Hi Sjoerd!
The heatsinks originate from a larger heatsink from which I cutted two pieces by using a Dremel multitool. In addition I tapped M3 fittings with a tapping tool. You can obtain nice heatsinks from for example old processor coolers (especially those Pentium I / 486 coolers). To be honest those heatsinks are not really necessary, but hey oversizing things is fun eh ;-).
Hi Kaplaars,
Traded in my 486 a while ago🙂
You are right! It's best when totally overdone🙂
Ps: I think you got the last two Mundorf caps in Europe!
Traded in my 486 a while ago🙂
You are right! It's best when totally overdone🙂
Ps: I think you got the last two Mundorf caps in Europe!
Last edited:
add in a DC blocking capacitor at the input.
Then set the output offset with the DC blocker in place.
It makes a big difference to the variation in output offset.
Mine wandered all over the place while I was still in the testing phase on the bench.
I used a 10uF film dielectric.
The offset still moves around more than any of my previous LTP style inputs, but manageable. Very temperature dependent.
Then set the output offset with the DC blocker in place.
It makes a big difference to the variation in output offset.
Mine wandered all over the place while I was still in the testing phase on the bench.
I used a 10uF film dielectric.
The offset still moves around more than any of my previous LTP style inputs, but manageable. Very temperature dependent.
Hi Andrew, I don't have that problem. Tested the boards and all is steady. I managed 0,2 to 2mvDC
But the NX will be powered on my DIY power conditioner and it is fitted with a DC blocker, just in case.
Sjoerd
But the NX will be powered on my DIY power conditioner and it is fitted with a DC blocker, just in case.
Sjoerd
Nice work Sjoerd! But be carefull with those RIFA's (if they are RIFA's, am not sure). They tend to explode/burn/cause ALOT of smoke after a few years because the dielectric is highly hygroscopic and probably not properly sealed anymore after many years. Those caps caused already caused quite a few times Syrian conditions over here in the lab ;-) Yes, even the X2, Y1 types ;-).
An externally hosted image should be here but it was not working when we last tested it.
.jpg){kind=link}
Thanks for the tip! These are brand new x1 and y1 types. Yes, Rifa. I will replace them in a few years 🙂
Andrew, the DC stability of these simple CFA amps is very good once the initial offset is dialed out - but you must do this AFTER the amp has fully warmed up. Note that from cold switch-on to fully warmed up a class A sx-Amp will take about 30 minutes. Only after its warmed up can you properly adjust the output offset to zero. On subsequent cold switch-on's, about 80% of the offset will disappear in the first 4-5 minutes and the rest over the next 25 mins.
The offset from subsequent cold switch-on's is then about 50mV, after 4-5 minutes about 10mV and after 30 minutes it should be low single digit - 2-3mV is about right.
If you suspect your source has any DC offset, you should use a DC blocking capacitor. 10uF is a good value and will ensure that the response at 20 Hz is not down by more than -0.1 dB
Sure, glad I could help 🙂. I have had very good results with those blue plastic ones (those you also have). These are really reliable, especially the BC ones. I think it has to do something with the sealing compound on those yellow trigger happy caps that somehow turns from the glass state to a mere liquid state after years. This makes them permeable to water and leading to problems.
Very curious about the DC-offset. Worst case scenario; it is then maybe not convenient to add a DC servo Andrew? Or could you improve it a bit by getting the sense transistor closer to the heatsink (that little BC847 if I recall correctly)?
I have currently all parts soldered except the output transistors. But I first need to find a cabinet for that (and some $$$ hehe). Here some pics:
Very curious about the DC-offset. Worst case scenario; it is then maybe not convenient to add a DC servo Andrew? Or could you improve it a bit by getting the sense transistor closer to the heatsink (that little BC847 if I recall correctly)?
I have currently all parts soldered except the output transistors. But I first need to find a cabinet for that (and some $$$ hehe). Here some pics:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Looks good Kaplaars! I hate the fact that you got those really nice caps on your psu ;-) Mine are in backorder.
Case is REALLY expensive. I was lucky to get hold of one that was leftover at Modu. So I payd a bit less. Still a steap amount though...
Sjoerd
Case is REALLY expensive. I was lucky to get hold of one that was leftover at Modu. So I payd a bit less. Still a steap amount though...
Sjoerd
DC servo's are very good - but both the sx and nx amplifiers are minimalist designs. Once you have set them up correctly they are fine and will perform without issue.
Note that on the nx-Amp, the offset settling time is much quicker
Nice looking boards by the way!
🙂
Note that on the nx-Amp, the offset settling time is much quicker
Nice looking boards by the way!
🙂
it is not anything to do with DC from the source.
It's the instability of the single ended input stage with respect to temperature.
This is compounded by sensitivity to the source impedance seen by the DC coupled connection.
The pair give rise to an output offset that wanders all over the place compared to an LTP style input stage that is AC coupled.
I had never built a DC coupled amplifier with a single ended input. The NX proved to me how bad that combination is.
Last edited:
That is a shame. But, problems are there to be solved; how could this problem be tackled according to your expertise? You suggest that it is a problem intrinsic to the design. However, from what I read in this topic you seem to be the only builder that has this problem. Could there not be a different explanation?
I currently have not the facilities to check how my NX behaves, but I am curious to what causes this. Have you tried to improve it by for example putting the sense transistor closer to the heatsinks? Is everything matched etc.?