[sdiy] JFET question - Sound leak in Mesa Road King II amplifier

Barry Klein barryklein at cox.net
Thu Sep 7 09:10:55 CEST 2017


I don't think jfets have a body diode, but there is a max. gate current spec 
of 50mA so a series resistor with each jfet gate might help out.  Could be 
an unexpected surge current on the drain too given so many things are being 
switched...
There are these cool component testers like for 20 bucks on ebay that will 
test these parts quickly out of circuit.  I bought the American first 
version, called the m3, and it has saved me a heck of a lot of time on 
things like this.

Barry


-----Original Message----- 
From: cheater00 cheater00
Sent: Wednesday, September 06, 2017 6:15 PM
To: synth-diy mailing list
Subject: [sdiy] JFET question - Sound leak in Mesa Road King II amplifier

Hi guys,
I've been trying to figure this one out for a year or so, but it's been 
elusive.

I have a Mesa Road King II tube amp. When I turn the gain all the way
down, there is still sound coming out. It's a slightly high-passed
version of the input. I even put a cable into the effect loop return,
with the other end of the cable disconnected - and sound was still
coming through. The cable should have made sure no audio comes through
from the peramp and input. Yet I still get room level output when the
head is connected to a Mesa oversized 4x12.

At first I thought it might have been the tubes coupling via their
voltage rails, but I haven't really tested that idea, because
decoupling each tube would be expensive (in part cost) and also
technically complex and a huge butcher job on the PCB.

I thought maybe the capacitors have dried out, but they seem ok-ish,
maybe a wee bit swollen, but they just might have been manufactured
that way.

I stumbled upon this video today, with a guy debugging an issue in a
Roadster, which is a scaled down Road King:
https://www.youtube.com/watch?v=4kNMtOmT76U

here he works on the JFETs used for muting the sound. Turns out he
shorted some pads on one of them and that made the whole amp silent.

Here are the relevant sections of the RKII schematics: 
https://imgur.com/a/astMS

You will see some J175's connecting from the signal chain to ground,
with the gate seemingly disconnected. All those gates are tied
together and connected to the pop reduction circuit ("mute circuit",
seen in the third diagram on the bottom right). The RKII uses a lot of
relays, and when they switch, normally there would be a loud pop in
the output - so during that time, the J175's momentarily short the
signal to ground, to make sure the pop doesn't propagate (and possibly
sustain due to built in spring reverb or outboard). The J175's are
controlled by the 6426 in the lower right corner of the third diagram,
but I'm not sure how that works exactly.

If you look at the comments of the video, you will see user mrkv4k
comments on the way the J175s are used. He says they are bound to be
damaged because the internal (parasitic) body diode gets biased too
much in this design. As a consequence, those JFETs die often and often
need to be replaced. And it is true, they often get replaced by the
techs. I didn't fully understand his explanation of using N-channel
transistors instead of P-channel. Could someone explain this to me?
Here are the comments:

"""
mrkv4k:
> This is the weirdest design with JFETs I have ever seen. IMHO they are 
> literary boud to malfunction at some point, because the method they 
> choosed to switch them On is just wrong!
> You should never use low impedance to ground to switch p-chanel JFETs, 
> because this way, they are gonna be damaged every time it's drain gets 
> above 1V.
> If you wanna ensure that they won't be damaged, and you don't wanna 
> rebuild the mute circuit with PNP instead of 6426, add one 4k7 resistor 
> between R128 and colector of NPN 6426 (on the transistor side).

The Guitologist:
> Maybe there's a job at Mesa in your future.

mrkv4k:
> I doubt that :D I am thousands of miles away and I design HW for acces 
> control systems (mostly RFID time-clocks). Building Tube amps is just a 
> hobby.
> I think that the problem is that more and more engineers design with some 
> simulation programs and they don't test extreme conditions on the real 
> thing. And some engineers are not aware of the fail-modes of basic 
> components, like with those JFETs, you learn how it works under standard 
> conditions (applying positive voltage on gate of p-JFET is going to 
> "shrink" the conductive channel), but ommit the non-standard conditions 
> (if you look at the schematic of JFET, there is an arrow on the gate, 
> which is actual "parasitic"/body diode that shouldn't be ever polarized in 
> conducting direction - and that is what can happen in this design).

The Guitologist:
> You would know better than I. I appreciate you sharing your expertise 
> here. Something to chew on for sure. So your suggested mod for this Mesa 
> Boogie Mute Circuit would be to add one 4k7 resistor between R128 and 
> colector of NPN 6426 (on the transistor side)? And this should lengthen 
> the life of the JFETs?

mrkv4k:
> Yes, if the problem is what I think it is, than it will help protect those 
> JFETs. Also, I checked  the numbers once more and I think 470R would be 
> better, since that is sufficient to supress that effect and it won't slow 
> down shutdown of  RM2. So I would check this solution, to see if the mute 
> circuit works fine. Other way would be to put those resistors to each 
> gate.
> What I think happends is that when you try to mute while there is high 
> signal (like when you switch on effect while playing), gates of JFETs  are 
> grounded via  low resistance of that NPN transitor. Since drain is at high 
> voltage, you will get spike of current going throught the body diode (just 
> until it's muted). Those body diodes are small and can handle only pulse 
> up to 50mA.
"""


My question is: could those JFETs - even if they are in OK-ish
condition (remember that's a really old amp with old, possibly crappy,
silicon in them), could they be leaking sound from the input (see
first schematic diagram, just next to guitar input jack), through the
gate, to the common gate "rail", and then to the gate of the J175MU on
the very right of the second schematic diagram and subsequently
through that chip to the power amp (V6B grid, part of the power amp's
input pair)? This is all in the JFETs' "off" state, where they are
*not* shorting the signal path to ground.

If I were to replace those JFETs, how could I limit this behavior?

Here are a few ideas - please let me know if they are sound:
- add 470 to 4k7 resistors in series with each JFET's gate. This
resistor would be placed between each gate and the bus that is common
to all gates.
- in addition to that resistor, add a capacitor from the junction of
the resistor and the JFET's gate to ground to make sure no sound makes
it into the gate rail. The gate itself is mostly a DC signal so that
should be blocked by the capacitor, consequently arriving at the gate
pin of the JFET, with a rounded leading edge. (please correct me if
I'm wrong).
- use a different transistor, maybe N-channel JFET instead of
P-channel, but I don't know how I would do that

I don't really know what characteristics of a JFET I would be looking
for. I had a stroll through the J175's data sheet (eg this one:
http://www.jameco.com/Jameco/Products/ProdDS/992409.pdf ) but I'm not
sure what number I would be looking to improve upon over what the J175
provides. Does anyone have tips?

Another user commented on possible FET replacements:

"""
Evil Gremlin:
> Seen same JFET issue. Replaced with more powerful MOSFETs and adjusted 
> gate voltage, so little static wouldn't be an issue in future. It's just 
> pull-downs, so basically you can put anything here. JFET, MOSFET, bipolar, 
> IGBT - whatever perverted fantasies you have :) It's just p-channels don't 
> require any other components.

The Guitologist:
> Did you have to adjust the values of the timing capacitors in the mute 
> circuit also? If you're changing voltage, it seems you might need to in 
> order for the mute not to last too long or too short.

Evil Gremlin:
> Nope, got lucky. Well, in theory mute time became slightly longer, because 
> of higher gate capacitance, but not noticeably :)
"""

Not sure how to select a transistor like this. How would I make sure
the transistor is a drop-in replacement and doesn't need me to change
the timing capacitors? Would I look for the same gate capacitance as
J175? Some other thing? What did The Guitologist mean with changing
voltage?

Thanks a lot. Any tips would be highly appreciated!

_______________________________________________
Synth-diy mailing list
Synth-diy at synth-diy.org
http://synth-diy.org/mailman/listinfo/synth-diy 




More information about the Synth-diy mailing list