After more resistor and jfet parameter tweaking of the circuit in post #143 I appear to have achieved even better (simulated) performance than Nelson posted: DF=132, THD=.018%, -3dB@200kHz. These numbers are somewhat meaningless until I can make an actual circuit that is close in performance.
I also have (some ) circ in simulation , being "better" than Papa's
off course that doesn't mean nothing , vs. Real McCay
off course that doesn't mean nothing , vs. Real McCay
I just dug around in my mosfet stash and found a pair of j162s and a pair of k1058s. I guess I better start wiring them up.
Looking at the datasheets, they do not have the right temperature coefficient which would lead to problems without source degeneration or something to prevent thermal runaway.
taking in account Iq in possible range of anything between 1A2 to 1A8 ?
neither of them is having zero point in that range
neither of them is having zero point in that range
I based my belief about the thermal coefficient on the Id vs. Vgs curves in the datasheet 2SJ201 pdf, 2SJ201 description, 2SJ201 datasheets, 2SJ201 view ::: ALLDATASHEET :::. It looks like Id increases significantly with temperature at the expected operating point. There might be other evidence to the contrary that I do not know about.
I believe nothing to worry about , if you - as usual - make final Iq setting at temp. equilibrium
thermal runaway is much greater problem in AB class amps with eenyweeny heatsinks , than in A class amps with big chunk of Al
though , you're most probably right , 1058 (and cousin) are more probable candidates , taking in account their neg. TempCo at Iq of interest
thermal runaway is much greater problem in AB class amps with eenyweeny heatsinks , than in A class amps with big chunk of Al
though , you're most probably right , 1058 (and cousin) are more probable candidates , taking in account their neg. TempCo at Iq of interest
Yes, I'm not surprised you can do this (still impressed though), but for us dumb dumbs, I don't think it would be a matter of building a circuit without them and expect it to work first go.
It would take a fare amount of trial and error and development time (or sheer luck) to get it to work, which most people here aren't prepared to do.
Is this the reason you spent quite a bit of time on reducing RF and high frequency CD noise?
Or unless you are using a trimpot in a tricky arrangement to achieve gate resistance, biasing, and loading of jfet input (three birds with one stone) and calling it no gate resistance (I can think of an arrangement to achieve this)
You could interpret that a few different ways. You are probably right about the feedback resistor though.
My interpretation was that this was his initial approach, then he decided "there is no challenge in this, even 2 picodumbs can do that (hahahaha). I need to do something different, I am going to build a minimal parts count amp that the cloners won't have a chance in hell of reproducing, muah hahahahaha."
So he stripped it back to a single pair of outputs and got rid of the cascode, removed every part and started reducing RF and high frequency CD noise, then only added back to the circuit what was necessary to make it work.
Last edited:
I didn't mention it earlier but if you absolutely want to match Nelson's harmonics (I don't think it's worth worrying about, it's still low and non-offensive) you can easily achieve it with a trimpot at source or jfets.
Last edited:
In my simulations I was able to match Nelson's harmonics by reaching into my bag of k170 JFETs and selecting one with just the right Idds, accomplishing the same result as using a trimpot.
At this point my circuit has only 7 resistors. Did Nelson really use only 5?
Aaah but did Nelson go for positive or negative phase second harmonic?
Hahahahaha
My recollection is P Channel Jfets (2sj74) has slightly more transconductance than N channel Jfets at same Idss value, and my Lateral mosfets measure higher transconductance on N channel devices giving a strong negative phase 2nd harmonic in this circuit (which is perfectly fine) but is that what Papa has?
I'm not concerned about it, just want to give you a headache. Hahaha 😀
Last edited:
The F7.pdf article shows negative phase H2 in "Distortion Waveform at 1W/1kHz" plot. And my earlier simulations had negative phase H2 at 22dB above H3. With careful Idss selection I have negative phase H2 at about 9dB higher than H3 and a THD of 0.018%. SIMULATIONS ONLY.
I missed that bit in the article (always pays to read article 10 times). At least my prediction was correct, Hahahaha. 😀
I personally would actually prefer -60dB of second harmonic not less but that is just me (I'm not building to please other people).
Hi Ihquam
After reading the article again, I still don't see the description of negative phase 2nd harmonic.
Could you explain what I am not seeing?
Thanks
🙂
Yes but I don't see the negative phase harmonic.
Don't be shy, just remember my brain cells aren't large in number, if you have some intelligence on this matter feel free to share.