channel switching jfets

[lowell]

I'm pursuing jfet switching in my next amp. Can someone review my arrangement in the schem? I cannot find the jfet symbol so I'm using the MOSFET symbol instead. I think the J112 will work. I'm going to tap off the bias supply w/ a diode, resistor, 15v zener, and cap for the negative control supply for the gate. My main concern is if M2 will cause any issues being in series w/ the clean channel signal. I'm new to jfet switching at this point but do have KOC's 1st book.

[J M Fahey]

Just do it and *you* tell us.
Thanks.

[lowell]

JM,
I'm building an entire amp. I'd like to have it all planned ahead of time AS MUCH as I can so I don't run into what I did on the last amp I made. I kept modifying and modifying, adding turret boards etc... I understand that it's unavoidable in new designs to some extent, but ampage being a great source of information and experience I'd like to get as much advice BEFORE I start building this amp. I want the layout to be RIGHT from the start. If you don't want to help that's fine. I'm not being lazy, I'm doing research in an attempt to save myself future headaches.

[d95err]

I haven't done much jfet switching, but I've read some of Kevins books.

First, M1 is going to mute both channels when "on".

I think there is also a problem with how the jfets are connected to -16V. You should not connect your control voltage directly to the jfet gates. You should have a resistor (e.g. 1Meg) or a backwards diode to ensure the gate sees only a very small current.

When the jfet gates are not seeing -16V, they should preferably see 0V or a positive voltage. If you disconnect them completely like in your schematic, you may get unpredictable results. I think it's even a bit more complex when using a series jfet.

Another thing I would be slightly concerned with are the voltages the jfets will be seing. Jfets typically break down at 40V or so. I can see a risk of the signal being bigger than that on both channels (in particular on the clean channel if someone pushes it with a booster or OD pedal). You must be absolutely sure the peak signal can never exceeds the maximum voltage of the jfets.

[loudthud]

JFETs sometimes switch slow if you just open circuit the gate. I like to use a resistor, 1 Meg usually works fine. The diodes (1N914) are needed so the "shunt" connected FETs don't interfere with the "series" connected FETs that share the same control line. The -25V control signal insures the FETs will still be Off with +/-15V signal. Excessive signal levels will damage the JFETs so signal clamps are needed.

[lowell]

Beautiful drawing and thanks for the info LT. If you don't mind can you take a look at my mosfet post for my PA? http://music-electronics-forum.com/t17120/

[J M Fahey]

Dear Lowell.
I think you did not get it, maybe I was too concise and thought you would get the message, specially considering your vast and undenied experience in Forum matters, both here and elsewhere.
Besides, I think you already know I *do* help all around, *if* within my means, I love the challenges, but also tend to rely heavily on personal experience and experiments, and much prefer to say nothing, rather tan "a friend of a friend of a friend heard that .... ".
Now to the technical doubt:
Fets (and MosFets and Multiplexers such as CD4053, CD40x6, DG212, etc.) are not normally used in tube circuits, simply because they are great switching up to the supply voltage (15Vpp) or up to (Vgate-Vp) , Vp being Vpinchoff, the Fet channel cutoff voltage.
Considering some Fets stand up to -40Vg , having Vp around 4 to 6V, they may control up to 34Vpp; better but no cigar.
This was stated correctly by d95err.
One unsatisfactory way around is clamping the signal , as Loudthud suggested, because it fights feedthrough well, but what does "clamping" mean, in practice?: clipping with diodes !!!!
Zeners or regular diodes connected to a certain clamping voltage end doing the same: soundwaves sharply clipped, closer to squarewaves than anything else, the opposite of what you hope to achieve with tubes.
I have been experimenting a lot with SS switching, trying to help a very good friend who wants to build a fully programmable "no pots" preamp, each pot being a 32 step switched resistor array, replacing all regular pots in the regular circuit, no padding , clamping, nothing.
Result?: utter, complete failure (so far).
The simple solution, padding levels to values that can be handled and re-amplifying them later, so each tube receives the same signal as in the classic preamps, kills the tone.
But, hey!! , maybe lowell makes it work !!!
Many times I've seen others succeed where I previously failed , and I know for sure you are a steady contributor here so my (long) answer was: "Hey lowell, interesting stuff, *I* can't help you much on this particular subject, in fact don't have much faith in it, but please *do* experiment and post the results for us to know, whether good or bad", which was resumed into: "Just do it and *you* tell us. Thanks."
Note: Fets (and bipolars) *can* switch voltages higher than those we are talking about, if they simply switch them to ground, but a series switcher such as M2 is a no no.
You want to minimize unknowns before building? Good!
But on *this* unknown, maybe there is not much people around which can help you; at most, refer you to some book.
Personal opinion?:I would breadboard and test it.
I mean the old original breadboards, a piece of wood, nails and some tinned wire, plus a couple tube sockets nailed for good; not push-in protoboards here.
OK friend, please breadboard only V1, V6b, V3a, (1 hour work), hook a generator or your guitar , load the free end of C37 with 10k to 100K to ground and hook there your scope, voltmeter, whatever.
You can also send that signal to any monitor amp hooked to a speaker, and proceed to switch those Fets on and off.
Try different gating voltages, straight connections (as drawn), back connected 1N4148's, 1M resistors , different signal levels, and you'll know much more than most of us.
Post your results.
Good luck with your project.
PS: next time I have nothing useful to say about something, I'll just , I promise.

[lowell]

Jm,
I have to admit what I got from your post was irritation at my "annoying" question. But I guess that's the issue these days w/ messaging, emails, forums etc... If you don't put a after every sentence you're an a##hole! But seriously, now I understand your post and thanks for explaining all that good info. You're right about the breadboard, I really need one of those around for this type of stuff. I've heard of the board w/ nails breadboard but never thought to actually do it. I will now. And I hope I can offer some helpful info on this stuff.... we'll see. Finally, you rocked w/ all the help on that Randall... thanks again. However I never did get a response from you to my last post on that one. Maybe I forgot to put a .

[J M Fahey]

Fine. I thought that Randall isue was finished, so I unsubscribed, otherwise my inbox bursts, but I'll check.