I flipped the power supply leads and the capacitor terminals to obtain the coveted negative DC 

You cannot view this attachment.

I have to remind myself that this stuff is not gonna do what I want it to do

Anyhow moving forward now maybe I can try to figure out a combination capacitor and resistor combination that would take the longest to discharge

Or a combo where I can input DC on one side and see it on the other side then slowly dissapate 

I had hit a brick 🧱 wall with this and needed some ideas so I tried it without the resistor just an open circuit

When I input DC now it's reflected on the other side of the capacitor

You cannot view this attachment.

Essentially Vin now equals Vout

You cannot view this attachment.

That said I bet now if I close the circuit that one side of the capacitor will discharge and be at zero

So I'm guessing my frustration at not being able to see isolated iterations of each sequence of the AC input circuit is because when I simply add DC and hold it there the other side discharges
Whereas input AC it's changing before the other side can react

Quote from: J M Fahey on August 21, 2026, 02:51:58 AMhat is somewhat common in MosFet amps for efficiency reasons.

I seem to recall that the MosValve 500, it is taken to a ridiculous degree. The VAS is something like 90+V dual power rails to 60-something for the output stage.

I was doing the diode clamping experiment and got in over my head so I removed the diode to make things simpler

This is what I had

You cannot view this attachment.

Vin and Vout are almost in phase

When I connect the signal generator to the input I clearly observe Vout swinging from positive to negative ad infinitum both on a scope and multimeter

In an attempt to observe each sequence individually I input positive DC to Vin

But get nothing for Vout

I know yeah capacitors block DC but trying to figure out how AC can influence Vout but not DC

 

Quote from: Kaz Kylheku on August 20, 2026, 01:28:22 AMThe MosValve 500 has an interesting take on it. There is a voltage amplification stage which runs on significantly wider power rails than the output stage. Negative feedback is taken from the voltage amplification stage, not the speaker output.

That should have two effects: one is that the impedance of the MOSFET output devices is not hidden from the speakers; i.e. it raises the output impedance a bit ("tube like").

Secondly, the MOSFETS will clip long before the VAS does and their clipping is outside of the NFB loop. Clipping being out of the NFB loop means that we don't get the hard clipping where the output is forced to stay linear until the very limit; the output stage is "allowed" to produce nonlinearities before clipping. This is again like a tube output stage without NFB.

That is somewhat common in MosFet amps for efficiency reasons.

Mosfets require a STRONG drive signal because you always waste part of the drive signal just turning them ON.
In bipolars, Vbe at top current (say 5 to 10 A, is a mere 1V, no big deal, so when drivers clip, emitters are only 1 V below.

But Mosfets are hard to drive, Gate typically needs 6-8 Volts    to reach such currents, just look at datasheets, which is unacceptable.
40V rails with bipolars can give you 38 Vp , so 26V RMS if supply is stable, so 85W RMS into 8 ohm (in theory). In practice, 60/8 because supply drops some 15%

Exact same supply and speakers, you can only get 32 Vp so 22 V RMS so 60 W RMS
With a real power supply which drops, some 40W RMS.
Power loss is significant so a good designer adds an extra 8-10V supply, floating on main rails

Here is VOX Q series MosFet power amp.



Notice there is a main split winding feeding D1 to D4 bridge and main filter caps PLUS an extra low voltage winding floating on it, feeding D5 and D6 which create extra 8-10V rails for the drive section, Tr2 and Tr3.

This very simple amp was very loud, lots of headroom.

Quote from: saturated on August 20, 2026, 09:32:22 PMAlso put probes across the diode to watch when it gets busy 

Why dont ya use an LED 

One less meter display to watch 

Looking at my signal generator I noticed wow I can really slow things down

You cannot view this attachment.

Now I'm wondering if I could construct the clamping circuit shown connect meters and actually see in real time voltages changing ala capacitor charging and discharging in real time 
Also put probes across the diode to watch when it gets busy 

Now I wonder is 0.03 hz gonna still be too fast to see 🙈

Or

Is the signal generator actually gonna deliver that low

The MosValve 500 has an interesting take on it. There is a voltage amplification stage which runs on significantly wider power rails than the output stage. Negative feedback is taken from the voltage amplification stage, not the speaker output.

That should have two effects: one is that the impedance of the MOSFET output devices is not hidden from the speakers; i.e. it raises the output impedance a bit ("tube like").

Secondly, the MOSFETS will clip long before the VAS does and their clipping is outside of the NFB loop. Clipping being out of the NFB loop means that we don't get the hard clipping where the output is forced to stay linear until the very limit; the output stage is "allowed" to produce nonlinearities before clipping. This is again like a tube output stage without NFB.

Yes, in SS aamps power  amp is just a BIG Op Amp, and clips as such, boring.

A few SS buikders such as Peavey Transtube, Pritchard, Quilter, Roland Blues Cube have power amps emulating tubes, but they are the exception.

So I got to thinking wow maybe I have a (fixed) bipolar power supply at my house

And I do  it's called a Peavey Audition Chorus 

Looking at the schematic it looks like if a bridge rectifier is constructed from four diodes then
-neg voltage can be obtained from the node in between the anodes
-positive voltage can be obtained from the node in between the cathodes

So I'm looking forward to making a circuit like that and checking it out