audio recording on IC -help wanted

[Grumpy OM]

Sean,

The bottom line is the internal temperature should not exceed that
specified for the class of insulation. If the unit isn't marked then
assume commercial grade class A which is 105C. I have used the
resistance change method recommended by UL and find it to be simple
and accurate. Also useful for motors and solenoids.

Roughly determine the load from the core size/weight. Measure the
primary and secondary cold resistances. The more accurate the better.

Let loaded unit cook for 15 minutes while monitoring for burning
odor or smoke which indicates an overload. Disconnect power and load,
and measure resistances again.

Use the following formula to determine the hot temp of each
winding.

R2/R1 = (x + T2) / (x + T1) solve for T2.

R1 = cold resistance
R2 = hot resistance
T1 = cold temp (ambient, usually 22C)
T2 = hot temp in deg C
x = 234.5

If T2 is less than 105C then increase cooking time and measure
again. The winding closest to the core should be hotter.

Grumpy

 

[Bill Bowden]

"Howard Henry Schlunder" <howard_hs@yahoo.com> wrote in message news:<3f39a431$1_5@corp.newsgroups.com>...

That transformer was not saturating and it is perfectly normal for big
transformers to draw significant current from the mains without any load
attached.

I don't know what's going on, it seems a little lightly built
for the job. The core cross section area is 1.25 by 1.75 inch
or .00141 sq meter. The laminations are all welded together
on a base plate so the eddy currents are quite high and probably
account for most of the heat. But I don't think it should draw
3 amps without only eddy currents as a load. The primary only
has 160 turns of looks like 12 gauge wire which is a little
over one turn per volt. Most power transformers are at least
2 turns per volt.

There is a transformer formula relating voltage, frequency,
turns, area, and flux density: Erms =4.44*FNAB

and using B=1.1 for transformer steel, I get:

Erms = 4.44 * 60 * 160 * .00141 * 1.1 = 66 volts rms.

Looks like a 66 volt transformer running at 120.

-Bill

 

[Terry]

Bill Janssen wrote:

Measure the winding temperature by measuring the resistance change when cold
and when hot.

Bill: If may paraphrase? "resistance change from cold to hot".
Sounds like a smart idea; using the resistivity/temperature
coefficient of copper before and immediately after power is
disconnected.
Later I must look that up and estimate what sort of numbers we
would be working with for typical gauges of wire. In my mind it
would be the 'ratio' of the resistance change that would indicate
the temperature difference; am I right there? (I'm not sure if
I've worded that correctly or very well?)
Winding resistance of transformer at say, room temperature to
start; then after testing at load?
I guess it would be most accurate, with most tools (meters)
available to us to measure the winding with the highest
resistance?
Cheers.
PS. I've got the remains of a Wheatstone Bridge somewhere! Wonder
if one would need to measure that accurately?

 

[Tilmann Reh]

Andras Tantos schrieb:

I finally have finished putting together a description of an idea of mine: a
new A/D converter circuit. It is based on a simple idea of converting
numbers from base-10 to base-2. I describe converters which are faster
and/or smaller and/or more flexible than the ones currently available. If
you could, please take a look at it at:
http://tantos.homelinux.org/andras/work/newad/index.htm

Any comment (good or bad) would be highly appreciated.

Assume some reasonable tolerances for the matching of the resistor
pairs, and then do some simulations to check monotony and missing
codes. I'm pretty sure you will be surprised...

--
Dipl.-Ing. Tilmann Reh
Autometer GmbH Siegen - Elektronik nach Maß.
http://www.autometer.de

==================================================================
In a world without walls and fences, who needs Windows and Gates ?
(Sun Microsystems)

 

[Wouter van Ooijen (www.vo]

Any comment (good or bad) would be highly appreciated.

Assume some reasonable tolerances for the matching of the resistor
pairs, and then do some simulations to check monotony and missing
codes. I'm pretty sure you will be surprised...

Your approach is similar to the weighted-resistor D/A converter. In
fact each of your bit steps is basically a 1-bit flash A/D + a 1-bit
D/A + substract. The problem is in the reference to the comparator and
substracting the exact amount, especially in the first steps. It must
be accurate to 2^N where N==the number of steps. For a low-end 8-bit
converter you would need 'only' some 0.4% accuracy, hence 0.1%
components (more than one component contributes to the error). Think
of what you would need for a high-end 16 bit A/D!


Wouter van Ooijen

-- ------------------------------------
http://www.voti.nl
PICmicro chips, programmers, consulting

 

[Jim Thompson]

On Thu, 14 Aug 2003 09:37:21 -0700, "Andras Tantos"
<andras_tantos@tantos.yahoo.com> wrote:

Hi all!

I finally have finished putting together a description of an idea of mine: a
new A/D converter circuit. It is based on a simple idea of converting
numbers from base-10 to base-2. I describe converters which are faster
and/or smaller and/or more flexible than the ones currently available. If
you could, please take a look at it at:
http://tantos.homelinux.org/andras/work/newad/index.htm

Any comment (good or bad) would be highly appreciated.

Thanks,
Andras Tantos

Old as the hills. I've used "your" technique on multiple occasions.

See:

p364 of
Analog-to-Digital/Digital-to-Analog Conversion Techniques
David F. Hoeschele, Jr.
John Wiley & Sons, Inc.
1968
Library of Congress #: 68-22305
SBN 471 40575 2

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| Jim-T@analog_innovations.com Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

For proper E-mail replies SWAP "-" and "_"

I love to cook with wine. Sometimes I even put it in the food.

 

[Andras Tantos]

Thanks for the answer!

Your approach is similar to the weighted-resistor D/A converter. In
fact each of your bit steps is basically a 1-bit flash A/D + a 1-bit
D/A + substract.

At least in most of the constructs, that's the case.

The problem is in the reference to the comparator and
substracting the exact amount, especially in the first steps. It must
be accurate to 2^N where N==the number of steps. For a low-end 8-bit
converter you would need 'only' some 0.4% accuracy, hence 0.1%
components (more than one component contributes to the error). Think
of what you would need for a high-end 16 bit A/D!

That's true, but, that's usually true (I think) for other A/D designs too:
you need a couple of elements that are at least as precise as the LSB
precision of the converter. What's important there is that the reference to
remain in the center of the two saturation voltages of the comparator. That
can be achieved though you need a slightly more complex design than the one
I've shown there. (Note, I've tried to convey the idea only. I'm sure an
actual implementation would be much more complex than the simple circuits
I've created.) In one of my designs (the first one) I use the same elements
over and over again for producing each bit. That would reduce the number of
precise elements needed by a great deal.

Regards,
Andras Tantos

 

[Kevin Aylward]

Andras Tantos wrote:

Old as the hills. I've used "your" technique on multiple occasions.



See:

p364 of
Analog-to-Digital/Digital-to-Analog Conversion Techniques
David F. Hoeschele, Jr.
John Wiley & Sons, Inc.
1968
Library of Congress #: 68-22305
SBN 471 40575 2

...Jim Thompson

I love to cook with wine. Sometimes I even put it in the food.

That's what I was afraid of . It's too trivial for me to be first
to find it out. Only I haven't seen any commercialy available
converters using this technique nor any documentation showing such a
circuit. Could you point out an on-line version of this document (if
their's any)? Also, do you know why this technique isn't more
wide-spread?

First.

bottom post.
sure that you
to make
You need


Its just about impossible to come about with anything new. Its *all*
been done before. What surprises me, is that you have went to a lot of
trouble in your web page. Its looks all nice and neat, yet you readily
admit that you have lack "industry experience" and lack "Spice
experience". Why then, would you suppose that you could have come up
with something that no one has looked at before?

Things are the way they are today, because most things have already been
rejected, and of couse, the millions of bad ideas dont get published,
usually. I would be amazed if some *really* new and usefull, electronic
topology came about today. Most "new" ideas are simply variations on an
old theme. If you think you have something new, ask yourself "why has no
one else seen this". Really try and understand *why* your idea could
have been missed, by 100'000's of other experienced designers.

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[Jim Thompson]

On Thu, 14 Aug 2003 11:06:17 -0700, "Andras Tantos"
<andras_tantos@tantos.yahoo.com> wrote:

Old as the hills. I've used "your" technique on multiple occasions.

That's what I was afraid of . It's too trivial for me to be first to find
it out. Only I haven't seen any commercialy available converters using this
technique nor any documentation showing such a circuit. Could you point out
an on-line version of this document (if their's any)? Also, do you know why
this technique isn't more wide-spread?

Regards,
Andras Tantos


See:

p364 of
Analog-to-Digital/Digital-to-Analog Conversion Techniques
David F. Hoeschele, Jr.
John Wiley & Sons, Inc.
1968
Library of Congress #: 68-22305
SBN 471 40575 2

...Jim Thompson

As other posters have pointed out, there's accuracy problems. The
best I've been able to do, with selected components, was 7-bit. Built
around 1977. I'll see if I still have the drawings... but I doubt it.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| Jim-T@analog_innovations.com Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

For proper E-mail replies SWAP "-" and "_"

I love to cook with wine. Sometimes I even put it in the food.

 

[Wouter van Ooijen (www.vo]

That's true, but, that's usually true (I think) for other A/D designs too:
you need a couple of elements that are at least as precise as the LSB
precision of the converter.

Have you read about (for instance) dual-slope A/D's? Or 1-bit A/D's?
Get a H&H and read the A/D chapter!


Wouter van Ooijen

-- ------------------------------------
http://www.voti.nl
PICmicro chips, programmers, consulting

 

[John Larkin]

On Thu, 14 Aug 2003 11:06:17 -0700, "Andras Tantos"
<andras_tantos@tantos.yahoo.com> wrote:

Old as the hills. I've used "your" technique on multiple occasions.

That's what I was afraid of . It's too trivial for me to be first to find
it out. Only I haven't seen any commercialy available converters using this
technique nor any documentation showing such a circuit. Could you point out
an on-line version of this document (if their's any)? Also, do you know why
this technique isn't more wide-spread?

Regards,
Andras Tantos

I'm fairly sure I've seen this done somewhere, both with the s/h boxes
and I think I've seen a fully static version. But lately most fast
ADCs are pipeline architectures. This is similar to your idea, but, at
each logic level, the input is quantized to, say, 3 or 4 bits and a
residual value is computed by subtracting the D/A'd value of the
quantization from the analog input, scaling up by a gain of 8 or 16,
and passing that on to the next stage. Errors pile up, so most of
these gadgets digitize to too many bits (sometimes with
not-truly-2:1-bit weighting!) and apply a final mathematical tweak.

There are lots of parts like this, in the 8-14 bit range with speeds
from 20 MHz up to over 100. They all have a pipeline delay (analog
sample to digital out), usually in the 6-7 clock range. The "876"
types are generic, 10 bits at 20 MHz, and the TI version costs about
$3.

John

 

[Tom Bruhns]

"Andras Tantos" <andras_tantos@tantos.yahoo.com> wrote in message news:<3f3bcf9d@news.microsoft.com>...

Old as the hills. I've used "your" technique on multiple occasions.

That's what I was afraid of . It's too trivial for me to be first to find
it out. Only I haven't seen any commercialy available converters using this
technique nor any documentation showing such a circuit. Could you point out
an on-line version of this document (if their's any)? Also, do you know why
this technique isn't more wide-spread?

Oh, but it is! The conversions are commonly not a single bit, because
it's so easy to do multiple bits, but have a look at, for example, the
Analog Devices 6645. Do a google search for "pipelined adc" and see
what you get. We've used a dithered, pipelined converter circuit in
100kHz bandwidth systems for a long time, and in fact did a 23 bit one
that runs at 10 or 20 megasamples/second. Have a look for the Agilent
E1430 and E1437 VXI modules.

I didn't look at all your implementation details, but you can do one
of these as a cascade of identical stages, or by feeding the analog
residue back to the same stage multiple times; I'm not sure you picked
up on that second idea.

For accurate, slow converters, I suspect you'll find that the
delta-sigma technique is more efficient.

Note that your leadin discussion of number conversions made me think
you were going to "discover" the successive approximation converter.

Cheers,
Tom

 

[Andras Tantos]

As other posters have pointed out, there's accuracy problems. The
best I've been able to do, with selected components, was 7-bit. Built
around 1977. I'll see if I still have the drawings... but I doubt it.

Have you done it with discrete components? I wouldn't be surprised than. As
other's pointed out they've used similar techniques and reaching 23 bit
precision so it seems it can be done.

Regards,
Andras Tantos

 

[CBFalconer]

Jim Thompson wrote:

andras_tantos@tantos.yahoo.com> wrote:

.... snip ...

Have you done it with discrete components? I wouldn't be
surprised than. As other's pointed out they've used similar
techniques and reaching 23 bit precision so it seems it can
be done.

Seems to me that 23 bit performance would require incredible
matching accuracy.

You don't depend on matching, just stability. You convert the
value into a time interval and measure that.

--
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

 

[Kevin Aylward]

Tom Bruhns wrote:

"Andras Tantos" <andras_tantos@tantos.yahoo.com> wrote in message
news:<3f3bcf9d@news.microsoft.com>...
Old as the hills. I've used "your" technique on multiple occasions.

That's what I was afraid of . It's too trivial for me to be first
to find it out. Only I haven't seen any commercialy available
converters using this technique nor any documentation showing such a
circuit. Could you point out an on-line version of this document (if
their's any)? Also, do you know why this technique isn't more
wide-spread?

Oh, but it is! The conversions are commonly not a single bit, because
it's so easy to do multiple bits, but have a look at, for example, the
Analog Devices 6645. Do a google search for "pipelined adc" and see
what you get. We've used a dithered, pipelined converter circuit in
100kHz bandwidth systems for a long time, and in fact did a 23 bit one
that runs at 10 or 20 megasamples/second. Have a look for the Agilent
E1430 and E1437 VXI modules.

I didn't look at all your implementation details, but you can do one
of these as a cascade of identical stages, or by feeding the analog
residue back to the same stage multiple times; I'm not sure you picked
up on that second idea.

For accurate, slow converters, I suspect you'll find that the
delta-sigma technique is more efficient.

Note that your leadin discussion of number conversions made me think
you were going to "discover" the successive approximation converter.

And you would indeed be correct. That's what he shows, in disguise. All
the design is doing is successively comparing each subtracted signal
depending on the prior signal. This is what the successive approximation
converter does. The only basic difference here is that instead of doing
the subtraction with the dac treated as the reference signal and
modifying this reference to the comparator, the signal itself is
modified and compared to a fixed reference. The obvious disadvantage is
that you now need N comparators, rather then just one.

When you start to look at the details, it can be seen that this approach
has a few shortcomings.

For example, you can't actually use the comparator outputs to drive the
subtract amplifier. It will never swing with high enougth accuracy to
the rails, in addition, the power supply noise will kill you. So, you
would have to use the comparator signal to switch in a reference voltage
to the subtract amplifier instead. Now you have a bank of switched,
analogue outputs, with R & R resistors, i.e. you now have a standard
dac, so we are back to where we started.

Issues:

Bandwidth. The signal goes through N stages of gain of 2.
Accuracy. Gain and offset errors multiply up through the chain. How
much?
Power. N precision amplifiers. N precision comparators.

In all of these key parameters, the technique is inherently worse, then
the conventional SAC.

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[Kevin Aylward]

CBFalconer wrote:

Jim Thompson wrote:
andras_tantos@tantos.yahoo.com> wrote:

... snip ...

Have you done it with discrete components? I wouldn't be
surprised than. As other's pointed out they've used similar
techniques and reaching 23 bit precision so it seems it can
be done.

Seems to me that 23 bit performance would require incredible
matching accuracy.

You don't depend on matching, just stability. You convert the
value into a time interval and measure that.

Oh?... Since the input would go through 22 gain of two stages, what
output offset would you expect from from the first stage input offset?

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[Mjolinor]

Bill: If may paraphrase? "resistance change from cold to hot".
Sounds like a smart idea; using the resistivity/temperature
coefficient of copper before and immediately after power is
disconnected.

The problem with using the AC is that the heating effect will not be just
related to current in the wire.
Better to do it with a DC current source and just measure the volt drop as
you increase the current. The readings from that will give you length of
wire and therefore a fairly accurate number of turns, assuming your into
length of spirals (don't fancy working that out on a square former). Current
would need to be kept small in order to be sure that the temerature wasn't
greater where the wires were in the middle of the windings. Using DC would
also remove any skin effect too. Getting a bit hypothetical here

 

[Watson A.Name - 'Watt Sun]

In article <o_Z_a.483$Ng.86320@kent.svc.tds.net>, file13@hotmail.com
mentioned...

I found an Archer "Wireless Oscillator" in my parts box. It was sold by
Radio Shack some years back, part number 277-237. I would like a datasheet
and schematics on it's use. I can't find anything using Google. Hopefully,
this will be the item I need to create a wireless AM broadcaster so I can
use it as a baby monitor or listen to my MP3's while out working in the
garden. (At least I assume that that is what this is used for.) It has 6
pins on the bottom, 3 on each end.

I would appreciate any information anyone can provide on this item.
~Tom

I suggest you don't use AM, mainly because it takes a long wire for a
decent antenna. But when you use a transmitter with a microphone for
a baby monitor, your neighbors can hear everything that goes on in the
bedroom. It's a bit less of a problem if you use an FM transmitter
with lower power that transmits on the low end of the FM band, around
88 MHz. You can limit the range to a hundred feet or so by reducing
the power. You can build a decent 2 transistor wireless FM microphone
from plans on the web for a few dollars. Or btter yet, buy a kit for
ten or fifteen dollars.

Also if you're listening to music you probably want stereo. You can
buy a Ramsey FM-10 FM Tx for $35 (see www.ramseyelectronics.com). My
Sony walkman didn't even have an AM band on it.

--
@@F@r@o@m@@O@r@a@n@g@e@@C@o@u@n@t@y@,@@C@a@l@,@@w@h@e@r@e@@
###Got a Question about ELECTRONICS? Check HERE First:###
http://users.pandora.be/educypedia/electronics/databank.htm
My email address is whitelisted. *All* email sent to it
goes directly to the trash unless you add NOSPAM in the
Subject: line with other stuff. alondra101 <at> hotmail.com
Don't be ripped off by the big book dealers. Go to the URL
that will give you a choice and save you money(up to half).
http://www.everybookstore.com You'll be glad you did!
Just when you thought you had all this figured out, the gov't
changed it: http://physics.nist.gov/cuu/Units/binary.html
@@t@h@e@@a@f@f@l@u@e@n@t@@m@e@e@t@@t@h@e@@E@f@f@l@u@e@n@t@@

 

[CBFalconer]

Kevin Aylward wrote:

CBFalconer wrote:
Jim Thompson wrote:
andras_tantos@tantos.yahoo.com> wrote:

... snip ...

Have you done it with discrete components? I wouldn't be
surprised than. As other's pointed out they've used similar
techniques and reaching 23 bit precision so it seems it can
be done.

Seems to me that 23 bit performance would require incredible
matching accuracy.

You don't depend on matching, just stability. You convert the
value into a time interval and measure that.

Oh?... Since the input would go through 22 gain of two stages, what
output offset would you expect from from the first stage input offset?

What gain - what stages? The fundamental analog method of
converting a value into a time is to charge a capacitor and
discharge it with a constant current. Measure the time between
starting the discharge and reaching zero via a single comparator
gating a clock into a counter. You trade conversion time for
accuracy. The primary problems are non-linearities, capacitor
hysterisis, odd-even count effects creating differential
non-linearity, etc.

--
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

 

[Jim Thompson]

On Fri, 15 Aug 2003 12:00:26 GMT, CBFalconer <cbfalconer@yahoo.com>
wrote:

Kevin Aylward wrote:
[snip]
Oh?... Since the input would go through 22 gain of two stages, what
output offset would you expect from from the first stage input offset?

What gain - what stages? The fundamental analog method of
converting a value into a time is to charge a capacitor and
discharge it with a constant current. Measure the time between
starting the discharge and reaching zero via a single comparator
gating a clock into a counter. You trade conversion time for
accuracy. The primary problems are non-linearities, capacitor
hysterisis, odd-even count effects creating differential
non-linearity, etc.

Dual-slope (and its variants) are *slow*, which the OP claims to have
licked.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| Jim-T@analog_innovations.com Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

For proper E-mail replies SWAP "-" and "_"

I love to cook with wine. Sometimes I even put it in the food.