switching audio with MOSFETs

[Walter Harley]

I need to mute some audio. The signal is ground-referenced AC with peak
voltages around +/-2V at most, usually more like 250mV; the source impedance
is a few hundred ohms; and the load is unknown (but at least 10k).

Requirements: For marketing reasons (having to do with how the device is
installed), I don't want to put anything in series with the audio, just a
shunt. That's not an absolute requirement, though. Parts cost needs to be
kept as low as possible, while maintaining high audio quality. I need to do
this with extremely low supply current (<5uA). The only supply voltage I
have access to is a single positive voltage which could be anywhere between
7v and 18v.

My thought was to do the following. I think I need to use two MOSFETs
because if I just use one, the body diode will conduct on peaks even when
I'm not trying to mute. Can anyone see any flaws, or perhaps a better or
simpler way to do this? With values shown, muting is only 28dB, but that
may be adequate. I could also use VN3205's, with lower Rds(on), to get to
45dB or so. To do better I'd need to put something in series with the
signal.

Other solutions I considered: (1) a CD4066, but no good with signals below
Vss (=ground). (2) a PhotoMOS relay, but that's expensive and the LED takes
too much current and anyway it's just a fancy way of switching the MOSFETs.
(3) a JFET, but Rds(on) is too high.




Vin -----------o---------- Vout
(Rs~150R) |
|
|
.---||-+
| ||<-
| ||-.
| |
| | Both MOSFET's are VN0106
| | (Typ. Rds(on) is 3 ohms)
| ||-'
| ||<-
Vctrl --o---||-+
(Rs~1MEG) |
|
|
===
GND
created by Andy´s ASCII-Circuit v1.22.310103 Beta www.tech-chat.de

 

[Active8]

On 10 Dec 2003 18:28:47 GMT, [email protected] said...

I need to mute some audio. The signal is ground-referenced AC with peak
voltages around +/-2V at most, usually more like 250mV; the source impedance
is a few hundred ohms; and the load is unknown (but at least 10k).

Requirements: For marketing reasons (having to do with how the device is
installed), I don't want to put anything in series with the audio, just a
shunt. That's not an absolute requirement, though. Parts cost needs to be
kept as low as possible, while maintaining high audio quality. I need to do
this with extremely low supply current (<5uA). The only supply voltage I
have access to is a single positive voltage which could be anywhere between
7v and 18v.

My thought was to do the following. I think I need to use two MOSFETs
because if I just use one, the body diode will conduct on peaks even when
I'm not trying to mute. Can anyone see any flaws, or perhaps a better or
simpler way to do this? With values shown, muting is only 28dB, but that
may be adequate. I could also use VN3205's, with lower Rds(on), to get to
45dB or so. To do better I'd need to put something in series with the
signal.

Other solutions I considered: (1) a CD4066, but no good with signals below
Vss (=ground). (2) a PhotoMOS relay, but that's expensive and the LED takes
too much current and anyway it's just a fancy way of switching the MOSFETs.
(3) a JFET, but Rds(on) is too high.

I thought the 4066 was bidirectional like the deal you have below,
which incidentally may be wrong depending on what you're trying to
say here.

You have the drains connected together. The schematic symbol
convention for MOSFETs is that the gate lead is drawn off the side
that the source is on and it looks like you substituted a "." and
"'" in place of a "+" to indicate drain. I see what you're doing,
but there is a better way. First, I doctor the symbol thusly:

| D
||-+
||<-+
-||-+|
| S



but I leave the S & D out

you could:

o----------------+-----------------o
|
|
||-+
||<-+
|----||-+|
| |
| |
| / |
o----+--|<---+ <-- Zener or a PVI opto pile
| / |
| |
| |
| ||-+|
| ||<-+
+----||-+
|
|
|
|
===
GND

but you're supply/control voltage constraints suggest you might
want to try



o------+-------->|----+----|<----+
| | |
| | |
| | |
| | |
| ||-+ |
| ||<-+ |
| o------||-+| |
| | |
| o---------+ |
| | |
| | |
+--------|<----+---->|----+------o

o--------------------------------------o

which would give you a V+ on 0v off kinda control.

HTH,
Mike


created by Andy´s ASCII-Circuit v1.22.310103 Beta www.tech-chat.de

 

[Walter Harley]

You have the drains connected together. The schematic symbol
convention for MOSFETs is that the gate lead is drawn off the side
that the source is on and it looks like you substituted a "." and
"'" in place of a "+" to indicate drain.

Yeh, I was relying mostly on the gate lead to indicate D and S. I find the
+ on both drain and source confusing because to me it makes it look like the
drain is connected to the substrate. Mebbe I'll try your symbols; they're
hard to misinterpret. I have a hard time with MOSFET symbols :-(

Anyway, you're right that I put the drains together. I thought about
putting the sources together, but I couldn't quite prove to myself that when
the MOSFETs were off, applying a voltage on the gate was going to result in
any particular voltage between gate and source - I guess that gate leakage
is less than "off" channel leakage but I don't really know that for a fact.
With the drains together, each source has a relatively low impedance
connection to ground, so a voltage on the gates is ground-referenced. In
retrospect, I guess that given the near-symmetry of drain and source, it
doesn't make much difference either way. That still trips me up; I need to
break the habit of thinking of drain and source as analogous to collector
and emitter.

you could:

o----------------+-----------------o
|
|
||-+
||<-+
|----||-+|
| |
| |
| / |
o----+--|<---+ <-- Zener or a PVI opto pile
| / |
| |
| |
| ||-+|
| ||<-+
+----||-+
|
|
|
|
===
GND

What's a "PVI opto pile"?

But anyway, what's the purpose of this? It looks like it's trying to work
around the problem I mentioned above, but I don't understand how it works.

but your supply/control voltage constraints suggest you might
want to try



o------+-------->|----+----|<----+
| | |
| | |
| | |
| | |
| ||-+ |
| ||<-+ |
| o------||-+| |
| | |
| o---------+ |
| | |
| | |
+--------|<----+---->|----+------o

o--------------------------------------o

which would give you a V+ on 0v off kinda control.

Except, the two diode drops whack the signal, don't they? Anyway, if I can
avoid it, I'd like to avoid putting something in series with the signal
path. (Because it makes the installation of my widget harder.)

Thanks,
-walter

 

[DG]

Do you use some kind of special software or editor to do that ASCII art?
Or is it all by hand? Just curious.

Dave

 

[Active8]

Do you use some kind of special software or editor to do that ASCII art?
Or is it all by hand? Just curious.

We used to do it by hand.

I usually leave the link to the software in (it copies to the
clipboard just below the schem so it's there when you paste it) but
the OP did that (atta boy some people cut it out ) so I took it
out. It's the same software. Someone posted a link to another one
called PyASCII. It's written in Python but it's still in devel and
ain't worth the amount you don't pay.

BRs,
Mike

 

[Active8]

Yeh, I was relying mostly on the gate lead to indicate D and S. I find the
+ on both drain and source confusing because to me it makes it look like the
drain is connected to the substrate. Mebbe I'll try your symbols; they're
hard to misinterpret. I have a hard time with MOSFET symbols :-(

Anyway, you're right that I put the drains together. I thought about
putting the sources together, but I couldn't quite prove to myself that when
the MOSFETs were off, applying a voltage on the gate was going to result in
any particular voltage between gate and source

That's why I set up a voltage between the source and gate. It's a
bidirectional analog switch, so don't waste too much time proving
it. Google for some app notes, you'll likely find just what you
want. But you might as well finish reading my replies clear down to
the bottom since they're there. there there. here here. aye, aye,
eyeeye. ploink! ooh! my eye! hey moe!

- I guess that gate leakage
is less than "off" channel leakage but I don't really know that for a fact.
With the drains together, each source has a relatively low impedance
connection to ground, so a voltage on the gates is ground-referenced. In
retrospect, I guess that given the near-symmetry of drain and source, it
doesn't make much difference either way. That still trips me up; I need to
break the habit of thinking of drain and source as analogous to collector
and emitter.

crap. my settings cut the quoted text. copy paste...
crap again, no quotes.

See correction to schem above.

What's a "PVI opto pile"?

try pvi5013r.pdf and an-1017.pdf at www.irf.com

the "pile" is a photovoltaic pile or "cell" part of a PVI opto
isolator. Think of it as a photocell for a Vgs control. It'll drive
a logic level MOSFET on if you stick it in place of the Zener
above. Cool beans.

There's some analog switches out there that may work in the shunt
config. DG300, DG305, DG400. They're bidirectional, anyway.

But anyway, what's the purpose of this? It looks like it's trying to work
around the problem I mentioned above, but I don't understand how it works.

You didn't really mention a specific prob, but the prob I saw was
connecting the drains together and I can't assume a "low" impedance
path from the top MOSFETs source to ground as you mention above.
When the thing is on, a 680 ohm audio input impedance will drop
enough voltage in relation to the MOSFET Rds(on) to lower the Vgs,
no?

It works, like so (not that it'll work in this shunt config, now
that I look again. It'll clip negative peaks): with 0V on the
control line, and the shunted sig crossing ground, you get 0V Vgs.
With the signal above the 0V (off) control voltage (MOSFETs still
on), the zener is conducting and you can only get a diode drop
across the g-s. The shunt cuts off. With a high voltage (+Vp audio
sig max + Vzener) on the control line, you'll have the zener
voltage as Vgs and the MOSFETs switch on if you pick the right
Zener. Enough to give you Vgs(on) and it'll do double duty to as a
Vgs limiter - not that your input swings are gonna hurt in this
app.

That low (< 5uA) supply current constraint is absurd. The zener
current will slam you. You can't leave it out without a negative (
< -Vp ). Who boxed you in like that? I'f you're getting the 7 to
18V from one of the audio circuits, you're up to 3 wires from
existing equip to shunt anyway and you might as well just put the
switch inline AFAIC.

And 5uA ain't gonna drive the LED in a PVI opto, either.


I started playing with a buffer or something to get the source to
follow the negative swing so the single MOSFET wouldn't conduct and
stopped. Threw in a diode instead...


o---------------------+--------------------------o
|
|
||-+
||<-+
Vc o--+----||-+|
| |
| | D
| +----->|---+
| | |
+----||-+| |
||<-+ |
||-+ |
| |
| |
| |
=== ===
GND GND
created by Andy´s ASCII-Circuit v1.22.310103 Beta www.tech-chat.de

Now the top MOSFET can't conduct through diode D on the negative
swing and you've got a reference for Vgs. That oughtta woik, eh?

You need to find some extra mA's. Those PVI optos (see below)
really simplify things.

but your supply/control voltage constraints suggest you might
want to try



o------+-------->|----+----|<----+
| | |
| | |
| | |
| | |
| ||-+ |
| ||<-+ |
| o------||-+| |
| | |
| o---------+ |
| | |
| | |
+--------|<----+---->|----+------o

o--------------------------------------o

which would give you a V+ on 0v off kinda control.

Except, the two diode drops whack the signal, don't they?

Yeah, It would. Solly. I'm switching big sigs in my head. File it
for later.

Anyway, if I can
avoid it, I'd like to avoid putting something in series with the signal
path. (Because it makes the installation of my widget harder.)

I'm picturing an inline splice block with one tap wire as opposed
to a box in the line. If it's the box in the line you're worried
about, ok, but analog switches are used in the audio path all the
time, so I wouldn't worry about infidelity ( nyuk, nyuk, nyuk )
The 4066 IIRC was used for audio switching, but it would need to be
DC biased up and isolated with blocking caps. Keep the sig above
0V.

Mike

Thanks,
-walter

 

[Walter Harley]

I tried it. It makes no difference whether the sources or drains are tied
together; to a first approximation, the FETs are symmetric, they don't know
the difference between S and D. Duh.

It works, like so (not that it'll work in this shunt config, now
that I look again. It'll clip negative peaks):

Yep, that's what I found. I hadn't been thinking very clearly about the
"off" (unmuted) state. With the gates grounded, if the signal has a
negative peak below -Vth, the FETs turn on. So I clip negative peaks,
starting around 2v. This is a bit too close to my normal (max) signal
levels for me to feel comfortable, so it's back to the drawing board...

That low (< 5uA) supply current constraint is absurd. The zener
current will slam you. You can't leave it out without a negative (
< -Vp ). Who boxed you in like that?

The device wants to run on a 9v battery without changing its shelf life
substantially.

I started playing with a buffer or something to get the source to
follow the negative swing so the single MOSFET wouldn't conduct and
stopped. Threw in a diode instead...

I think it ends up with the same problem. when the signal goes below -Vth,
the top MOSFET (which can't tell the difference between its S and its D)
turns on; that exposes the signal to the bottom MOSFET, same thing happens;
and I clip. Really, I just need to find a way to get the Vc to be negative
rather than ground. Maybe a tiny little charge pump. Or maybe just a
rectified power supply driven by the signal itself, although there are some
problems with that (which I won't get into here since I'm intentionally
being a little bit vague).

Thanks,
-walter

 

[Active8]

On 12 Dec 2003 19:47:38 GMT, [email protected] said...

I tried it. It makes no difference whether the sources or drains are tied
together; to a first approximation, the FETs are symmetric, they don't know
the difference between S and D. Duh.



Yep, that's what I found. I hadn't been thinking very clearly about the
"off" (unmuted) state. With the gates grounded, if the signal has a
negative peak below -Vth, the FETs turn on. So I clip negative peaks,
starting around 2v. This is a bit too close to my normal (max) signal
levels for me to feel comfortable, so it's back to the drawing board...



The device wants to run on a 9v battery without changing its shelf life
substantially.



I think it ends up with the same problem. when the signal goes below -Vth,
the top MOSFET (which can't tell the difference between its S and its D)
turns on;

I'm thinking it doesn't have a current path or source reference
potential because of the diode. You can't say the MOSFET is on if
it isn't passing current. The things are off so the top guy has no
reference and there's no current or potential at the diode's anode.
You'd have to get current flowing through the diode (which won't
happen at -Vp) before the bottom guy can see -Vp. Or you'd have to
have the bottom guy switched on and that's isn't going to happen
with 0V on it's gate becuse it has a solid source reference.

I wish you'd just lash it up and try it so I don't have to ;-) I
just got some MOSFETs in the mail but this is you're baby. Please
tell me this isn't some ill conceived mute button to shut down the
audio so you can hear the phone ring. Any idiot can get a flashing
strobe at radio shack that serves the same purpose, not that it
would help the blind.

that exposes the signal to the bottom MOSFET, same thing happens;
and I clip. Really, I just need to find a way to get the Vc to be negative
rather than ground. Maybe a tiny little charge pump. Or maybe just a
rectified power supply driven by the signal itself, although there are some
problems with that (which I won't get into here since I'm intentionally
being a little bit vague).

Thanks,
-walter

Everything I've seen reguires a Vss more negative than -Vp and
that's for series switches. This shunt config is something I
haven't even considered prior to this.

Since you finally coughed up enough info to mention a 9V batt, (7-
18V you mentioned before) those are good for about 325mAH so you've
got plenty of juice to turn on a PVI opto for say, 16 hours,
roughly. The "shelf life" will depend on duty cycle. Will it be on
most of the time or off? "Shelf-life" implies off for a substantial
amount of time and a PVI opto won't be drawing LED current during
that time. That 7-18V statement would imply 2 9V batts and there's
your negative voltage 9-0-(-9).

Mike

 

[Walter Harley]

I'm thinking it doesn't have a current path or source reference
potential because of the diode. You can't say the MOSFET is on if
it isn't passing current. The things are off so the top guy has no
reference and there's no current or potential at the diode's anode.
You'd have to get current flowing through the diode (which won't
happen at -Vp) before the bottom guy can see -Vp. Or you'd have to
have the bottom guy switched on and that's isn't going to happen
with 0V on it's gate becuse it has a solid source reference.

Hi, Mike.

There's a voltage between the gate and the top FET's drain, because they
share a common ground. When the signal at the drain is less than -Vth, then
as far as the FET is concerned, its gate is forward biased compared to one
end of its channel (it can't tell which end is which, for practical
purposes), and the channel starts conducting. When that happens, the source
of the bottom FET is now exposed to -Vth, meaning its gate is forward
biased, and the same thing happens.

The diode doesn't have anything to do with that picture, because it's
reverse biased if it's biased at all. But even in principle, there's
nothing it could do to help: the problem is between the gate and drain of
the top FET.

I wish you'd just lash it up and try it so I don't have to ;-)

I did try it without the diode, and it behaves that way. If you'd like I'll
be glad to put the diode in

Please tell me this isn't some ill conceived mute button to shut
down the audio so you can hear the phone ring.

No, no. Fear not. It's a little bit more noble than that

I'm probably being absurd by trying to protect the details of the device I'm
working on, since it's really more just a widget than an invention, and it
feels pretty "obvious" to me. But I've learned that my judgement of what's
patentable (or of what's worth patenting, which is different) is lousy, so I
need to do a little more research before I disclose. Sorry!

Since you finally coughed up enough info to mention a 9V batt, (7-
18V you mentioned before) those are good for about 325mAH so you've
got plenty of juice to turn on a PVI opto for say, 16 hours,
roughly. The "shelf life" will depend on duty cycle. Will it be on
most of the time or off?

I don't have access to the midpoint of the batteries - the supply is just
like I said, from my perspective a single voltage that could be anywhere
from 7 to 18v. The thing I'm working on is an in-the-field addition to an
existing piece of equipment, thus all the silly constraints.

With some cleverness elsewhere, I might be able to mute for just a few
seconds at a time, with weeks in between. If so, then you're right, the PVI
opto would work just fine. Might be the right direction to look.
(Although, I think they're a bit more expensive than a pair of MOSFETs.)

-w

 

[Active8]

On 13 Dec 2003 01:37:04 GMT, [email protected] said...

I did try it without the diode, and it behaves that way. If you'd like I'll
be glad to put the diode in

You didn't ground the sources, did you?

If not, I guess it won't matter if you used 2 MOSFETs. The diode
idea came to me because I know a single MOSFET would conduct and
its source has a path to ground (unidirectional switch). Now if you
connected 2 MOSFETs up without the diode and it clipped then so
much for that idea, but it would've never turned on, either since
there's no source reference.

No, no. Fear not. It's a little bit more noble than that

I'm probably being absurd by trying to protect the details of the device I'm
working on, since it's really more just a widget than an invention, and it
feels pretty "obvious" to me. But I've learned that my judgement of what's
patentable (or of what's worth patenting, which is different) is lousy, so I
need to do a little more research before I disclose. Sorry!

Try www.tinaja.com and if you can find the link on when2pat and
whatever else he's written on patents, you may just save yourself a
lot of hassle. I'm sure if the patent office issued a patent on a
mute switch from 2 MOSFETs, it won't be enforceable and if you use
any ideas you get here, solly. Public domain, obvious solution,
prior art...

If you have to cut into the line to tap it with the shunt, why not
just add 2 chip caps and put a positive dc bias on the line right
at the tap?

Mike

 

[GPG]

All methods thus far involve a dc shift on the signal.
Try a J-fet biased off by neg on gate, bias 0V for shunt.

 

[Walter Harley]

All methods thus far involve a dc shift on the signal.
Try a J-fet biased off by neg on gate, bias 0V for shunt.

I don't have access to a negative voltage, unless I create one myself with a
charge pump or whatever. Trying to avoid the extra cost of that.

 

[Walter Harley]

You didn't ground the sources, did you?

If not, I guess it won't matter if you used 2 MOSFETs. The diode
idea came to me because I know a single MOSFET would conduct and
its source has a path to ground (unidirectional switch). Now if you
connected 2 MOSFETs up without the diode and it clipped then so
much for that idea, but it would've never turned on, either since
there's no source reference.

But there's still a *drain* reference. The MOSFET can't tell the difference
between its source and its drain; if there's + bias from gate to *either* of
those pins, it'll turn on.

Try www.tinaja.com and if you can find the link on when2pat and
whatever else he's written on patents, you may just save yourself a
lot of hassle. I'm sure if the patent office issued a patent on a
mute switch from 2 MOSFETs, it won't be enforceable and if you use
any ideas you get here, solly. Public domain, obvious solution,
prior art...

Thanks, I'll do that.

The MOSFET muting is not the (conceivably) patentable part of the device.
(Although I'm starting to think that if I could figure a way to mute audio
with low current and no negative supply it just might be!) It's just a part
I need to get right for the whole thing to work.

If you have to cut into the line to tap it with the shunt, why not
just add 2 chip caps and put a positive dc bias on the line right
at the tap?

With a shunt, I can just solder an extra wire on at one end. No need to cut
anything. But cutting and caps may be the only answer...

 

[Active8]

On 13 Dec 2003 05:32:59 -0800, [email protected] said...

All methods thus far involve a dc shift on the signal.
Try a J-fet biased off by neg on gate, bias 0V for shunt.

OP said:

3) a JFET, but Rds(on) is too high.

 

[Active8]

On 13 Dec 2003 18:08:21 GMT, [email protected] said...

But there's still a *drain* reference. The MOSFET can't tell the difference
between its source and its drain; if there's + bias from gate to *either* of
those pins, it'll turn on.

you seem to be fixating on that problem.

i was talking about the thing not being able to turn on WHEN YOU
WANT IT TO because there's no source ref. have you bothered to
search for bidirectional analog switches? you need a potential
between the sources and gates to switch AC.

i said:

[...] but it would've never turned on, either since
there's no source reference.

Try www.tinaja.com and if you can find the link on when2pat and
whatever else he's written on patents, you may just save yourself a
lot of hassle. I'm sure if the patent office issued a patent on a
mute switch from 2 MOSFETs, it won't be enforceable and if you use
any ideas you get here, solly. Public domain, obvious solution,
prior art...

Thanks, I'll do that.

He knows from what he's talking.

The MOSFET muting is not the (conceivably) patentable part of the device.
(Although I'm starting to think that if I could figure a way to mute audio
with low current and no negative supply it just might be!

someone would've done it by now. That's why I've never seen it
done, I suppose.

Oh, wait. I have. It's (again) called a bidirectional analog switch
and it goes in series with the line. It normally has a negative
supply, but maybe the aformentioned caps and bias could deal with
that. but in this case, there's a buffer to make the Vgs(on) ( and
thus the Rds(on) ) contant to prevent distortion. A PVI would solve
everything but you've ruled that out with your current budget spec.
I told you there's audio switches in existing equipment. My boom
box has 'em, but those D cells have more maH than a 9V.

) It's just a part
I need to get right for the whole thing to work.



With a shunt, I can just solder an extra wire on at one end. No need to cut
anything. But cutting and caps may be the only answer...

and here I was thinking you wanted a simple block splice (and tap)
in the line so the end user didn't have to mess with an extra
chassis. hell, if you're inside the thing already WTF is a little
mod? Even with a little block splice with RCA connectors between a
tuner and amp... a couple chip caps are trivial.Mike

 

[Walter Harley]

The MOSFET muting is not the (conceivably) patentable part of the device.
(Although I'm starting to think that if I could figure a way to mute audio
with low current and no negative supply it just might be!

someone would've done it by now. That's why I've never seen it
done, I suppose.

Oh, wait. I have. It's (again) called a bidirectional analog switch
[...]

Oh, this is getting ridiculous!

I think I'm convinced that patenting is a bad idea. Judging from Don's
page, hardware patents are a different world than software patents, which I
know more about. The software patent world has gotten bizarre enough that
my judgement of what prior art is, and what is or is not obvious, is quite
different than that of the PTO - they grant, and other companies honor,
patents that are just absurd. But it sounds like my intuitions about
hardware are closer to correct, which is refreshing, but means that my idea
isn't worth patenting. So, I'll disclose it and we can talk more lucidly.

But I'm going to do it in a different thread, because this one's starting to
take a lot of scrolling in my newsreader to maintain

I'll title the new thread something like this one.

 

[Walter Harley]

But I'm going to do it in a different thread, because this one's starting to
take a lot of scrolling in my newsreader to maintain

Changed my mind - I think there's not really a lot left to discuss. So I'll
just post the context of the problem here, for the sake of clarity.

The device I'm working on is an aftermarket power controller for electric
bass preamps. Electric basses (and some other instruments) contain preamps
and EQ circuits, powered by one or two 9v batteries. These preamps are
switched "on" by plugging in the instrument cable: the cable has a mono 1/4"
phone plug, and the jack is a stereo (TRS) jack with the negative side of
the battery connected to the ring terminal. So, when the cable is plugged
in, the battery negative is shorted to ground, and the preamp works. If
there are two batteries, they are invariably in series, that is, a single
18v supply, not a split supply.

In normal operation, the preamps draw very little current; typical battery
life is several hundred hours. Since people generally play the bass at most
an hour a day, and many people just noodle around once every week or two,
this is long enough that sometimes people buy, use, and sell a bass without
ever realizing that it had a battery. No problem.

However, sometimes someone accidentally leaves the cable plugged in. A
hundred hours isn't very long, in that case; you come back to the bass at
the next practice session and the batteries are drained. Many folks have
more than one bass, some passive, some active, and it's confusing to
remember which ones you need to unplug and which you can leave plugged in.
We're bass players, not keyboardists ;-)

I'm working on a device that can be installed into an existing bass, ideally
by someone with no electronics skill and minimal experience with a soldering
gun. (This means that if they don't have to cut and reconnect any existing
wires, it's better.) The device will sit there monitoring signal and turn
the instrument's power OFF when there's no signal for a few minutes.
Obviously, the device itself has to draw a lot less power than the
electronics it's controlling; and its environment is rather restrictive and
has some unknowns. That's the context for all of this.

(There are other possible solutions, such as phantom power from an external
source. I feel confident enough in my understanding of this particular
small market segment that I think this particular solution is the right one
to pursue.)

I've got the power controller part working fine. Detect signal from
pickups, turn power on and off with appropriate delay. The problem is that
there's a thump as the preamp turns on and off, even if I do the transition
very gradually. So, I need to mute the output, either during turn-on and
turn-off, or all the time it's off.

So, that's what this is all about! I think the best idea thus far is to use
a photoMOS coupler as a shunt, and restrict its use to the brief period of
power transition.

-w

 

[Active8]

On 13 Dec 2003 23:38:40 GMT, [email protected] said...

Changed my mind - I think there's not really a lot left to discuss. So I'll
just post the context of the problem here, for the sake of clarity.

The device I'm working on is an aftermarket power controller for electric
bass preamps. Electric basses (and some other instruments) contain preamps
and EQ circuits, powered by one or two 9v batteries. These preamps are
switched "on" by plugging in the instrument cable: the cable has a mono 1/4"
phone plug, and the jack is a stereo (TRS) jack with the negative side of
the battery connected to the ring terminal. So, when the cable is plugged
in, the battery negative is shorted to ground, and the preamp works. If
there are two batteries, they are invariably in series, that is, a single
18v supply, not a split supply.

In normal operation, the preamps draw very little current; typical battery
life is several hundred hours. Since people generally play the bass at most
an hour a day, and many people just noodle around once every week or two,
this is long enough that sometimes people buy, use, and sell a bass without
ever realizing that it had a battery. No problem.

However, sometimes someone accidentally leaves the cable plugged in. A
hundred hours isn't very long, in that case; you come back to the bass at
the next practice session and the batteries are drained. Many folks have
more than one bass, some passive, some active, and it's confusing to
remember which ones you need to unplug and which you can leave plugged in.
We're bass players, not keyboardists ;-)

I'm working on a device that can be installed into an existing bass, ideally
by someone with no electronics skill and minimal experience with a soldering
gun. (This means that if they don't have to cut and reconnect any existing
wires, it's better.) The device will sit there monitoring signal and turn
the instrument's power OFF when there's no signal for a few minutes.
Obviously, the device itself has to draw a lot less power than the
electronics it's controlling; and its environment is rather restrictive and
has some unknowns. That's the context for all of this.

Er, I'm not trying to be the prick that I am, but how do you get
the power controller in the supply line without cutting into any
existing wires?

(There are other possible solutions, such as phantom power from an external
source. I feel confident enough in my understanding of this particular
small market segment that I think this particular solution is the right one
to pursue.)

I've got the power controller part working fine. Detect signal from
pickups, turn power on and off with appropriate delay. The problem is that
there's a thump as the preamp turns on and off, even if I do the transition
very gradually. So, I need to mute the output, either during turn-on and
turn-off, or all the time it's off.

Vedy interestink! Wouldn't you expect it to thump when it's plugged
in the normal, unmodified way, all else being equal? power on is,
after all, power on and the tip should be making contact at that
point. Maybe not quite. (run and find plug and jack) When I plugged
up hot with the amp on and the guitar volume up, it didn't sound
real purty. I can't hear your "thump", either.

Ok. .25 plug, stereo jack. T&R both make contact with the plug tip
before the R slips over the insulator and contacts the plug S. So
the signal path is established before the power is applied.

You have another problem.

Does it do this if you just disconnect and connect the battery
while the cable's plugged in? Try it with the bass volume up *and*
down.

So, that's what this is all about! I think the best idea thus far is to use
a photoMOS coupler as a shunt, and restrict its use to the brief period of
power transition.

Or fix the problem.

Mike

 

[Walter Harley]

Er, I'm not trying to be the prick that I am, but how do you get
the power controller in the supply line without cutting into any
existing wires?

This is much more fun, being able to actually talk about it! The answer to
that particular question is: I take over the battery connector. So, I sit
between the battery and the rest of the electronics.

Vedy interestink! Wouldn't you expect it to thump when it's plugged
in the normal, unmodified way, all else being equal? power on is,
after all, power on and the tip should be making contact at that
point. Maybe not quite. (run and find plug and jack) When I plugged
up hot with the amp on and the guitar volume up, it didn't sound
real purty. I can't hear your "thump", either.

Does your guitar have active electronics? Interestingly, active electric
guitars are much less common than active electric basses. (I say
"interesting" because the guitar would benefit more from the additional
tonal possibilities, I think.)

The active ones do pop when you plug 'em in, but it's not usually an issue
because people aren't plugging and unplugging during a show. But this
device would cause the power to turn on and off at unexpected and
inopportune times, like when you're plugged into a few kW of PA system.

Ok. .25 plug, stereo jack. T&R both make contact with the plug tip
before the R slips over the insulator and contacts the plug S. So
the signal path is established before the power is applied.

You have another problem.

Does it do this if you just disconnect and connect the battery
while the cable's plugged in? Try it with the bass volume up *and*
down.

When I plug a mono plug into a stereo jack (with voltage between R and S),
first the plug T touches the jack S. Then plug T touches jack R, and plug S
touches jack S (pop!). Then plug T touches jack T, plug S touches both jack
R and jack S (electronics on).

But anyway, the scenario of interest is that the bass is sitting there,
plugged in to an amp that's turned on, and either the player starts playing
(pop!) or the timeout expires and the bass turns off (pop!). They expect
there to be noise when plugging and unplugging, but not when the bass is
just sitting there.

If I disconnect and connect the battery (with a debounced MOSFET switch)
while the cable's plugged in, I get a pop. Less of a pop if I allow the
voltage to ramp gradually up and down over 0.25s, but still present.
(Actually more of a 'thump' if the voltage ramps rather than being turned on
and off abruptly.)

-w

 

[Active8]

On 14 Dec 2003 07:57:01 GMT, [email protected] said...

This is much more fun, being able to actually talk about it! The answer to
that particular question is: I take over the battery connector. So, I sit
between the battery and the rest of the electronics.

Sounds dark and lonely

Does your guitar have active electronics? Interestingly, active electric
guitars are much less common than active electric basses. (I say
"interesting" because the guitar would benefit more from the additional
tonal possibilities, I think.)

No, it didn't though I had thought of adding an active tone control
at one time. I was just trying to get an idea of what kind of
"thump" you're getting. Even my stereo thumps under certain
conditions.

The active ones do pop when you plug 'em in, but it's not usually an issue
because people aren't plugging and unplugging during a show. But this
device would cause the power to turn on and off at unexpected and
inopportune times, like when you're plugged into a few kW of PA system.



When I plug a mono plug into a stereo jack (with voltage between R and S),
first the plug T touches the jack S. Then plug T touches jack R, and plug S
touches jack S (pop!).
pop?

Then plug T touches jack T, plug S touches both jack
R and jack S (electronics on).

With the plug and jack I just examined, the plug T eventually
contacts both the jack T&R before the plug is all the way in. Then
the jack R finally slides past the PT to contact PS.

So maybe (just maybe) the signal and any charge on the outpout cap
is somehow discharged/shorted within the preamp and maybe not.

But anyway, the scenario of interest is that the bass is sitting there,
plugged in to an amp that's turned on, and either the player starts playing
(pop!) or the timeout expires and the bass turns off (pop!). They expect
there to be noise when plugging and unplugging, but not when the bass is
just sitting there.

Is it the same "pop"/"thump" as above? Creating a problem?
Perceived problem?

If I disconnect and connect the battery (with a debounced MOSFET switch)

I meant to do it with your circuit out of the picture. Just plug
the bass into the amp and try connecting and disconnecting the batt
at different bass volumn levels. You might find it makes the same
noise and if it's that bad and you want the mute thingy, so be it.
You know the other people in the bar aren't going to hear it while
you're on break because the juke box or other music should be
playing.

while the cable's plugged in, I get a pop. Less of a pop if I allow the
voltage to ramp gradually up and down over 0.25s, but still present.
(Actually more of a 'thump' if the voltage ramps rather than being turned on
and off abruptly.)

If there's no transient coming from/caused by your power
controller, then this should also happen when you do the above
test. I have to ask this shit because if I had a dime for everyone
who claimed to know electronics, I'd be rich. Design and
troubleshooting are 2 different things, also.

I'd like to see a schem of the active bass circuitry to get an idea
of why it "thumps" if this "thump" is *that* friggin' dramatic.

-mike