Need a sample & hold, low frequency to voltage circuit, no output drift.

[[email protected]]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

I have a generated audio range of 50 Hz to 300 Hz

I would like to correlate the Output voltage range as:
5v @ 50 Hz - to - 8v @ 300 Hz

A new update or sample pulse would be once every 10 minutes. The
output voltage must hold solid, in other words absolutely drift/decay
free between 10 minute intervals.

I tried a 1 Meg resistor and a 1 uf film capacitor on a TL082 OP Amp
input with the unacceptable output voltage decay shown below.

Decay Time
5.58v 12:33
5.57v 12:36
5.56v 12:41
5.55v 12:46
5.54v 12:53

The ideal chip would be 8 pin and drift free on the Output. It needs
to operate from a single 9 volt battery, so minimum current drain is
important.

Any chip or circuit recommendation is appreciated.


Thank you for your insight and sharing your knowledge


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[[email protected]]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

I have a generated audio range of 50 Hz to 300 Hz

I would like to correlate the Output voltage range as:
5v @ 50 Hz - to - 8v @ 300 Hz

A new update or sample pulse would be once every 10 minutes. The
output voltage must hold solid, in other words absolutely drift/decay
free between 10 minute intervals.

I tried a 1 Meg resistor and a 1 uf film capacitor on a TL082 OP Amp
input with the unacceptable output voltage decay shown below.

Decay Time
5.58v 12:33
5.57v 12:36
5.56v 12:41
5.55v 12:46
5.54v 12:53

The ideal chip would be 8 pin and drift free on the Output. It needs
to operate from a single 9 volt battery, so minimum current drain is
important.

Any chip or circuit recommendation is appreciated.

I can do the frequency to voltage conversion fine. What I need is an 8
pin chip that can work at 9 volts with low current drain. It must take
a 4 to 8 volt sample voltage at the input and hold that same value on
the Output indefinitely or until it receives an update pulse.

Thanks again,

* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[Bob Eld]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

I have a generated audio range of 50 Hz to 300 Hz

I would like to correlate the Output voltage range as:
5v @ 50 Hz - to - 8v @ 300 Hz

A new update or sample pulse would be once every 10 minutes. The
output voltage must hold solid, in other words absolutely drift/decay
free between 10 minute intervals.

I tried a 1 Meg resistor and a 1 uf film capacitor on a TL082 OP Amp
input with the unacceptable output voltage decay shown below.

Decay Time
5.58v 12:33
5.57v 12:36
5.56v 12:41
5.55v 12:46
5.54v 12:53

The ideal chip would be 8 pin and drift free on the Output. It needs
to operate from a single 9 volt battery, so minimum current drain is
important.

Any chip or circuit recommendation is appreciated.


Thank you for your insight and sharing your knowledge


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

Is the held voltage AC or DC? It sounds like DC the way you describe it but
the discussion of 50Hz and 300Hz is confusing. Is it the peak of these that
is held or some sort of synthesis voltage that is held. More importantly,
what are you trying to do, stabilize a frequency?

If it is DC and you want to hold it indefinitely, digitize it and hold it in
RAM for ever or as long as you want. A PIC processor with built in 10 bit
ADC will do this to 0.1% accuracy. It will require an DAC to return the held
digital word to analog. That's a two chip solution

An other way would be to phase lock the synthesized frequency to a divided
down crystal oscillator to create a stable output frequency. No attempt
would be made to hold a DC voltage.

 

[[email protected]]

Is the held voltage AC or DC? It sounds like DC the way you describe it but
the discussion of 50Hz and 300Hz is confusing. Is it the peak of these that
is held or some sort of synthesis voltage that is held. More importantly,
what are you trying to do, stabilize a frequency?

If it is DC and you want to hold it indefinitely, digitize it and hold it in
RAM for ever or as long as you want. A PIC processor with built in 10 bit
ADC will do this to 0.1% accuracy. It will require an DAC to return the held
digital word to analog. That's a two chip solution

An other way would be to phase lock the synthesized frequency to a divided
down crystal oscillator to create a stable output frequency. No attempt
would be made to hold a DC voltage.

I hope to get a manufactures number for a single chip that operates on
a single supply at nine volts.

It must sample a voltage between 5 and 8 volts of smooth DC and hold
that value on the output indefinitely or until an update pulse is sent
to the chip.

The desired chip updates are only needed every 10 minutes or so.

Basically I am monitoring the "thermal drift" that is detected in a
theremin musical instrument; which is heard in the audio as a slow
audio frequency drift as the theremin sits idle. My circuity already
compensates for this but I need a chip that Inputs a voltage sample
from my freq to DC voltage converter circuit at a given moment and
hold that same voltage on the Output indefinitely.


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[Randy Day]

Bob Eld wrote:

[snip

Is the held voltage AC or DC? It sounds like DC the way you describe it but
the discussion of 50Hz and 300Hz is confusing. Is it the peak of these that
is held or some sort of synthesis voltage that is held. More importantly,
what are you trying to do, stabilize a frequency?

If it is DC and you want to hold it indefinitely, digitize it and hold it in
RAM for ever or as long as you want. A PIC processor with built in 10 bit
ADC will do this to 0.1% accuracy. It will require an DAC to return the held
digital word to analog. That's a two chip solution

The dsPIC's have built-in DAC's as well as ADC's.

Another option might be to use the PIC ADC to sample
the AC directly, count the number of peaks and output
a DAC voltage. That would eliminate the F-V conversion
circuit and make it a real one-chip solution.

 

[[email protected]]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

After a little research please allow me a re-word my original
question.

Can you give me a manufactures part number for a chip on the market
that can operate on a single 9 volt supply? It needs to be able to
"Sample" a DC voltage on the Input like 5.67 volts, Hold that voltage
on the Output with an accuracy of +/- .01 volts then maintain that
exact value for at least10 minutes or another Update pulse is sent to
Sample again or the power is turned off. The Input DC voltage sampling
range slowly drifts between 5 volts and 8 volts.

One approach seems to be a single channel ADC to DAC, maybe on one
chip?

But what part #, I can't seem to find a choice from searching with the
keywords Sample & Hold?

Thanks again,


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[Jasen Betts]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

can you fake it with a microcontroller with PWM output driving a
low-pass filter.

I have a generated audio range of 50 Hz to 300 Hz

I would like to correlate the Output voltage range as:
5v @ 50 Hz - to - 8v @ 300 Hz

to do that in single chip may need to power the microcontroller
from +9v to VCC and +4V to VDD.

Bye.
Jasen

 

[[email protected]]

I can do the frequency to voltage conversion fine. What I need is an 8
pin chip that can work at 9 volts with low current drain. It must take
a 4 to 8 volt sample voltage at the input and hold that same value on
the Output indefinitely or until it receives an update pulse.

---

Vin>---[ADC]---[LATCH]---[DAC]-->Vout
| |
SAMPLE>--+--------+

Thank you John, that gave me the proper logic for the information I
began searching for. Google is not my friend on this one. ;-(



* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[[email protected]]

BTW, why are you limited to an 8 pin chip?

I am not limited to 8 pins. It was just a thought that
there must be the ideal chip out there I could use that would
temporarily match the Output voltage with the Input voltage at each
sample Input update pulse.

Two pins for 9 volt "power",

One pin for a DC voltage level analog "Input",

One pin for a DC voltage level analog "Output",

One pin for adjusting a "Clock Frequency".

One pin for Outputting a clock pulse which goes to the
"sample update" pin, I would not use this, it
is done by my existing circuit or manually with
a conditioned pulse.

One pin to "enable" the update sample & latch using
leading edge and trailing edge of the clock
pulse for temporary digital storage.

One pin for what else the chip would need.


Thank you JF.

* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[Robert Monsen]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.

I have a generated audio range of 50 Hz to 300 Hz

I would like to correlate the Output voltage range as:
5v @ 50 Hz - to - 8v @ 300 Hz

A new update or sample pulse would be once every 10 minutes. The
output voltage must hold solid, in other words absolutely drift/decay
free between 10 minute intervals.

I tried a 1 Meg resistor and a 1 uf film capacitor on a TL082 OP Amp
input with the unacceptable output voltage decay shown below.

Decay Time
5.58v 12:33
5.57v 12:36
5.56v 12:41
5.55v 12:46
5.54v 12:53

The ideal chip would be 8 pin and drift free on the Output. It needs
to operate from a single 9 volt battery, so minimum current drain is
important.

Any chip or circuit recommendation is appreciated.


Thank you for your insight and sharing your knowledge


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

What you are really talking about is sampling some kind of tuner every
10 minutes.

There are various FtoV chips, like the TC9401 or the AD650, that will
convert a frequency into a voltage.

TC9401: http://ww1.microchip.com/downloads/en/DeviceDoc/21483d.pdf

Then, you want to simply record the voltage once every 10 minutes, and
output the changes that might occur.

Using a traditional sample and hold is probably not the right solution
here. I'd say that you want to store the voltage digitally, so it'll
be easy to see the changes over time. Recording the result using a
microcontroller makes lots of sense. They generally have non volatile
memory, so you can wake it up when you want a sample, and then put it
back to sleep to save power. Many microcontrollers have fairly
accurate atod converters.

Note that since you are trying to capture changes with temperature,
you'll need to ensure that the FtoV converter and voltage reference
for the AtoD doesn't also drift with temperature. I believe this will
be nearly impossible without doing extensive temperature testing, and
then table lookups in software. The temperature drifts will be
different for each device. The drift over time will be an issue as
well. However, the AD650 has a temperature coefficient of 250ppm/C, so
a 10C drift will only cause a 0.25% drift in frequency. The external
parts will also need to be matched, according to the data sheet here:

http://www.analog.com/static/imported-files/data_sheets/AD650.pdf

You could also use an AtoD converter after the FtoV chip, and pass the
input into a PC, recording the result digitally over time. This has
the same problems, and you also need a PC.

Another solution might be to simply record these 'frequency pings'
over time, using the sound input on a PC. You could analyze the drift
over time using some software. This is probably much easier and more
accurate than you will get using the above method, since it probably
won't drift with temperature. However, you will need a PC.

You could use a beagleboard, and write the software to run on linux.
Not a bad solution for $150 + power adapter. Any control might be
accomplished using the sound output. You could also hook up a monitor
to visually display the tuning over time.

http://beagleboard.org/

Sounds like an interesting project.

Regards,
Bob Monsen

 

[[email protected]]

---
I think you're going to find it from very difficult to impossible to
find a COTS single chip solution for your application.

If you can disclose some details about your application I'm sure at
least one of us will be able to help you wind up with something
practical and workable.

Are you using a microcontroller at all?

No


All my projects are eventually made public, here is a sneak peek, hope
it does not make it even more confusing. The technique works perfect
other than what I believe is called capacitor droop!

http://www.oldtemecula.com/theremin/tan/index.htm

* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[[email protected]]

By the way, have you seen the video (DVD) called An Electronic Odyssey (the
story of Leon's life)? It's well worth getting if you're a Theremin fan.

Yes a wonderful movie!

Hmmm.... Could you explain the concept of a "tuned" Theremin?

Thank you for your interest. A theremin is musical instrument played
without touching anything.

Normally the thereminist tunes a theremin to place a specific note,
lets say, musical note "A1" eighteen inches from the pitch antenna.

Due to thermal and environmental conditions such as humidity this
moves or drifts away from this original set location.

I would like to donate a theremin to a local kids museum where tuning
the instrument is beyond the ability of people trying it out.

I can have it self tune if I can overcome the capacitor droop problem
of a basic sample & hold circuit.

Please visit my theremin webpage to understand my passion for the
instrument.


http://www.oldtemecula.com/theremin/index.htm


* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

 

[Bob Eld]

No


All my projects are eventually made public, here is a sneak peek, hope
it does not make it even more confusing. The technique works perfect
other than what I believe is called capacitor droop!

http://www.oldtemecula.com/theremin/tan/index.htm

* * * *

Christopher

Temecula CA.USA
http://www.oldtemecula.com

How is the hunt for a long term single chip sample and hold going? I agree
with John, I doubt you are going to find one, a least I don't know of one.

The solution to your problem could be a microprocessor, but once you have
taken that step, the whole circuit with the 555's etc. in the above URL can
be replaced with a single chip processor solution. Everything you are trying
to do can be accomplished within a processor, reference frequencies,
detection of Theremin oscillation and calibration, control and tuning of
same, long term holding of values, accurate timing of any number of events,
accurate crystal controlled frequencies, and so on, the list is endless.

I think you may be trying to pound a square peg into a round hole with a
less than ideal analog solution.

 

[whit3rd]

I am working on a hobby project and would like to find a "single chip"
solution or "circuit" for a sample & hold frequency to voltage
converter with a no drift output.  

As others have said, a true 'no drift' solution will generally involve
latches (i.e. it's a digital device you want). For this kind of
thing,
a tracking A/D converter is the gizmo you want. This consists of a
clock,
a comparator, and up/down counter feeding a D/A converter.
To track, compare the target voltage with the D/A converter output,
and count UP if the DAC is low, DOWN otherwise. To hold, just
disconnect the clock from the counter. Inexpensive microprocessors
like PIC16F785 have 10-bit A/D conversion and can do this in
software (use the converter to 'servo' an op amp integrator to make
an analog output).

I think an analog solution will work, too. If the time isn't days,
the drift can be low enough if you use a low-input-current op amp,
an analog switch, and polystyrene or teflon film capacitors.
The capacitor and a resistor and one op amp form an
integrator; switch this integrator between an error current source
(just a resistor from an amplified error-voltage signal) and an
offset trim source (millivolts-from-ground trimmer with a 1M limit
resistor) using an SPDT connected analog switch.
The offset trim should be able to zero the drift of the output.
I've made 10 mV/day ramp generators this way.

 

[Robert Monsen]

On Thu, 28 Aug 2008 23:32:52 GMT, [email protected] wrote:

<snipped>

I wonder if you could use a GP2D12 Sharp IR distance sensor to build
one of these? The device is sensitive out to about 50cm. You would
need a way to linearize the output (probably a lookup table of some
kind), and a way to generate a sine wave at the proper frequency. One
way would be an intersil 8038 or its bretheren, using a PLL to set the
frequency.

However, there is no information in the datasheet about temperature
sensitivity.

Regards,
Bob Monsen