Alpha Nirvana 39w 8ohm Class A Amp

This is what I have already done cause I am sure that the 100R on the snubber is the proper one, although AndyR has placed them both.

Spiros, I suggest the situation is:
  • if you only use one 100R resistor - then it should be on the snubber board (with a wire link in that position on the main PCB).
  • but there is no sonic penalty if you use both 100R resistors - like I did.
 
Andy,
I could guess that there will be no problem placing both resistors but I wanted to do what is recommended by the designer. I knew you had used both of them. Take a look of my snubbers. I think it's Ok for my first time soldering smds. Thanks for your concern.
P1090007.JPG
 
Andy,
I could guess that there will be no problem placing both resistors but I wanted to do what is recommended by the designer.

Trouble is, Spiros ... in this build, Hugh seems not to be the only designer. 😵

Sure, he produced the original cct diagrams - for both the initial 8R version ... and then the subsequent 4R version. But then his circuit was modified by the guys that produced the PCB and its associated cct diagram - so, for example, we ended up with:
* a completely different component numbering system
* and Molex connectors on the PCB (instead of spades). 🙁

Now, @Hugh came to listen to my build (having 2x 100R Gate resistors) ... and didn't notice anything wrong with the sound. So, as the original designer, I am interested to know - given I am going to take them apart, to install 0.15R Source reses (instead of the 0.12R reses I currently have) ... to reduce the heat load on the heatsinks - is it worth replacing the 100R reses on the main PCB with wire links?
 
You are right Andy. I don't think that placing the two resistors it would be a big deal. Personally I am not going to install the molex and the smds on the main board. I have already soldered the 200R trimpot and I am not going to replace it with a resistor. If everything works fine, I will make all the recommended corrections. As far as the heat of the amp, you can build monoblocks with the appropriate heat sink dimensions.
 
Hi Spriros (if you will allow me to use your name?)
Andy is pretty right; but one must be careful with too much gate resistance.
For the larger mosfets I use 100R (total) and for the IRF mosfets, only 150W, I use 150R.
And yes, Andy's amplifier sounded absolutely sublime, really, really good. At the time I said that it was probably the best sound I had ever heard. However, I designed it, it was my baby, although he did build it very well so maybe neither of us should say it is the best we've heard. Very hard to be objective about the subjective, huh?

Hugh
 
As far as the heat of the amp, you can build monoblocks with the appropriate heat sink dimensions.

But there are 2 concerns re. Class A heat, Spiros:
1. are the transistors/mosfets getting too hot?
and
2. is the room getting uncomfortably hot?

#1 is dealt with by having sufficient heatsinking.

However, it's #2 that I'm concerned with. 3a bias current x 42v across the '+' and '-' DC rails ===> 126w per channel. I have 4 channels - so 504w goes into the room when I'm using my AN 4Rs. Hence the problem in summer.

I would like to reduce this - but not too much, given:
a. my spkrs are 4 ohms, and
b. whilst the AN 4Rs don't clip, currently ... I don't want to significantly reduce their power output - in case this brings on frequent clipping. So I'm hoping to drop my bias current by 10% (to, say, 2.7a) - which would reduce the total heat into the room to 454w ... and extend the time period I can use them, by a month or so.
 
@andyr

I don’t know the exact specifics of your active speaker setup but have you considered just utilizing the AN design for midrange/treble while using a Class AB design for your other two channels?

Although the AN has the best bass from a single ended Class A design I have heard, it is still not the equivalent of some Class AB designs.

I am sure you have contemplated this dilemma given how much waste heat is present in single ended Class A designs and this is compounded when you are designing for 4 ohm loads.

Best,
Anand.
 
@andyr

I don’t know the exact specifics of your active speaker setup but have you considered just utilizing the AN design for midrange/treble while using a Class AB design for your other two channels?

Although the AN has the best bass from a single ended Class A design I have heard, it is still not the equivalent of some Class AB designs.

I am sure you have contemplated this dilemma given how much waste heat is present in single ended Class A designs and this is compounded when you are designing for 4 ohm loads.

Hi Anand,

You might say I do use my AN 4Rs for midrange & treble ... in that each 2-way 'main' hands over to a sub at 110Hz (48dB slopes). So the ANs are not used for much bass!

Each sub has an 800w Class D amp driving it.

And the reason I needed to build the 4R version was that I had 3-way Maggies at the time I decided to build the ANs - and my ribbons were 2 ohms (which the AN 4R handled very well)! When I moved from the Maggies to my 'zero baffle' spkrs ... as I had the 4R version of the ANs ... I had to use 4 ohm drivers. 🙂
 
  • Like
Reactions: poseidonsvoice
But there are 2 concerns re. Class A heat, Spiros:
1. are the transistors/mosfets getting too hot?
and
2. is the room getting uncomfortably hot?

#1 is dealt with by having sufficient heatsinking.

However, it's #2 that I'm concerned with. 3a bias current x 42v across the '+' and '-' DC rails ===> 126w per channel. I have 4 channels - so 504w goes into the room when I'm using my AN 4Rs. Hence the problem in summer.

I would like to reduce this - but not too much, given:
a. my spkrs are 4 ohms, and
b. whilst the AN 4Rs don't clip, currently ... I don't want to significantly reduce their power output - in case this brings on frequent clipping. So I'm hoping to drop my bias current by 10% (to, say, 2.7a) - which would reduce the total heat into the room to 454w ... and extend the time period I can use them, by a month or so.
Andy,
The reason I mentioned the monoblocks is because I have built the Lang 20W amp 2A QC in monoblocks and the heat sink were really "enjoying" that. The temperature had been reduced a lot. Note, that, although I intented to place a fan, it was finally useless because the sinks were just a little warm. After that i had decided to get rid of the 2 large chassis and place everything in one.
The heat had come back but I could put my hands on it for 5 to 10 seconds which I thought It was fair enough. You know that the temperature in Greece during summertime is unbearable.
So, most of the times, I switch the air condition on. My tube amplifiers suffer every summer but they are still "alive and kicking".
I don't think that lowering the QC 0,3A will make a great difference. In my humble opininion, You could think of getting rid of SLBs and their heating mosfets and use the old fashioned way of FWB + CRC filters, calculating the transformers secondaries for 20-21V.
Besides the AN 4R must work in the designer's specs, or else it would be another common class A amp.
Best regards,
Spiros
 
Hi Spriros (if you will allow me to use your name?)
Andy is pretty right; but one must be careful with too much gate resistance.
For the larger mosfets I use 100R (total) and for the IRF mosfets, only 150W, I use 150R.
And yes, Andy's amplifier sounded absolutely sublime, really, really good. At the time I said that it was probably the best sound I had ever heard. However, I designed it, it was my baby, although he did build it very well so maybe neither of us should say it is the best we've heard. Very hard to be objective about the subjective, huh?

Hugh
Hello Hugh,
You can surely use my name.
This is my first build using lateral mosfets. In my previous builds I have used the Hitachi 2SK135 and 2SJ50. All of the schematics with those mosfets and similar ones, use 100R gate stoppers.
In my AN 4R I ordered IXTQ36N30P and IXTQ36P15P from mouser.
As I wrote before, If it work fine, I will make some tests. I can easily add the extra 100R, removing the short from the upper side of the board and listen to any differences.
I will keep my finger crossed till the final test.
Thanks a lot for your comments
Spiros
 
  • Like
Reactions: AKSA
Hi guys. The VgS of my MOSFETs (FQAs) measured in situ powered and playing music isn't exactly balanced. I have 3v3, 3v4, and 4v3,4v6 for the different N and P channels. Should these be matched either by selecting better matched devices or by any adjustments to any values of any other components.?
Or forget about it!
 
Yes sir. Not my work. I got them from a member who'd laid out his own Zen Mod Plethora boards. I figured these little boards would be useful.! There's a few pads on there for different purposes that to be honest is beyond me until I learn more. So they will be shorted for the AN39.
His intention was for the RC type bullet connectors but I wanted to come in at 90deg so I'll use these with a ring connector on the wires.
 
Hi guys. The VgS of my MOSFETs (FQAs) measured in situ powered and playing music isn't exactly balanced. I have 3v3, 3v4, and 4v3,4v6 for the different N and P channels. Should these be matched either by selecting better matched devices or by any adjustments to any values of any other components.?
Or forget about it!
@jimk04

This isn’t a complementary or push pull design in the classic sense. This is a single ended Class A design. The P-channel transistor (and associated circuitry) is the current source and the N-channel transistor is your main amplifying transistor.

The only matching you need is on the KSA992 input transistors both Vbe and Hfe preferably matched.

Thanks,
Anand.
 
Hi Jim,

Anand nailed it; this is a somewhat odd output stage because the upper nmos processes the audio signal and sets the voltage at the load over the entire waveform; whilst the lower pmos mosfet is the variable current source and it ensures the nmos always passes current and sinks current at the negative half cycle from the load.
This is best described as a Single Ended Push Pull - SEPP. It looks like a push pull but
offers a single ended harmonic profile with dominant second harmonic.
Your observation of Vgs nmos v. Vgs pmos comes back to the devices themselves; the two devices have different biasing levels and this can be seen on the data sheets. There is no matching needed here.

Congratulations for building another Alpha Nirvana 39!

Hugh