Maker Pro
Maker Pro

breadboarding and stray capacitance

M

Merlin

Jan 1, 1970
0
Hi,

I have been experimenting with making an oscillator circuit running around
400-500 kHz (some of you probably recognize this from previous questions). So
far out of 3 different designs I keep getting practical results that are far
below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have a
certain amount of capacitance in them. But could the breadboard be the reason
for a large discrepancy between expected and observed frequencies at frequencies
around 500kHz?

It seems that calculated vs observed differences get worse as I increase in
frequency. It is relatively close around 80kHz but as I ramp up the frequency
the output amplitude decreases and eventually the oscillator stops. The best I
seem to get is 140kHz to 250kHz depending on which design I use. In both cases
the calculated values should give around 480kHz.

I have thought of using a proto board of some sort instead of breadboard but I
was hesitant about spending the time soldering together another failure.

Any practical insite would be appreciated.

Merlin.
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Merlin <[email protected]>
wrote (in said:
Hi,

I have been experimenting with making an oscillator circuit running around
400-500 kHz (some of you probably recognize this from previous questions). So
far out of 3 different designs I keep getting practical results that are far
below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have a
certain amount of capacitance in them. But could the breadboard be the reason
for a large discrepancy between expected and observed frequencies at frequencies
around 500kHz?
It depends how you lay out your circuit. Capacitances in 500 kHz
oscillators tend to be hundreds of picofarads or more, so there are few
places where a few tens of picofarads can make a substantial difference.

Maybe you should post a schematic and a clear picture or sketch of your
board to alt.binaries.schematics.electronics (a.b.s.e).
 
R

Rene Tschaggelar

Jan 1, 1970
0
Merlin said:
Hi,

I have been experimenting with making an oscillator circuit running
around 400-500 kHz (some of you probably recognize this from previous
questions). So far out of 3 different designs I keep getting practical
results that are far below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have
a certain amount of capacitance in them. But could the breadboard be
the reason for a large discrepancy between expected and observed
frequencies at frequencies around 500kHz?

It seems that calculated vs observed differences get worse as I increase
in frequency. It is relatively close around 80kHz but as I ramp up the
frequency the output amplitude decreases and eventually the oscillator
stops. The best I seem to get is 140kHz to 250kHz depending on which
design I use. In both cases the calculated values should give around
480kHz.

I have thought of using a proto board of some sort instead of breadboard
but I was hesitant about spending the time soldering together another
failure.

Any practical insite would be appreciated.

What kind of values did you calculate ? L ? C ?
The best way of prototyping such circuits is to
take an unetched piece of print material and
work 3D on it. Place short wires to connect components.
The copper surface is GND then.

Rene
 
J

John Larkin

Jan 1, 1970
0
Hi,

I have been experimenting with making an oscillator circuit running around
400-500 kHz (some of you probably recognize this from previous questions). So
far out of 3 different designs I keep getting practical results that are far
below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have a
certain amount of capacitance in them. But could the breadboard be the reason
for a large discrepancy between expected and observed frequencies at frequencies
around 500kHz?

It seems that calculated vs observed differences get worse as I increase in
frequency. It is relatively close around 80kHz but as I ramp up the frequency
the output amplitude decreases and eventually the oscillator stops. The best I
seem to get is 140kHz to 250kHz depending on which design I use. In both cases
the calculated values should give around 480kHz.

I have thought of using a proto board of some sort instead of breadboard but I
was hesitant about spending the time soldering together another failure.

Any practical insite would be appreciated.

Merlin.

What kind of oscillator? If you're using opamps, many common opamps
run out of gain and slew rate around 1 MHz. I doubt the breadboard is
the problem.

John
 
J

John Jardine

Jan 1, 1970
0
Merlin said:
Hi,

I have been experimenting with making an oscillator circuit running around
400-500 kHz (some of you probably recognize this from previous questions). So
far out of 3 different designs I keep getting practical results that are far
below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have a
certain amount of capacitance in them. But could the breadboard be the reason
for a large discrepancy between expected and observed frequencies at frequencies
around 500kHz?

It seems that calculated vs observed differences get worse as I increase in
frequency. It is relatively close around 80kHz but as I ramp up the frequency
the output amplitude decreases and eventually the oscillator stops. The best I
seem to get is 140kHz to 250kHz depending on which design I use. In both cases
the calculated values should give around 480kHz.

I have thought of using a proto board of some sort instead of breadboard but I
was hesitant about spending the time soldering together another failure.

Any practical insite would be appreciated.

Merlin.

The 'Waffle' 'proto' type breadboards have about 0.8pF between adjacent
contact strips. This is near enough unoticable to out beyond 50MHz. Some
component in your circuit is running out of steam. You'll need to show the
circuit.
regards
john
 
P

Paul Burridge

Jan 1, 1970
0
Hi,

I have been experimenting with making an oscillator circuit running around
400-500 kHz (some of you probably recognize this from previous questions). So
far out of 3 different designs I keep getting practical results that are far
below the calculated ones.

I have been breadboarding the circuits and I know that breadboards have a
certain amount of capacitance in them. But could the breadboard be the reason
for a large discrepancy between expected and observed frequencies at frequencies
around 500kHz?

I *very* much doubt it. Your strays should be pretty much negligible
compared to the component values you use at that frequency.
 
M

Merlin

Jan 1, 1970
0
thanks for all the replies. Here are a few more details about the circuit I am
currently using:

This is the web page with the schem that I am using:
http://home.att.net/~theremin1/Circuit_Library/wienbridge.htm

The calculated values I am using for C and R are 100pF and 3.3k respectively
which should give me an Fout of about 482kHz. The best I am getting is about
120kHz.

The op amp is a LMV932MA, the FET is an NTE equivalent to the 2N3819 (NTE312)
and the power supply is 2 AA batteries for +3V only.

Here are the specs:
NTE312 - http://www.nteinc.com/specs/300to399/pdf/nte312.pdf
LMV932MA - http://www.national.com/pf/LM/LMV932.html

Also I have noticed that changing 1330 or 3320 values not only changes the
amplitude but also affects the frequency of the output. I had thought that
these values would have little to no effect on the frequency.

Up until now my layouts have been quite compact on the breadboard but now I'm
thinking I should space things out. At any rate, comments and suggestions on
how to get this thing closer to the calculated values are definitely welcome.

M.
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Merlin <[email protected]>
wrote (in said:
thanks for all the replies. Here are a few more details about the
circuit I am currently using:

This is the web page with the schem that I am using:
http://home.att.net/~theremin1/Circuit_Library/wienbridge.htm

I thought you were using an LC oscillator.
The calculated values I am using for C and R are 100pF and 3.3k
respectively which should give me an Fout of about 482kHz. The best I
am getting is about 120kHz.

The op amp is a LMV932MA, the FET is an NTE equivalent to the 2N3819
(NTE312) and the power supply is 2 AA batteries for +3V only.

Here are the specs:
NTE312 - http://www.nteinc.com/specs/300to399/pdf/nte312.pdf
LMV932MA - http://www.national.com/pf/LM/LMV932.html

The op-amp has a gain-bandwidth product of 1 MHz. That is your trouble:
you need a gain of 3, which you can perhaps get at some frequency around
300 kHz, but with a very large phase-shift which will throw your
frequency calculation out of the window.

The Wien Bridge is not a recommended circuit for 500 kHz.
 
J

Jeff Stout

Jan 1, 1970
0
As others have suggested (John Jardine, John Larkin) your op-amp is running
out of steam. The LMV932 has a Gain Bandwidth product of 1.5MHz. At 480KHz
this will give you a gain a little better than 3. Which if memory serves is
all needed by the Wien Bridge just to operate. You need at least 10 times
more gain to get this to work at 480KHz.

If your in love with the LMV932, stick another one in the loop with a
positive gain of 10.

Jeff Stout
 
J

Jeff Stout

Jan 1, 1970
0
Jeff Stout said:
If your in love with the LMV932, stick another one in the loop with a
positive gain of 10.

Sometimes I'm so dense. I just finished saying that the opamp ain't good
for a gain of 3, then I say add another amp and configure it for a gain of
10. Flame away...

Jeff Stout
 
B

Ben Bradley

Jan 1, 1970
0
In sci.electronics.design said:
thanks for all the replies. Here are a few more details about the circuit I am
currently using:

This is the web page with the schem that I am using:
http://home.att.net/~theremin1/Circuit_Library/wienbridge.htm

The calculated values I am using for C and R are 100pF and 3.3k respectively
which should give me an Fout of about 482kHz. The best I am getting is about
120kHz.

You've definitely reached the upper frequency limit of that circuit
with that op-amp. Going to PCB or different prototyping method won't
change the problem. It might be workable to use an opamp with a higher
gain-bandwidth product. This is probably past the upper limit of where
I would use a Wein Bridge oscillator circuit anyway.
The op amp is a LMV932MA, the FET is an NTE equivalent to the 2N3819 (NTE312)
and the power supply is 2 AA batteries for +3V only.

Here are the specs:
NTE312 - http://www.nteinc.com/specs/300to399/pdf/nte312.pdf
LMV932MA - http://www.national.com/pf/LM/LMV932.html

With a gain-bandwidth product of 1MHz, this circuit isn't going to
work at all at the frequencies you want. You may not find an op-amp
that meets or even comes close to your specs. I really think some
other circuitry (not sure what offhand) would work better.
Also I have noticed that changing 1330 or 3320 values not only changes the
amplitude but also affects the frequency of the output. I had thought that
these values would have little to no effect on the frequency.

At much lower frequencies, they probably wouldn't, they probably
would only change the amplitude.
Up until now my layouts have been quite compact on the breadboard but now I'm
thinking I should space things out.

The layout for this circuit doesn't much matter.
At any rate, comments and suggestions on
how to get this thing closer to the calculated values are definitely welcome.

Don't use this circuit. Maybe you could reverse-engineer the thing
in the pictures you posted (trace all the connection with a DMM, one
with a 'beep' for conductivity would be very convenient), without
knowing what the 8-pin device is, and then maybe we could figure out
what it should be from the rest of the schematic. Resistors can of
course be measured, and if you don't have a C meter (which may not
work well in-circuit anyway), capacitor values can be grossly (no pun
intended) estimated by their size and type.
Actually, I want to know what it is, because I see about a dozen
capacitors (more than the sum of all other components!), and I wonder
what they all do.

It also depends on the impedances in the circuit as well as
frequency. In this circuit, impedances are low enough that breadboard
capacitance is insignificant.
 
J

John Larkin

Jan 1, 1970
0
thanks for all the replies. Here are a few more details about the circuit I am
currently using:

This is the web page with the schem that I am using:
http://home.att.net/~theremin1/Circuit_Library/wienbridge.htm

The calculated values I am using for C and R are 100pF and 3.3k respectively
which should give me an Fout of about 482kHz. The best I am getting is about
120kHz.

The op amp is a LMV932MA, the FET is an NTE equivalent to the 2N3819 (NTE312)
and the power supply is 2 AA batteries for +3V only.


The NE5532 on your schematic is an 8 MHz part, so should oscillate to
1 MHz or so, but will need bigger supplies. The LMV932 is too slow.

Three volts is kinda low for driving the fet, espacially if you bias
the source halfway up.

There are lots of fast low-voltage r-r opamps around these days.

John
 
G

Guy Macon

Jan 1, 1970
0
Unless you are working with high frequencies, the stray capacitance
of the usual white breadboarding system isn't a problem. The reason
those boards get such a bad rap is because of numbskulls who don't
follow directions. There is a maximum lead size, and if you use a
component with a thicker lead, it spreads the contacts, and makes
that hole intermittent with normal leads. This creates a time bomb
for future users of that board.
 
J

John Jardine

Jan 1, 1970
0
John Woodgate said:
I read in sci.electronics.design that Merlin <[email protected]>


I thought you were using an LC oscillator.

The op-amp has a gain-bandwidth product of 1 MHz. That is your trouble:
you need a gain of 3, which you can perhaps get at some frequency around
300 kHz, but with a very large phase-shift which will throw your
frequency calculation out of the window.

The Wien Bridge is not a recommended circuit for 500 kHz.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

Yes. In the olden days, the Wien network's low Q (i.e. large frequency shift
for small network phase change) made it pretty much useless above a few
hundred kHz when used with the relatively slow amplifiers then easily
available. Poor old Wien got a bad press and it's stayed that way to this
day.
Even then though, (late 50's) the Marconi TF1370 could do 10Hz to 10MHz
using a straight forward Wien bridge, (it even used toobs!).
The technology has moved on. Modern video amps (etc) are fast enough to
offer little phase shift upto into the tens of MHz, allowing application of
the Wien network into these R.F areas.
There's little else available that offers such a wide tuning range with
minimal components.
Bring back the Wien network. All is forgiven!.
regards
john
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that John Jardine
Even then though, (late 50's) the Marconi TF1370 could do
10Hz to 10MHz using a straight forward Wien bridge, (it even used
toobs!).

There were other video oscillators which were similar. But toobs have
BIG gain-bandwidth products compared with op-amps.
The technology has moved on. Modern video amps (etc) are fast
enough to offer little phase shift upto into the tens of MHz, allowing
application of the Wien network into these R.F areas. There's little
else available that offers such a wide tuning range with
minimal components.

The OP has never indicated a need for a wide tuning range.
Bring back the Wien network. All is forgiven!.

Yes, where appropriate. Not here, I think.
 
M

Merlin

Jan 1, 1970
0
Once again, thanks to all for the replies. I'm catching up, remembering, and
learning a lot. It has been quite a while since I have done anything serious in
electronics.

Ok, I definitely missed the gain-bandwidth being to small (actually forgot
multiply the gain needed, 3, by the bandwidth, 500KHz, and was thinking 'hmm,
I'm at 500KHz and it can handle over 1MHz.. no problem'). It looks like the
LMV932 is out of it for sure.

But a couple of questions to top off my learning so I know how to choose an op
amp for the next project (if it fits):
1. The specs say a GBW of 1.4 MHz. Jeff Stout suggested that I need at least 10
times this gain to get the circuit to work at 480KHz. Why so much larger? 3 *
500K = 1.5MHz seems to be within the range of the specs. I know that in some
cases you use the rule of thumb of doubling the calculated bandwith but I don't
know exactly why in this case or how much headroom to leave.

2. Some are saying that 500KHz is kinda past the limit of wein bridges, others
are saying it should be ok up to 1MHz with a fast enough op amp. Why would wein
bridges not handle 500KHz (with a fast op amp or discrete)?

3. John, you mention that 3V is kinda low for the FET. Admitantly I have little
understanding of this part of the circuit. I know that it is essentially an AGC
to adjust the gain down as oscilation begins but I don't know the specifics of
driving the FET and what voltages should be used. Any insight to the working of
this part would be appreciated.

4. When it is mentioned that the LMV932 is too slow are you just refering to the
GBW or are the other parameters that I should be looking at? I will look and
see what I can find.

Sorry for all the newbie questions and thanks for all the help.

Merlin.
 
M

Merlin

Jan 1, 1970
0
Ben said:
[snip]

Don't use this circuit. Maybe you could reverse-engineer the thing
in the pictures you posted (trace all the connection with a DMM, one
with a 'beep' for conductivity would be very convenient), without
knowing what the 8-pin device is, and then maybe we could figure out
what it should be from the rest of the schematic. Resistors can of
course be measured, and if you don't have a C meter (which may not
work well in-circuit anyway), capacitor values can be grossly (no pun
intended) estimated by their size and type.
Actually, I want to know what it is, because I see about a dozen
capacitors (more than the sum of all other components!), and I wonder
what they all do.

Hi Ben,

For this project I don't have the time to reverse engineer the circuit although
like you I am damn curious as to what it really is. Once I have a solution (or
abandon) the current project I plan to come back to this and see what I can
find. I will post the results for your viewing pleasure.

Merlin.
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Merlin <[email protected]>
1. The specs say a GBW of 1.4 MHz. Jeff Stout suggested that I need at
least 10 times this gain to get the circuit to work at 480KHz. Why so
much larger? 3 * 500K = 1.5MHz seems to be within the range of the
specs. I know that in some cases you use the rule of thumb of doubling
the calculated bandwith but I don't know exactly why in this case or how
much headroom to leave.

The reason you need a lot more bandwidth is phase-shift, as I've already
pointed out. If you look at most op-amp data sheets, you'' see a curve
of phase-shift against frequency. In an oscillator circuit, this phase-
shift is converted into a frequency shift - downwards.
2. Some are saying that 500KHz is kinda past the limit of wein bridges,
others are saying it should be ok up to 1MHz with a fast enough op amp.
Why would wein bridges not handle 500KHz (with a fast op amp or
discrete)?

If you want to make an oscillator that tunes from say 150 kHz to 500
kHz, the a Wien bridge is one suitable technique. But for fixed-
frequency or narrow tuning range oscillators at 500 kHz, LC oscillators
are preferable. They don't need a gain of 3, just 1.0000001, and can be
made to stay tuned to a given frequency.
 
J

Jeff Stout

Jan 1, 1970
0
----- Original Message -----
From: Merlin <[email protected]>
Newsgroups: sci.electronics.design
Sent: Thursday, March 04, 2004 2:20 PM
Subject: Re: breadboarding and stray capacitance


[snip]
But a couple of questions to top off my learning so I know how to choose an op
amp for the next project (if it fits):
1. The specs say a GBW of 1.4 MHz. Jeff Stout suggested that I need at least 10
times this gain to get the circuit to work at 480KHz. Why so much larger? 3 *
500K = 1.5MHz seems to be within the range of the specs. I know that in some
cases you use the rule of thumb of doubling the calculated bandwith but I don't
know exactly why in this case or how much headroom to leave.

You can make certain simplifying assumptions when calculating the circuit
gain from an opamp circuit: infinite input impedance, infinite gain, zero
output impedance. You can then write KCL's or KVL's and solve for gain or
other system parameters. However, in this situation you are breaking one of
the assumptions on which circuit gain is calculated; infinite opamp gain.

For example, the gain from an inverting opamp circuit is:

Vo/Vs = -R2/R1 (1)

where R1 is the source resistance, and R2 is the feedback resistance. But
if you take the infinite opamp gain assumption out of the picture, the
circuit gain can be shown to be:

Vo/Vs = -R2/(R1 + (R1 + R2)/Av) (2)

where Av is the opamp gain. If Av goes to infinite, you will return to the
idealized equation (1) discribed above. But if the opamp gain (Av) is 10,
or 5, or 1, the overall circuit gain could be quite a bit less than that
predicted by the original equation (1).

This is why I suggested a factor of 10 bandwidth more than common since
would suggest.

[snip]
4. When it is mentioned that the LMV932 is too slow are you just refering to the
GBW or are the other parameters that I should be looking at? I will look and
see what I can find.

Yes, although sometimes they hide this parameter in the data sheet.
Sometimes, you have to read it off a graph, or infer it from some other
number.
Sorry for all the newbie questions and thanks for all the help.

Merlin.

Jeff Stout
 
Top