switching audio with MOSFETs

[Walter Harley]

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.

Interesting. I went down and grabbed a standard ol' open-frame Switchcraft
stereo jack and a Switchcraft mono plug, and confirmed with a continuity
tester that the plug T definitely leaves the jack S before it touches the
jack R; that is, the first thing that touches the jack R is the plug T, and
when they touch, the plug T is not touching anything else. At the moment
this contact occurs, plug S is contacting jack S.

With no plug, the circuit looks like this:
(jack S) ----- (electronics) ----- (+ bat -) ------- (jack R)

So, when the plug T touches jack R, and plug S touches jack S, it completes
a circuit and roughly 9v is placed across whatever the plug is connected to
(i.e., the bass amp). As soon as the plug is inserted the rest of the way,
this is no longer true.

Perhaps this varies with different jack and plug styles. The Switchcrafts
are pretty common, though (although the closed-frame ones are slightly more
common in basses than the open-frame ones, IME).

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

Let me be clear: this device is not in beta, it's just on my workbench. So
far the only bass player that has experienced this is me. But having built
and repaired a lot of equipment for bass players, I'm pretty confident that
if the bass was just sitting there, a few minutes after a song (e.g., during
a dead spot in a theatrical performance) and suddenly it made a loud "pop",
that would be perceived as a significant problem.

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.

Yes, same noise.

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.

In a noisy bar, no problem. In a Broadway theater pit, big problem. In one
of my roles, I work sound for a dinner cabaret production
(http://www.teatrozinzanni.com). There are moments where someone's on a
tight rope or trapeze, and the audience is dead silent. If the bass
suddenly went "pop" there'd at least be a very pissed-off stage manager, and
the distraction could even cause a performer to be injured.

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.

Well, a design with no troubleshooting ain't worth a lot

Yes, the conclusion I came to was that the transient was out of my control;
I could not rely on the electronics to be silent during a power-on or
power-off event. This is not surprising, since they're virtually all
single-supply devices biased at 1/2Vcc with DC blocking caps on the output.
I had hoped that by gradually ramping the supply up or down I could evade,
but what I found was that I have to go so slowly that there would be a
noticeable turn-on lag.

I take no offense from the questions; I'd be asking the same ones if I were
trying to help someone, and I appreciate your help. For the sake of detail,
though: I do have an EE degree, and I've been doing analog electronics as a
hobby for about 25 years. I'm not a great engineer, but I do claim to have
a reasonable understanding of at least some of the rudiments. (However, I
never really worked with MOSFETs until recently.)

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.

You can see one typical example at
http://www.cafewalter.com/cafewalter/fetpre/htfld.gif. That's not the one
that I happen to be testing with at the moment; but the differences aren't
huge. Practically all the active pres I've looked at are based on TL062's,
with an input buffer followed by an active EQ.

The thump is not *that* friggin' dramatic; it's about the same level as the
music. But it needs to be practically inaudible, I think, for this device
to be marketable.

 

[Active8]

On 15 Dec 2003 05:56:21 GMT, [email protected] said...

Interesting. I went down and grabbed a standard ol' open-frame Switchcraft
stereo jack and a Switchcraft mono plug, and confirmed with a continuity
tester that the plug T definitely leaves the jack S before it touches the
jack R; that is, the first thing that touches the jack R is the plug T, and
when they touch, the plug T is not touching anything else. At the moment
this contact occurs, plug S is contacting jack S.

With no plug, the circuit looks like this:
(jack S) ----- (electronics) ----- (+ bat -) ------- (jack R)

So, when the plug T touches jack R, and plug S touches jack S, it completes
a circuit and roughly 9v is placed across whatever the plug is connected to
(i.e., the bass amp). As soon as the plug is inserted the rest of the way,
this is no longer true.

Perhaps this varies with different jack and plug styles. The Switchcrafts
are pretty common, though (although the closed-frame ones are slightly more
common in basses than the open-frame ones, IME).

I figure they're the same exactly as far as make/break and all.
Remember, I found that the PT touches JT&R before it's in all the
way so there's another circuit completed at that point.

Let me be clear: this device is not in beta, it's just on my workbench. So
far the only bass player that has experienced this is me. But having built
and repaired a lot of equipment for bass players, I'm pretty confident that
if the bass was just sitting there, a few minutes after a song (e.g., during
a dead spot in a theatrical performance) and suddenly it made a loud "pop",
that would be perceived as a significant problem.



Yes, same noise.

Ok. That's what I figured. Gotta check. That eliminates *your*
circuit as the problem. And turning the bass volume down doesn't
help? I wouldn't leave my axe on stage with the volume up. Any loud
noise would get the strings vibrating. Even setting it down makes a
noise.

In a noisy bar, no problem. In a Broadway theater pit, big problem.

I can see being lazy at home, but a pro might have enough
discipline to take active measures to prevent this if you put a big
bold warning on the instrucion book. Bold red print, lightning
bolts, flaming cherubs of the (who was the destroyer? Abbadon?)
type.

OTOH, no thump *would* be nice. It's sometimes hard to strike a
balance between functionality and preventing the eventuality of
stoopid user tricks. With my software, I started getting tired of
shit like validating input to the umpteenth degree. What are the
chances that someone's going to enter a letter where it's obvious
that a number needs to be entered? The embedded project I'm
doing... no critical app or safety hazards involved, so if the EU
wants to enter an invalid parameter, too bad. There's not enough
memory to add that much error checking or more than a simple
"Error" and "Memory write failed." string. I just ignore invlid
keystrokes.

But...


Hell I just sent off a slew of questions to a customer. The black
box functions were clearly defined but no mention of how it needed
to be powered (I asked AC? Battery? Hampster?), external
connectors, etc. What is a short beep? Long beep? Someone actually
got reamed by a customer because he didn't like the sound of the
beep.

In one
of my roles, I work sound for a dinner cabaret production
(http://www.teatrozinzanni.com).

Looks like a place where you could rent yourself out to women who
don't get it at home

There are moments where someone's on a
tight rope or trapeze, and the audience is dead silent. If the bass
suddenly went "pop" there'd at least be a very pissed-off stage manager, and
the distraction could even cause a performer to be injured.

No net? What if a kid screams at the circus and the cannon goes off
as the projectile is getting inside? No one's ever gotten so
absorbed by the action they dropped their glass?

Well, a design with no troubleshooting ain't worth a lot

It is to me. I like it when it works the first time and I've had
great luck via proper planning.

But I think I've heard the pop that plagues you and there's nothing
that could've prevented that snag.

Yes, the conclusion I came to was that the transient was out of my control;
I could not rely on the electronics to be silent during a power-on or
power-off event. This is not surprising, since they're virtually all
single-supply devices biased at 1/2Vcc with DC blocking caps on the output.
I had hoped that by gradually ramping the supply up or down I could evade,
but what I found was that I have to go so slowly that there would be a
noticeable turn-on lag.

And the mute time won't be noticeable? What if the TTP (time-to-
pop) varies from model to model? :-0

I take no offense from the questions; I'd be asking the same ones if I were
trying to help someone, and I appreciate your help. For the sake of detail,
though: I do have an EE degree, and I've been doing analog electronics as a
hobby for about 25 years. I'm not a great engineer, but I do claim to have
a reasonable understanding of at least some of the rudiments. (However, I
never really worked with MOSFETs until recently.)

ME either. Same with thyristors. I started a business 12yrs ago
that required little electronic knowlege. Good money, no one
looking over my shoulder. Gotta start it back up to fuel this
electronics/computer biz 'cause there isn't enough of that work
around here.

You can see one typical example at
http://www.cafewalter.com/cafewalter/fetpre/htfld.gif. That's not the one
that I happen to be testing with at the moment; but the differences aren't
huge. Practically all the active pres I've looked at are based on TL062's,
with an input buffer followed by an active EQ.

I like those TLxxx's. Check it out. When the plug is in, look at
the path from JT through R15 C12... R13, C2 R2 and all that. That's
an RC time constant that might varie from model to model, thus
unpredictable TTP. If that's the source of your woes, the volume
control won't help.

That piezo pickup is a new one on me.

Mike

 

[Active8]

On 15 Dec 2003 05:56:21 GMT, [email protected] said...

Interesting. I went down and grabbed a standard ol' open-frame Switchcraft
stereo jack and a Switchcraft mono plug, and confirmed with a continuity
tester that the plug T definitely leaves the jack S before it touches the
jack R; that is, the first thing that touches the jack R is the plug T, and
when they touch, the plug T is not touching anything else. At the moment
this contact occurs, plug S is contacting jack S.

With no plug, the circuit looks like this:
(jack S) ----- (electronics) ----- (+ bat -) ------- (jack R)

So, when the plug T touches jack R, and plug S touches jack S, it completes
a circuit and roughly 9v is placed across whatever the plug is connected to
(i.e., the bass amp). As soon as the plug is inserted the rest of the way,
this is no longer true.

Perhaps this varies with different jack and plug styles. The Switchcrafts
are pretty common, though (although the closed-frame ones are slightly more
common in basses than the open-frame ones, IME).




Let me be clear: this device is not in beta, it's just on my workbench. So
far the only bass player that has experienced this is me. But having built
and repaired a lot of equipment for bass players, I'm pretty confident that
if the bass was just sitting there, a few minutes after a song (e.g., during
a dead spot in a theatrical performance) and suddenly it made a loud "pop",
that would be perceived as a significant problem.



Yes, same noise.



In a noisy bar, no problem. In a Broadway theater pit, big problem. In one
of my roles, I work sound for a dinner cabaret production
(http://www.teatrozinzanni.com). There are moments where someone's on a
tight rope or trapeze, and the audience is dead silent. If the bass

<snip>

Sorry. hit a Ctrl - something by accident. I was going to add to my
stoopid user tricks rant...

A common mistake that people make when trying to design something
completely foolproof is to underestimate the ingenuity of complete
fools.
-Douglas Adams

I like that.

Mike

 

[Walter Harley]

OTOH, no thump *would* be nice. It's sometimes hard to strike a
balance between functionality and preventing the eventuality of
stoopid user tricks. With my software, I started getting tired of
shit like validating input to the umpteenth degree. What are the
chances that someone's going to enter a letter where it's obvious
that a number needs to be entered?

Depends on the application. Certainly, one doesn't always do what one means
to do, even if it's obvious: people get distracted, people hit the wrong
buttons. A program that crashed when invalid data was entered would always
be unacceptable to me, not to mention being exactly the sort of security
hole that people always fault MSFT from writing by the bucketload. A
program that simply ignored the keystrokes? Well, might be okay or not,
depends on the application.

In the case of my device, the whole thing is essentially a nicety - after
all, people could just remember to unplug the cable when the instrument
isn't in use, just like they have to at present. So I think to be
marketable (and I'm not trying to take the world by storm with these, just
trying to add them to my little product line), it needs to be very
well-behaved.

No net? What if a kid screams at the circus and the cannon goes off
as the projectile is getting inside? No one's ever gotten so
absorbed by the action they dropped their glass?

No net. Of course the performers are not relying on absolute silence; but
also, they're expecting noises from the audience more than from the stage.
Performers do sometimes get hurt; distractions don't necessarily cause, but
they contribute. On a good day, they're working well within their margin of
safety, but that margin gets eroded by a variety of factors: if it's humid
and their hands or feet get sweaty, if they are sick or tired, if the
audience is doing distracting things, if a cue is screwed up
or their props are misplaced, and so forth.

I like those TLxxx's.

The TL0xx series is nice (as are the TLE20xx updated versions). The TL06x
in particular, though, has a noise profile that I find kind of ugly - it's
got a kind of grainy, muddy sound. Not surprising given its very low power
consumption and relatively (for the power) high bandwidth.

That piezo pickup is a new one on me.

Piezo pickups mounted under the saddle (the contact point of the strings at
the bridge) are becoming moderately popular. Two problems with them are (1)
they are very sensitive to mounting and (2) they are *extremely* sensitive
to the impedance they see. In the particular circuit I showed you, the
piezo sees an impedance of as low as 12k (IIRC), which is about two orders
of magnitude too low for it to have decent bass response. On the other
hand, if you show them a high impedance (10MEG), they can generate high
enough voltages that a 9v system can't reproduce them without clipping.

 

[Active8]

Depends on the application. Certainly, one doesn't always do what one means
to do, even if it's obvious: people get distracted, people hit the wrong
buttons. A program that crashed when invalid data was entered would always
be unacceptable to me, not to mention being exactly the sort of security
hole that people always fault MSFT from writing by the bucketload. A
program that simply ignored the keystrokes? Well, might be okay or not,
depends on the application.

Yeah, accidents. I hate the way you have to do a

On Error Goto in VBA and remember to not let a function fall
through to the error handler. Snippet managers work.

What I try to do (I prefer C++) is keep a try catch block snippent
and I've disciplined myself to use it. Then I write functions that,
if there's an error, just go back to the last state. In other
words, even if i leave my catch block empty, the call stack unwinds
to the calling proceedure. So there's no error message, but nothing
adverse happens. I also use error handling classes, so I can just
type in a string and it ends up in an error message box.

In the case of my device, the whole thing is essentially a nicety - after
all, people could just remember to unplug the cable when the instrument
isn't in use, just like they have to at present. So I think to be
marketable (and I'm not trying to take the world by storm with these, just
trying to add them to my little product line), it needs to be very
well-behaved.

What other products are in your line?

The TL0xx series is nice (as are the TLE20xx updated versions). The TL06x
in particular, though, has a noise profile that I find kind of ugly - it's
got a kind of grainy, muddy sound. Not surprising given its very low power
consumption and relatively (for the power) high bandwidth.

TLE20xx... noted.

Piezo pickups mounted under the saddle (the contact point of the strings at
the bridge) are becoming moderately popular. Two problems with them are (1)
they are very sensitive to mounting and (2) they are *extremely* sensitive
to the impedance they see. In the particular circuit I showed you, the
piezo sees an impedance of as low as 12k (IIRC), which is about two orders
of magnitude too low for it to have decent bass response. On the other
hand, if you show them a high impedance (10MEG), they can generate high
enough voltages that a 9v system can't reproduce them without clipping.

Ah, so.

See Ya,
Mike

 

[Walter Harley]

What other products are in your line?

http://www.cafewalter.com. And I do mean "little product line"

Eventually I hope to do electronics fulltime. Analog was my love as a kid
and my goal as an undergrad, then I spent a while in academia studying
cognitive psych in grad school and finally ended up in the software
industry.

 

[Active8]

On 17 Dec 2003 07:54:30 GMT, [email protected] said...

http://www.cafewalter.com. And I do mean "little product line"

Gotta start somewhere.

And food! Recipes. Food. Damnit! I was fine and now I'm hungry.

Eventually I hope to do electronics fulltime. Analog was my love as a kid
and my goal as an undergrad, then I spent a while in academia studying
cognitive psych in grad school and finally ended up in the software
industry.

Furman. We have Furman's Foods cannery here. Did some pgming for
them. They go by Furmano's on their labels like the Italians in NY
are to stupid to read the address on the back. I guess it worked
for a while, but some broker's costing them business. I don't know
many people who read labels.

Well. looks like we've almost managed to get completely OT. That's
the norm around here

Mike

 

[John Woodgate]

I read in sci.electronics.design that Walter Harley

Piezo pickups mounted under the saddle (the contact point of the strings
at the bridge) are becoming moderately popular. Two problems with them
are (1) they are very sensitive to mounting and (2) they are *extremely*
sensitive to the impedance they see. In the particular circuit I showed
you, the piezo sees an impedance of as low as 12k (IIRC), which is about
two orders of magnitude too low for it to have decent bass response. On
the other hand, if you show them a high impedance (10MEG), they can
generate high enough voltages that a 9v system can't reproduce them
without clipping.

You can use a lower resistive input impedance AND reduce the output
voltage while maintaining the bass response by putting a capacitor
across the input. For example, if you need a 1 Mohm load for good bass,
but you then get 4 V rms out (which will clip an amplifier with a 9 V
supply), make the resistance 0.5 Mohm and put a capacitor equal in value
to the pickup capacitance in parallel with the input. You get the same
bass response and 2 V rms out, with just 0.5 Mohm input resistance.

You can push this a lot further; 9 times the pickup capacitance gives
0.4 V out across 100 kohms. Eventually, with a big enough shunt
capacitor, the equivalent series resistance of the pickup cuts the high
frequency response, so you can't push too far.

 

[Walter Harley]

You can use a lower resistive input impedance AND reduce the output
voltage while maintaining the bass response by putting a capacitor
across the input. [...]

Thanks, John, that's a helpful observation.