Announcement

Collapse
No announcement yet.

Using Jfets for switching in tube amps - high voltage concerns

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by uneumann View Post

    Sure - zeners would help, but the transient coming through the cap is potentially full B+ voltage. Yes, current is limited, but voltage can fry semis at low currents - as with static zaps.
    In a preamp stage, as we are talking about here, there will no such (switch on ) full voltage B+ transient. The rise time will be very slow due to the filtering action of the preamp supply. A 10K/22uf is typical. Thus the current will be something like 100uA but more importantly the power delivered to the JFET will be quite low. ESD's zap semis because the transient is fast so the power is high but of short duration. JFETs being bipolar are quite tough.


    There are options that seems better and safer. The DN3545 is a mosfet with 10 ohm on resistance and rated at 450 volts. Or how about solid-state relays - they're rated at 300-400V. I suspect JFETS are attractive because they're cheaper. But if you really want to be sure something is safe and reliable, maybe cheaper isn't the only criteria to consider.
    A MOSFET is not without merit but remember the gate is very easily destroyed on these so care needs to be taken in handling and circuit design. They are also more expensive, and yes, the low cost of a JEFT is an big attraction.

    If you run a JFET without zener protection I don't see any harm coming to JFET due to the low current. The drain / gate junction can be forward biased on an overload and this needs to be considered in the gate drive. This will also lead to the the coupling capacitor charging and that could have unintended consequences. In the other direction lies the breakdown region but with low currents and so reversible

    SSR are surely most the most robust but cost 20x to 40x that of a JEFT on quick look.

    It seem to me that provided the largest voltage swing is a volt or two less than the used gate off voltage, JFETs are a great solution, zeners a matter of preference I think. If the swing is greater go for a MOSFET but protect the gate. For easy to use, go SSR.

    EDIT: On reflection a single MOSFET won't work due to the intrinsic body source-drain diode. They are only suitable for unipolar applications.
    Last edited by nickb; 02-20-2020, 06:21 AM.
    Experience is something you get, just after you really needed it.

    Comment


    • #17
      Originally posted by Gaz View Post
      Thank you, Nick. Working out this real-world example for me is really helpful. I didn't know how to do the math to find the current. So even though there is a transient of 400v (for example) because it's only at 40 microamps it will likely be okay? I guess I don't really understand the relationship...

      Thanks again
      I didn't mean to hit send on that one as it was half baked but I see that I did!
      Experience is something you get, just after you really needed it.

      Comment


      • #18
        SSR are surely most the most robust but cost 20x to 40x that of a JEFT on quick look.
        It's more like 3x to 4x which is not much at all. Here are some high voltage DIP4 SSR prices compared to a typical J175 mute JFET.

        https://www.mouser.com/Electromechan...e&Ns=Pricing|0

        https://www.mouser.com/_/?Keyword=J175&Ns=Pricing|0

        Comment


        • #19
          Cost isn't a big factor to me, I mean the whole amp probably cost $800 in parts or more, but I do understand of course why the big companies do it.

          What I'm wondering is if I can use this same Peavey/Mesa "trigger" circuit with a higher voltage device for muting, or will this only work with a P-channel FET? Although Nickb has pointed our the DC transient is probably not a big concern, there is so much anecdotal evidence of these fets failing I'd like to find a less delicate mute circuit. Plus, being able to handle voltage peaks above the gate voltage would be great. Then this could be used in "vintage" style amps with switchable elements where the signal voltage stays high all the way to the PI.

          Could I use the same triac trigger to close an SSR for example? Having never used one, I'm assuming the SSR's are faster than mechanical relays and click-less.

          Comment


          • #20
            What I'm wondering is if I can use this same Peavey/Mesa "trigger" circuit with a higher voltage device for muting...
            Yes, you can use it. Connect SSR's LED cathode where JFET gate is connected. Run a 1k5-2k2 resistor from SSR's LED anode to 12V supply and then connect the switching part in parallel with the sound path. Add back to back zeners for extra protection (optional). Done.
            Check out the datasheets for the switching part capacitance but it's usually very low like 5-10pF which is inaudible.

            Comment


            • #21
              Originally posted by Gregg View Post
              Yes, you can use it. Connect SSR's LED cathode where JFET gate is connected. Run a 1k5-2k2 resistor from SSR's LED anode to 12V supply and then connect the switching part in parallel with the sound path. Add back to back zeners for extra protection (optional). Done.
              Check out the datasheets for the switching part capacitance but it's usually very low like 5-10pF which is inaudible.
              Gregg, thanks for the reply, and I didn't expect that! Yeah, no sense in worrying about the low voltage Jfets then! I'm a little perplexed why bigger companies don't use an SSR triggered by a TRIAC. It would seem to give a lot more freedom. Expense I suppose?

              And with a 400V part, zeners do seem unnecessary right? I'm confused about value zener would even protect a part rated so high (sorry for my ignorance if I'm missing a potential failure point here).

              Thanks again.

              Comment


              • #22
                Originally posted by Gregg View Post
                It's more like 3x to 4x which is not much at all. Here are some high voltage DIP4 SSR prices compared to a typical J175 mute JFET.

                https://www.mouser.com/Electromechan...e&Ns=Pricing|0

                https://www.mouser.com/_/?Keyword=J175&Ns=Pricing|0
                Lucky you! Here in the UK cheapest SSR was about $2.00 yet I can get JFETS for $0.05.
                Experience is something you get, just after you really needed it.

                Comment


                • #23
                  yet I can get JFETS for $0.05.
                  Where do you buy JFets?
                  - Own Opinions Only -

                  Comment


                  • #24
                    $0.05 is very low for anything.

                    I checked Farnell's prices and they look kind of same as Mouser's:

                    https://uk.farnell.com/panasonic-ele...503695?st=aqy2

                    https://uk.farnell.com/search/prl/re...d&sort=P_PRICE

                    Comment


                    • #25
                      Originally posted by Gregg View Post
                      $0.05 is very low for anything.

                      I checked Farnell's prices and they look kind of same as Mouser's:

                      https://uk.farnell.com/panasonic-ele...503695?st=aqy2

                      https://uk.farnell.com/search/prl/re...d&sort=P_PRICE
                      Which equates to USD2.00 as taxes are 20%

                      I checked my inventory for JFETs and that was the price I paid. Mind you, I see it was a reel of 3000 from a few years back. In current prices $0.15-$0.20 might be closer for small qtys. That's still at least 10x.

                      Anyway, the OP doesn't care about the price so it's of no consequence.
                      Experience is something you get, just after you really needed it.

                      Comment


                      • #26
                        Originally posted by Gregg View Post
                        Yes, you can use it. Connect SSR's LED cathode where JFET gate is connected. Run a 1k5-2k2 resistor from SSR's LED anode to 12V supply and then connect the switching part in parallel with the sound path. Add back to back zeners for extra protection (optional). Done.
                        Check out the datasheets for the switching part capacitance but it's usually very low like 5-10pF which is inaudible.
                        Gregg, got the Panasonic SSR and hooked it up like this, however it's not working quite right. When the panel switch is thrown that swithes the mechanical relays (and triggers the TRIAC), the sound gets muted, but stays muted, rather than just muting the "pop." If I wait several seconds I can flip the switch again to switches channels, but then it gets stuck in mute. Any ideas?

                        Comment


                        • #27
                          Firstable lease check if everything is connected properly. Stays muted for how long? Longer time constant?
                          Disconnect SSR from the triac and check manually if the if it's working properly (repeatedly short SSR's LED negative end to gnd). If it doesnt get stuch then follow back down the circuit.

                          Comment


                          • #28
                            Thanks. Ok, so everything seems to be hooked up correctly with the SSR. I can disconnect from the triac and mute the signal as I ground LED. But when I hook up the triac it doesn't mute fast enough to stop the relay click and it gets stuck, muted. If I cycle through the channel switch a few times it will unmute, but the mute again.

                            I also noticed if I removed the 12V supply from the Triac the same behavior continues. Did I destroy the triac perhaps?

                            Comment


                            • #29
                              Probably not. If you can break the circuit to release it, then it activates the next time you chanel switch, then it works. The triac won't release until the current drops to zero. Perhaps your SSR is a current path, the LED inside perhaps?

                              That resistor isn't there for teh triac, which will conduct just fine without it. The resistor is there to provide gate voltage to hold the JFET OFF.

                              You do realize, I hope, that the triac grounds the gate voltage of the JFET to allow it to turn ON.

                              What does the data sheet say about the rise time at turn on?

                              I'm a little perplexed why bigger companies don't use an SSR triggered by a TRIAC.
                              Why would they? The part costs a lot more, and the JFET works just fine and is super simple.
                              Education is what you're left with after you have forgotten what you have learned.

                              Comment


                              • #30
                                The 1M/100nF combo has a 100msec time constant which means that after it shorts the gate (LED in your case) to ground it should release shortly after that.
                                With disconnected SSR measure the voltage at the 1M/100nF junction. Then try a regular LED with resistor to 12V and cahtode connected as I suggested earlier and let us know.

                                Comment

                                Working...
                                X