That is a very good question. Without the right test socket, matching SOT23 devices can be difficult and unfortunately the semiconductor world is steadily moving away from leaded small signal devices. The demand is already very low for these packages (eg TO 92, TO71 etc) and maintaining old assembly machines is difficult and getting more expensive.
Fortunately, you can get test sockets for SOT23 quite cheaply - I paid £20 IIRC for the one below from Peak Electronic Design (The prices have gone up a bit since then).
Here is the link https://www.peakelec.co.uk/acatalog/pca23-peak-component-adapter-sot23.html
That said, I can confirm I DID NOT match the input devices on the build with the measurements shown in the first thread. For the nx2, match the input devices to eke out the last few ppm of THD performance. I will post some new measurements up later today or tomorrow where I used an external attenuator - the distortion at 100W out into 8 Ohms is 20ppm on the one channel and below 30ppm ppm the other channel. At 50W out, both channels are <20ppm on my latest measurements. If the transistors are matched, you can get 15~20ppm on both channels at 100W. So, I would not fret too much about it because no one will hear these levels of distortion in any event.
Fortunately, you can get test sockets for SOT23 quite cheaply - I paid £20 IIRC for the one below from Peak Electronic Design (The prices have gone up a bit since then).
Here is the link https://www.peakelec.co.uk/acatalog/pca23-peak-component-adapter-sot23.html
That said, I can confirm I DID NOT match the input devices on the build with the measurements shown in the first thread. For the nx2, match the input devices to eke out the last few ppm of THD performance. I will post some new measurements up later today or tomorrow where I used an external attenuator - the distortion at 100W out into 8 Ohms is 20ppm on the one channel and below 30ppm ppm the other channel. At 50W out, both channels are <20ppm on my latest measurements. If the transistors are matched, you can get 15~20ppm on both channels at 100W. So, I would not fret too much about it because no one will hear these levels of distortion in any event.
Last edited:
I doubt that I would hear anything below 0.1-1% range anyhow even if I tried so in real life it is not critical when we are talking a few tenths of ppm THD.
Its just that it is so easy (haha) to buy a couple of extra transistors and match. Well maybe more tricky in this design because it is N/P polarity and hence not from the same strip and inherently different. And in the overall picture not much extra work.
Good tip with the adapter, it is available at Digikey et al for around 30€ ex VAT. I guess it would be possible to find cheap knockoffs at the various Chinese sites aswell.
There is also the option to use miniatyre test clips intended to fit on SMD chips. They are insanely small and require some magnification for most of us 🙂
Its just that it is so easy (haha) to buy a couple of extra transistors and match. Well maybe more tricky in this design because it is N/P polarity and hence not from the same strip and inherently different. And in the overall picture not much extra work.
Good tip with the adapter, it is available at Digikey et al for around 30€ ex VAT. I guess it would be possible to find cheap knockoffs at the various Chinese sites aswell.
There is also the option to use miniatyre test clips intended to fit on SMD chips. They are insanely small and require some magnification for most of us 🙂
I've finalised the nx2 write-up and specifications - link etc in the first post of this thread.
Andrew,
Congratulations again on completing yet another amplifier design! The NX2 looks much more fleshed out than the original NX. I am highly curious how different it would sound compared to a KX2 into lower impedance loads. I would recommend that you upload the drill file spec for the heatsinks to your first post and also remind builders that they can custom order the heatsinks from GianLuca at Modushop drilled for KX2 and NX2 use - am I correct?
Kind regards,
Anand.
Congratulations again on completing yet another amplifier design! The NX2 looks much more fleshed out than the original NX. I am highly curious how different it would sound compared to a KX2 into lower impedance loads. I would recommend that you upload the drill file spec for the heatsinks to your first post and also remind builders that they can custom order the heatsinks from GianLuca at Modushop drilled for KX2 and NX2 use - am I correct?
Kind regards,
Anand.
Hi Andrew,
Please check C4, C6 and C8. These are configured differently compared with the positive supply.
Regards,
VS
Please check C4, C6 and C8. These are configured differently compared with the positive supply.
Regards,
VS
Last edited:
Obviously, I was only referring to page 13. Check again and you'll see - negative vs positive supply.
Last edited:
These look ok to me VS. On the negative supply, the capacitor + has to go to 0V. Or have I missed something here? 🙂
Anand, thanks for the feedback - chassis details added to the first post. The nx2 is a much more capable amplifier in terms of drive capability, low distortion, and noise due to the heftier output stage, increased loop gain, and TMC. The layout has also been much improved compared to the original.
It's not about the polarity 🙂 The decoupling on the positive supply has 1000 uF // 1 uF. The decoupling on the negative side has only 1000μF. The opposite occurs with +/- 10 V Zener regulation.
Last edited:
While it may not be critical, the 1 µF/100 V MLCC at 50 V DC has a capacitance decrease of around 57%. Personally, I would use film capacitors for the intended purpose instead.
https://weblib.samsungsem.com/mlcc/mlcc-ec-data-sheet.do?partNumber=CL31B105KCHNNN
https://weblib.samsungsem.com/mlcc/mlcc-ec-data-sheet.do?partNumber=CL31B105KCHNNN
Enough. If you want to trade insults then go and find somewhere else to play. A few posts have been deleted. First and final warning.
These capacitors are for HF decoupling, and they are 1uF precisely because the capacitance has a voltage relationship. But, since they are at a constant DC voltage, they do the job I want them to do. I would not use film in this location - X7R are great for HF decoupling. But, if you want to discuss capacitors, please start a new thread - this is about the nx2 amplifier 🙂
Nick, the NJL devices offer a little more capability, but they are nearly double the price (NJL1302 is over £10 here). I've been using the NJW for 10 years now with excellent results.
but they have only 0.3 seconds thermal inertia, compared to conventional thermal stabilization with 1-2 minutes via radiators. I tried them, and will never go back to conventional ones, now I use only ThermalTracks [ https://www.patreon.com/posts/99013335 ]