Compared to what? 😀
Could be that gm of your sit model is a bit off so trip is different.
I've always wondered about this.... saying that one is 20% and the other 80%...
Do they both affect the "sound" the same way?
I mean, is just a little bit of SIT affect the sound more than the MOSFET? Say, just 10%, or are we looking at a linear effect?
Not taking into account stuff like power and speaker efficiency and so on....
Did anyone think about installing a knob in the SIT5 like the SIT1? In this case to affect the balance, say from 10 to 30%, of the SIT's contribution. Or perhaps just a three position switch.
Yes, I did. You can change a few things with a knob in place of R2. Things like amp working class, thd spectra, thd at low power, thd at high power, output power.. actually too many things 🙂
When set to work in class ab a current limiter of some kind would keep things safe.
^ Awesome. How far have you tried it?
You could address the myriad of changes and the non linearity (and poor tracking between pots) by using a switch that selects between resistors that are closely matched between the channels.
In essence, don't use a pot.
Or, how about a plug in carrier for R2?
You could address the myriad of changes and the non linearity (and poor tracking between pots) by using a switch that selects between resistors that are closely matched between the channels.
In essence, don't use a pot.
Or, how about a plug in carrier for R2?
Nelson writes some posts earlier
"....... In the diagram below, the output coupling caps have a DC bias and each carries the bulk of the AC from their attached Fet, but in reverse phase. In this manner the 2nd harmonic each would have is cancelled."
In my Spice the currents in the two output caps are not reversed. Only when I turn one cap about 180° the two currents are reverse....
Oh, this spice has more traps than I thought, I need a Ki....
:--))
I always thought LTspice is clever enough to set at least the DC currents itself in the right way using the difference of the voltages left and right of cap or resistor.
"....... In the diagram below, the output coupling caps have a DC bias and each carries the bulk of the AC from their attached Fet, but in reverse phase. In this manner the 2nd harmonic each would have is cancelled."
In my Spice the currents in the two output caps are not reversed. Only when I turn one cap about 180° the two currents are reverse....
Oh, this spice has more traps than I thought, I need a Ki....
:--))
I always thought LTspice is clever enough to set at least the DC currents itself in the right way using the difference of the voltages left and right of cap or resistor.
A good-ears friend and I recently did an A/B listening comparision of my SIT5 build vs. my SIT-3X (https://www.diyaudio.com/community/threads/the-sit-3x-amplifier.353999/). Other than some sight left/right channel gain differences the two amplifiers sounded the same.
The two amplifiers are very similar, but very different in the way that the PFET contributes to the output AC current.
In post #199 https://www.diyaudio.com/community/threads/first-watt-sit5.418023/post-7886871 I presented a modification to the SIT5 topology to include the Xpot of the SIT-3X. When resistor R2 is shorted (0 Ohms) and the X-pot is adjusted, we get the SIT-3X behavior. I call this the SIT-5X. This hybrid of the SIT5 with the SIT-3X has the advantage of using the SIT5 dual output capacitor and R1 trick of setting the bias current without needing a bias servo. My SIT5 OS PCB layout provided for this option, but only recently did I get around to bench testing it.
Shown below are simulations of simplified SIT5, SIT-5X, and SIT-3X output stages. The waveforms are for 1 Watt@1kHz into an 8 Ohm load. The big difference is that the PFET dominates the output of the SIT5, whereas the SIT dominates the output of SIT-5X and SIT-3X.
The two amplifiers are very similar, but very different in the way that the PFET contributes to the output AC current.
In post #199 https://www.diyaudio.com/community/threads/first-watt-sit5.418023/post-7886871 I presented a modification to the SIT5 topology to include the Xpot of the SIT-3X. When resistor R2 is shorted (0 Ohms) and the X-pot is adjusted, we get the SIT-3X behavior. I call this the SIT-5X. This hybrid of the SIT5 with the SIT-3X has the advantage of using the SIT5 dual output capacitor and R1 trick of setting the bias current without needing a bias servo. My SIT5 OS PCB layout provided for this option, but only recently did I get around to bench testing it.
Shown below are simulations of simplified SIT5, SIT-5X, and SIT-3X output stages. The waveforms are for 1 Watt@1kHz into an 8 Ohm load. The big difference is that the PFET dominates the output of the SIT5, whereas the SIT dominates the output of SIT-5X and SIT-3X.
Shown below are simulations of simplified SIT5, SIT-5X, and SIT-3X output stages. The waveforms are for 1 Watt@1kHz into an 8 Ohm load. The big difference is that the PFET dominates the output of the SIT5, whereas the SIT dominates the output of SIT-5X and SIT-3X.
SIT5 Sweeps
SIT-5X Sweeps (plot titles incorrectly say SIT5, rather SIT5X)).
SIT5 Sweeps
SIT-5X Sweeps (plot titles incorrectly say SIT5, rather SIT5X)).
Last edited:
The final two images show the distortion spectra for 1 Watt, 1kHz, and 8R load, for the SIT5 and SIT-5X.
SIT5
SIT-5X
SIT5
SIT-5X
