. Nice :)

Stay tuned!!!

UcD Modulator

SODA and UcD side by sideGot the UcD modulator daughter board assembled. It’s the UcD on the right with the DIP comparator. It works quite well and played the first time, something I have found with UcD all the times I have tried this kind of modulator .. it’s just so easy to get to oscillate.

With the chosen components it has a little ess gain than the SODA implementation. It also seems that noise level is a bit les, but this might just be because of the lesser gain ….

Have made some first measurements on distortion (without any special filter as you really should use measuring switching amps … ) and it seems that the SODA version have a little less distortion. But this is only the first trials, so I will not be that supprised if it turns out otherwise! :roll:

For the SODA I measured the first harmonic at some 0.16% for the UcD it was 0.23%. With a filter this would probably be a lot less, on the other hand it was measured at 100 Hz 😆

New Modulator PCB

New vs. old PCB

New vs. old PCB

Been too long since I posted anything here. Have been bussy …. but not with DIY :-(
Anyway a bit of DIY have been done :-)
I made a new and improved PCB for the SODA modulator. Now both the opamp and the compaator is SMD, as I’ve settled for MAX913. Though LT1711 is better on paper, it is just to hard to make work. Also it consumes way too much current, which will be a prblem when trying on higher rail voltages, as the smal smd regulators will not be able to handle it.
Made some further improvements to the filtering (LC) of the rails for the two ICs and also did some further filtering in the loop.
Now switching af a bit less than 300 kHz
The new pcb is the one on the left 😉
Have PCBs ready for a second chanel ….. stereo thay call it :mrgreen: …. got to try it out soon.
Also have PCBs made for UcD like modulators ……. stay tuned …




Building speakers are normaly a long and dificult process. It’s not just selecting the right drivers and putting these into a well designed enclosure. The real challange is the crossover.

The crossover will have a high impact on the resulting sound, and requires a lot of knowledge, possibility to make good measurements, possibility to make simulations and a good deal of patience.

Using a digital crossover can make the process much shorter and the chance of a succesful result much higher.

Using a line level crossover, either digital or analog, of course requires more amplifiers, one for each driver, but in general these then don’t need to be as big as otherwise.

But making you own digital filter board is of course not easy at all, and making a suitable PC program that can be used for ajusting the thing, is even harder.

I just came across this small board made by which is sold for only 99 USD. It has two analog inputs and four analog outputs, enough for either a four way active filter or stere 2 way.

The PC software used for ajusting the filter, comes for 10 USD. Given the time and cost of development it looks quite a good deal.

The board is based on the Analog Device ADAU1701 chip, which is the hart of the design.

Only things that are missing, I think, is ability to play from a digital source, through either SPDIF or USB, or both :)

But the SPDIF interface should be something you could add quite easily, as the ADAU1701 has 4 I2S inputs …

Would love to try this out with a small two way speaker and 4 amps …. -P

Have any of you tried out these small wonders????

Class-d …. it’s alive

Class-d amp running
Class-d amp running

The design is a self oscillating amp (SODA … read more here, use Google to reanslate from German), and it simply didn’t want to start oscillating.

Just running the modulator without inclusion of the MosFet output stage, it worked like a charm, but including the output stage just didn’t work.

The remedy is to feed back a little current from the modulator / comparator output to the input of the modulator, then it just workes 😉

 So one more resistor, and it just plays.

 Well I also found some other smal faults and improvements. One being the necessary level shifter, that shifts the +-5V volt output from the modulator / comparator to 5V above the negative rail, to feed it into the MosFet driver ( here IR2110). It simply didn’t work fast enough.

 I altered the circuit with a small 1N4148 and a 1k resistor in parallel with the input resistor to allow for at high current return path.

Take a look at a video of the beast her on YouTube: DIY Class-d SODA amplifier test

Class-d Motherboard and modulators

Class-d Bother Board

Class-d Motherboard

As you have probably seen on my web site, I have been working on class-d amplifiers for some time.

I have succesfully made a 400W full bridge for a subwofer, and a small mini amp based on the UcD design presented by Bruno Putzeys.

I have also for some time been working on a half bridge design, but have not ben so sucessful. Can only tell, it’s not that easy 😉

I have made a couple of models, but all had a bad tendency to be somewhat noisy, having a hissing sound at idle. As the goal is to make something better sounding than my 180 W dual mono LC-Audio The-End, this is of course not acceptable.

I did a trial version with separately regulated supplies for all stages, but it didnt do anything to the noise. Therefore my conclusion is that it is the modulator topology it self that I’m using that is not good enough. Therefore I decided to make a motherboard with all the switching, the filter, the power supply etc. all except the modulator, and a connector where I can place and try out verious modulator designs.

Tried out the new design with a simple SODA modulator, but something is not working as i should. It did start to modulate at some 350 kHz (which is a bit high) but it was not easy to get to start, switching the amp on and of  a lot of times to make it run.

SODA Modulator
SODA Modulator

I was actually a bit nervous that it had burned the whole thing, but then again nothing had become warm or haf smelled ……. To test the motherboard it self, I made up a small oscillator, doing 100 kHz and connected it to the modulator connector. Mother board seems to work just fine, so the fault must be on the modulator.

I will now try to cook up a modulator on a bread board connected to the motherboard.
Stay tuned :)

Sony PS3 as source in my Home Theater

PS3 Slim 250 Gb

PS3 Slim 250 Gb

Just bought a PS3. Wow ….

You simply get som much for your money.

Blu-ray is just such an upgrade from DVD. Now I see the full potential from my Sanyo PLV-Z700 pojector. Just saw the new Transformers movie … so much fun :)

The details are just so rich, and it also seems to improve the black level (mayme because of the HDMI 1.3 deep color).

Sound also seems to improve, but that might just be the good soundtrack on this particular movie.

Also games looks fantastic, playing Grand Tourismo on a 77″ big screen with surround soind 😀

And I haven’t even yet been trying out the streaming facilities or the capability to play movies from a hard drive.

What a steal …… go buy it …. 5 out of 5 stars

Modding Denon DVD 3910, new op-amps



Output section Denon DVD3910

Output section Denon DVD3910

I have had my current DVD player for about 5 years now. I’m quite happy with the CD sound, but I wanted to try out whether it would improve the sound by changing the op-amps in the stereo output stage to something better.



The player is using a PCM1796 DAC per channel in differential mode. This feeds into a differential I/U stage and a filter section using in all 3 dual op-amps per channel. The op-amps used are MJM2068.


There are many pages on the net describing the sound of op-amps and as many pages about moding your CD or DVD to better sound.

Well I ordered some samples of the new National op-amps LM4562 which have some of the industry’s best distortion performance , and has been announced as one of the best sounding op-amps for audio. I got 4 samples delivered for free :) (4 is the max and you have to pay for shipping).


I changed all the opa-mps in the filter section, which doesn’t take more than 15 minutes if you know which opamps to change :) (You’ll need a service manual with the schematics and print layout).


This is what I found:


Step 1

Changing IC303-IC306 from NJM2068 to LM4562

 Sound improvement/change:

  • Less aggressive in the treble
  • Larger sound stage, but moved further away
  • Can play louder without discomfort
  • More relaxed sound
  • More details which are more distinct and easier to pinpoint in the sound stage

 Improvement (on a scale from 0 to 5) : 2

Next I changed the I/U converter opamp to a OP275. I have used OP275 before, and have found their sound to be a little less correct, but creating a bigger feeling of room, which I think was what was needed now.




Step 2

Changing IC301 – IC302 from MJM2068 to OP275G

Sound improvement/change:

  • Sound was maybe more direct, and more present.
  • A bit less relaxed.
  • Room was maybe a bit bigger but maybe also a bit less well defined

Improvement: 0-1



Overall I think the improvements are ok for the little money spent. It’s way better than changing cables (which I think belongs on a 0 on my scale). The more expensive LM 4562, seems to be better than the standard MJM2068 in all respects, and gave a good improvement. Changing the last two MJM2068 to OP275, is not that clear. The change is so small that you have to set up a real A/B situation to make a fair comparison, so to me MJM equals OP275.


I think I’ll try changing the I/U op-amp to something else maybe the AD8066 which have been said to be very good especially in the bass, or maybe also use the LM4562, which for sure did pose an improvement.

 Should you do it … yes it’s worth a try, but don’t expect miracles …. do something for your room to achieve that :)

 Good reading about op-amps:



My scale of improvements:

  • 0 = None
  • 1 = Small but noticeable
  • 2 = I can also hear the difference tomorrow!
  • 3 = Definitely better
  • 4 = Big … my wife tells me unprovoked that something have changed for the better!

5 = Huge  … new speakers, new room

IMG_6788 (Large)

Using DirectFET from IRF for DIY amplifiers

TO-220 vs. DirectFETThe new series of FETs from International Rectifier called DirectFet, are not really DIY friendly :roll:

Look at the comparison between a normal TO-220 stle of MosFet, here IRF640 that I have used in my 400W amp. Besides is the tiny IRF6785 DirectFET which should we good for the same kind of amp size!!!

These MosFets are surface mounted devices with a “ball grid array” of 3 points at the bottom, for Gate and Source.

The top is a metal shield that is also acting as the Drain conductor.

How do you use therse for DIY? And why should you?

Well for a Class-D amplifier these divices have many advantages. Very litle lead induction, as there are no leads. They are very small which can result in a very compact layout, which will lower the EMI noise it will be emitting when switching.

But how do you solder these?

Well I just bought two samples, to have a closer look. They are indeed very small …. tiny I would say.

I have been thinking about using some larger vias (1mm hole) positioned right under the balls for D and G.

Then you should be able to solder from the other side. Maybe you need to put a thin layer os solder on the top side before putting the fet on, and starting to sloder from the other side ……

Have anyone out ther tried something like this and has som experience??

IR6785 top side

Here is the DirectFET (IRF6785) seen from the top. As can be seen from the paper it is placed upon, it’s only some 5 x 5 mm in size!

As for heatsink you can just use a small one with some double adhesive thermal conducting tape between. But as they are this small a heat sink will be needed, even tor test purpose. With class-d using TO-220 devices, I normaly don’t put on a heat sink when running the prototype on a low voltage (2x15V)

 IRF6785 Bottom view

Here is the DirectFET seen fom below. Look at the small contact points! …… can this be used for DIY?? ……. it has to be tried out :)

Hi-Fi & Sourround Copenhagen 2009

The yearly Hi-Fi show in Copenhagen was held last weekend (Nov. 14-15). The show is held at the Scandic Hotel and covers the ground floor + 1 and second floor. The rooms used at the 1. and 2. floor are normal small hotel rooms, so only a very limited number of people can listen at a time (6 -8 sitting). The rooms at the ground floor are generally very large, so all in all, most of the rooms does not resemble a normal living room much.

I always go there, as it’s the only such show in the vicinity, and there are normally a lot to see and hear. Most years, I go home and think that nothing I heard was much, or even, any better that what I have at home. Only twice I heard some thing that really was a special experience. First time was hearing the Wisdom Audio Rush M75 … wow, what dynamics, and what bass.

Second time was last year, where I heard the Wilson Sophia set up in a small room, playing on some Audio Research valve equipment (but using CD). That was just such a great combination, very wide stereo image, perfect mid and high, and a bass that just rocked.

Best Sound

This year, the best sound was delivered (to my mind) by a large set up of the new Dyanudio Consequence, playing on some very large McIntosh equipment.

I didn’t get to sit down, but I had a pretty good position in the middle and about 5 meters from the speakers. No doubt that these speakers are really well constructed. Very relaxed, but very detailed sound. Great imaging, and a low end to die for (had some advantage in playing in a large room).


Wilson Audio / Audio Research

I of course wanted to experience something similar as last year from Wilson, and a pair of the new Sasha was playing again on Audio research equipment.

This time on the ridiculously large power amps 610T amps. The relatively large room was crowded, and the heat from the big amps where not to be mistaken.

I was in and out several times, trying diferent listening positions, but nothing came near any special experience! The sound was good, clean and all … but not much better overall that my own at home!!

Gemme Audio

I also visited Gemme Audio, listening to the Vivance. I must admit that it was quite amassing to hear how much good sound can be had from a single Fostex FE108EZ unit. It was deep (I first didn’t believe it was just the Fostex unit playing), clear and with good highs. But the sound changed somewhat when moving out of the sweet spot (not surprisingly).

But thing on the small unit it’s quite mind blowing. On the other hand, the price is ridiculous, the size is not small at all (only the unit is), so why go to this length to get a sound that you could get, smaller (and probably better) using more units? Interesting though. The box is basically (as I understand) a back loaded horn construction, where a plate is covering the mouth and a small vent tube is inserted.

Otherwise there was not that much to report. There was way too much analog hype, exotic cables and hot air.

We’ll try again next year :)