bjosephs
Well-known member
I know this is very much a recording equipment forum but I have a guitar amp application that I’m looking for the benefit of experience on.
I’m working on a programmable tube preamp. Part of that is achieving some way of switching cathode capacitors with a microcontroller. Jumping past LDRs and Relays due to size I’m investigating the use of P channel JFETs as used in many Mesa and some Marshall amps.
The J175 (rds on max of 125 ohms) is the most common I can see in schematics, though I’ve found a Marshall DSL using J174. The “problem”, if you can call it that, is that the guaranteed Vgs off of these parts is 6v or higher. And a microcontroller is going to be toggling 0 to 5v natively. So they won’t be fully “off” but perhaps they get close enough?
I can use an NPN switch to interface but I have found the J176 which has a max Vgs off of 4v in exchange for a higher rds on of 250 ohms max. So I wired a couple up and collected some characteristic data which is attached. I measured their switching capability as raw resistance and as a voltage divider since there will likely be a small positive voltage on the drain from leakage through the cap - not sure how that affects the curves. And in the interest of seeing how good of a transistor switch is need I rechecked the V and R across the FET with a residual .5v on the gate to simulate a weak pull down.
I’m curious if others that have used JFETs this way can comment on their implementation choices, the role of part-to-part variability in deciding which of these options is best, and if thermal or voltage variation should be of concern. My gut says that as long as the FET gets above, say, 20k consistently that’s a good enough “off” and anything approaching 100 ohms is a good enough “on”. It doesn’t appear that I need any sort of gate resistance and could tie the uC pin directly to the gate.
I’m working on a programmable tube preamp. Part of that is achieving some way of switching cathode capacitors with a microcontroller. Jumping past LDRs and Relays due to size I’m investigating the use of P channel JFETs as used in many Mesa and some Marshall amps.
The J175 (rds on max of 125 ohms) is the most common I can see in schematics, though I’ve found a Marshall DSL using J174. The “problem”, if you can call it that, is that the guaranteed Vgs off of these parts is 6v or higher. And a microcontroller is going to be toggling 0 to 5v natively. So they won’t be fully “off” but perhaps they get close enough?
I can use an NPN switch to interface but I have found the J176 which has a max Vgs off of 4v in exchange for a higher rds on of 250 ohms max. So I wired a couple up and collected some characteristic data which is attached. I measured their switching capability as raw resistance and as a voltage divider since there will likely be a small positive voltage on the drain from leakage through the cap - not sure how that affects the curves. And in the interest of seeing how good of a transistor switch is need I rechecked the V and R across the FET with a residual .5v on the gate to simulate a weak pull down.
I’m curious if others that have used JFETs this way can comment on their implementation choices, the role of part-to-part variability in deciding which of these options is best, and if thermal or voltage variation should be of concern. My gut says that as long as the FET gets above, say, 20k consistently that’s a good enough “off” and anything approaching 100 ohms is a good enough “on”. It doesn’t appear that I need any sort of gate resistance and could tie the uC pin directly to the gate.
