In this schematic of a Roland Blues Cube guitar amp there are many instances of diodes in series with FET gates (D1, D2, D3, on page 6). Can someone explain their function? I don't understand how they are biasing the FET since no current should be flowing in the gate.
roland_blues_cube30_60.pdf
roland_blues_cube30_60.pdf
Normally the leakage current of the diode is sufficient to bias the FET. It looks 'wrong' but that's how it works.
I knew we had had this issue crop up before. This was a simulation of a similar type of circuit, however the simulation relies critically on all aspects of the devices being modelled correctly.
PRR elaborated on the technique a few posts further down:
DC offset on NAD 3120 preamp output
PRR elaborated on the technique a few posts further down:
DC offset on NAD 3120 preamp output
Those are switching FETs.
These ones are N Channel (notice the internal arrow pointing from Gate to Channel which is the doped Silicon strip joining Source to Drain); explanation is same for P channel Fets but reversing polarities.
"Zero biasing"=0V applied to Gate means the same on both cases because Zero has no polarity or sign.
** To *open* switch (so no audio gets through) you need to apply negative voltage to the Gate, above a certain value called Vp=V pinchoff.
The name is self explaining: imagine a water/air hose where you squeeze/"pinch" it so much that flow stops.
Any negative voltage above Vp will stop electron flow and so open the Fet switch.
Typical Vp values hover around -5V and you can safely apply -10 to -15V , so they do that.
** To close switch you simply "do not pinch it off" so leaving Gate open is enough.
2 particular cases which are often ignored:
a) even a negative voltage can leave switch open, IF it´s not enough to "pinchoff" .
That´s why some amps squeal or switch to distorted channel or unmute or ... when turned off: supply voltages drop taking a few seconds and at some point "all Fet switches turn on" while amplifier can still work "a little" and squeal.
b) Gate can even be biased positive, switch will still close.
Won´t damage Fet but generally should be avoided because some switching voltage can "get through" and appear as an annoying "thump" mixed with signal.
Both problems can be solved with the reverse oriented Gate series diode:
* if voltage is Negative, it goes through with no problem, reverse biases Gate and turns Switch OFF.
Since Gate is reverse biased, no current goes through so no Thump.
* if voltage is Positive, Diode blocks it, again no Thump, Gate "floats" , in principle at or near 0V and Fet switch closes.
Remember a Fet switch is not perfect; when ON it does not show 0 ohms but some channel resistance, think from 50 to 200 ohms, sometimes even a little more.
Both zero bias channel resistance and Vp are shown in the datasheets, although sometimes in a cryptic way 🙄 ; but the generic values I wrote above will give you an idea of what to expect.
In general series open switch function is reasonably good; grounding signal not that much because of residual resistance.
If used to activate a Gain or Master control, basically same as a regular Volume potentiometer, you may notice it does not *really* reach zero volume, a little audio remains even on 0 ... that´s the nature of the beast 🙄
Personally I stopped using Fets as mutes in my own designs, a humble saturated bipolar transistor does it better because it can show way lower dynamic impedance (think 10 ohms instead of 150) but to each his own.
I bet Roland can order specially matched lower resistance Fets 😀
Thank you for the explanation. The effect of the diode on the gate voltage seems small to me. I ran a simulation of this type of FET switch with and without the diode and there is about 0.25V difference in gate bias between the two in the switch ON state (at around +50V before the 1Meg supply resistor). But that might be a significant difference, I don't really know. Is there a maximum positive Vgs for JFETs?
In the switch OFF state, with the diode in the circuit, the diode on voltage is tiny (microvolts) owing to the tiny reverse gate current of the FET.
The diode 1ss133 is special . It has reverse current 0.5ua leakage up to 80v ,which permits the fet to discharge its gate -channel capacitor and switch back on .
Always speaking of Nch Fets (reverse polarities for Pch ones), "maximum positive Vgs" does not apply, because it´s a plain diode, a plain PN junction if forward biased.
I will plain pass forward current like any forward biased diode and drop what a diode drops, less than 0.6V .
Way less because current is tiny.
There's further discussion on this switch topology here:
Curious about Passing Audio Signal with a FET Switch (Simple - No Flip Flop)
Curious about Passing Audio Signal with a FET Switch (Simple - No Flip Flop)
- Status
- Not open for further replies.