Hifisonix NX2-Amplifier: 100W CFA Power Amp

Jump to Latest
Bonsai

Bonsai

www.hifisonix.com
Joined 2003
hifisonix.com
Paid Member
#1
Attached below are the preliminary write-up and some measurements for the new nx2-Amp.

Specifications :-

1749316578365.png
IMG_4997.jpeg
IMG_4918.jpeg
 

Attachments

Last edited:
  • Like
Reactions: VIPINSAINI20, maxlorenz, GumGum9000 and 16 others
#4
63dB Loop gain at 1K is pretty conservative for TMC. I usually get similar figures with single pole compensation.
It might be some compromise made for the 2.2uf capacity load.
Have you ever encountered such capacitive load in real live? Usually, I don’t verify beyond 0.1uf.
 
  • Like
Reactions: lineup
#11
Bonsai said:
I'm putting this up as a place holder for the new nx2-Amplifier (hereinafter just called the 'nx2') which has been in gestation for some 15 months now. I've had quite a few nudges from various people via email asking about progress. I can understand that some folks have fallen into believing the nx2 is 'vapourware' (if you've worked in semiconductors or software product development, you'll now what that means!).

I started work on this in September 2023, assembled the initial proto and decided it needed more refinement - details which I will cover when the documentation is released. However, things came to a grinding halt when TraxMaker (part of the CircuitMaker CAD suite) stopped working in early 2024. I made the decision to switch to KiCAD (a good move!) and had to redesign the board from scratch. After another 2 proto/dev cycles, the amplifier module PCBs are finally where I want them to be, and quick bench testing shows peak mains hum harmonic components at below -125 dBr reference 60W RMS into 8 Ohms (the power limit of my test lash-up) and <35ppm distortion.

I am currently finalising the PSU +inrush PCB and that will go off for manufacture sometime next week. In parallel with this, the mechanical layout is also being finalised.

Part of the delay in this project has also been the fact that I became a grandfather (yeah, I'm old) for the second time and that entailed putting a lot of stuff on the backburner, and the decision to move closer to family, which we will do in the next 3 weeks. Its chaos here with boxes all over the place, packing etc.

For all of the above reasons, the nx2 won't be completed and released for at least another 2 months, but hopefully you wiol get an idea of whats n the pipeline from this intro. The boards will be available on the hifisonix web shop if you want to build one (but only after I release the completed amp + documentation).

Here are the specs (note these may change once the completed, housed amplifier is re-measured)

View attachment 1443516

I've attached some photos below - taken in my workshop today in amongst the packing boxes. (note the front plate is still being assembled so looks a bit manky - it still has to be cleaned, screws fitted and then polished)



I’m sure this will be a nice update to a strong design. I’ll be interested in whether you’ve kept the simple current limiter approach or gone for a SOA VI detector to feed the SSR module.

View attachment 1443518
Click to expand...
 
#12
jxdking said:
63dB Loop gain at 1K is pretty conservative for TMC. I usually get similar figures with single pole compensation.
It might be some compromise made for the 2.2uf capacity load.
Have you ever encountered such capacitive load in real live? Usually, I don’t verify beyond 0.1uf.
Click to expand...
The nx2 is a CFA - the loop gains are generally lower than blameless type VFAs.

A worst case load I usually cater for is an electrostatic which usually lie between 1uF and 3uF worst case load. I agree most conventional speakers are nowhere near this.
 
  • Like
Reactions: akimmet and jxdking
#20
I have read through it more in detail and have a thought that popped up from a practical point of view, that we probably can adress eventually in the forum.

Switching to SMD is necessary due to parts availibilty I presume, but matching SMD small signal transistors require some form of jig . I wonder how to setup that?

I know there is ZIF socket adapters for testpurposes for 3-pin SOT packages such as SMD transistors (both small signal and power) but they are probably quite expensive. It is easier with TO-packages.

A SOT 23-6 (6 pin for tiny as h-ll MCU) programming adapter easily runs of at 100Euro/piece.

It cant be the first time this challenge shows up and I am sure the capacity of Diyaudio have the answer, apart from poor ba-d that sits in his lab matching and selling in the Swap meet section 🙂
 
Last edited:
You must log in or register to reply here.
Top Bottom