That legendary Field-Strength Meter: the latest...

[Paul Burridge]

It worked first time!

Having implemented most of the suggestions given here on this group,
the bloody thing is as stable as they come and performs pretty much
exactly as advertised in the LT Spice simulation. Amazing...
What was done in the second attempt:
New PCB with large ground area and increased plus and minus rail
areas;
Mylars dumped in favour of ceramics for all RF parts.
More bypass caps added in areas where they make a difference
Component leads shortened as much as possible
Coils moved much further apart from one another
Better stabilisation of power supply

I have to admit to being sceptical that just doing these few things
would made so much difference, but I guess there's no question of it
now.
Just for the record, 1mV P-P RF in at 40Mhz gives a mid-range meter
reading of S5.
Many thanks once again to everyone who assisted.

 

[Tweetldee]

It worked first time!

Having implemented most of the suggestions given here on this group,
the bloody thing is as stable as they come and performs pretty much
exactly as advertised in the LT Spice simulation. Amazing...
What was done in the second attempt:
New PCB with large ground area and increased plus and minus rail
areas;
Mylars dumped in favour of ceramics for all RF parts.
More bypass caps added in areas where they make a difference
Component leads shortened as much as possible
Coils moved much further apart from one another
Better stabilisation of power supply

I have to admit to being sceptical that just doing these few things
would made so much difference, but I guess there's no question of it
now.
Just for the record, 1mV P-P RF in at 40Mhz gives a mid-range meter
reading of S5.
Many thanks once again to everyone who assisted.

Paul, now we come full circle to your original post.... what do we think
about your abilities as a designer. I think you've taken all the necessary
hits about remaining humble until you've really proven your grit, and using
your head as your primary design tool instead of Spice.
You've seen that a simulator is a great tool, but it should only be used to
check your efforts after you've done a proper design.. using your pencil,
paper, calculator, whatever is handy. Leave the simulator in icon fom
until you're satisfied that you've done your part with your education,
training and experience.
Now that you have a working project under your belt, the question is: what
do YOU think of your abilities as a designer? If you didn't have that Spice
simulator (or whichever of its cousins you have), would you have been able
to design your circuit on paper, and build it with good expectations that it
would work?
--
Tweetldee
dgmason44 at comcast dot net (Just subsitute the appropriate characters in
the address)

Never take a laxative and a sleeping pill at the same time!!

 

[Paul Burridge]

Paul, now we come full circle to your original post.... what do we think
about your abilities as a designer.

Erm, yeah, well don't worry about that. Please!

I think you've taken all the necessary
hits about remaining humble until you've really proven your grit, and using
your head as your primary design tool instead of Spice.

Yes, except in this particular case of course, it was someone else's
design...

You've seen that a simulator is a great tool, but it should only be used to
check your efforts after you've done a proper design.. using your pencil,
paper, calculator, whatever is handy. Leave the simulator in icon fom
until you're satisfied that you've done your part with your education,
training and experience.
Now that you have a working project under your belt, the question is: what
do YOU think of your abilities as a designer?

Not terribly impressive, I'm afraid. There's a *lot* more to the
design side of the subject than I ever imagined. Up until relatively
recently, I'd never really thought there was that much difference in
the abilities of designers and competent service engineers (when I say
"service engineers" I mean the sort that *do* actually have a
thoroughly good grasp of electronics, not the type that just get
called out to replace boards). It really instills in one a high regard
for the profession of electronic design.

If you didn't have that Spice
simulator (or whichever of its cousins you have), would you have been able
to design your circuit on paper, and build it with good expectations that it
would work?

All I can say is that you're right, IME. It's very tempting to design
a circuit in Spice. I've tried it many times and pulled component
values out of the air just because they "looked about right" and in
every case it's ended in failure. Then I get the calculator, pencil
and paper out and analyse the voltages and currents by first or second
approximation and invariably, the "out-of-the-air" values are *miles*
out. So yes, Spice is great when used as it should be but it doesn't
design things for you. Pencil paper and calculator rules.

 

[Leon Heller]

Erm, yeah, well don't worry about that. Please!




Yes, except in this particular case of course, it was someone else's
design...




Not terribly impressive, I'm afraid. There's a *lot* more to the
design side of the subject than I ever imagined. Up until relatively
recently, I'd never really thought there was that much difference in
the abilities of designers and competent service engineers (when I say
"service engineers" I mean the sort that *do* actually have a
thoroughly good grasp of electronics, not the type that just get
called out to replace boards). It really instills in one a high regard
for the profession of electronic design.




All I can say is that you're right, IME. It's very tempting to design
a circuit in Spice. I've tried it many times and pulled component
values out of the air just because they "looked about right" and in
every case it's ended in failure. Then I get the calculator, pencil
and paper out and analyse the voltages and currents by first or second
approximation and invariably, the "out-of-the-air" values are *miles*
out. So yes, Spice is great when used as it should be but it doesn't
design things for you. Pencil paper and calculator rules.

You can make things easier by using a speadsheet to replace pencil,
paper and calulator. It's better in some ways because you can play with
different values and see what happens. Then check the design with SPICE.
I've got several spreadsheets I use for doing things like designing bias
circuits, inductors, etc.

Leon

 

[James Meyer]

You can make things easier by using a speadsheet to replace pencil,
paper and calulator. It's better in some ways because you can play with
different values and see what happens. Then check the design with SPICE.
I've got several spreadsheets I use for doing things like designing bias
circuits, inductors, etc.

Leon

"Like a spreadsheet" is exactly the way I use spice circuit simulation.

I'm very comfortable with pictures. Words, less so. And math, only
when there's no other way. Starting with a picture, a schematic drawing, suits
my way of thinking best. I think I must have an editic memory for pictures.

I can tell when a circuit "looks right" in schematic form and then use
spice to do the grunt work of telling me the component values necessary to make
it "work right".

Jim

 

[Jim Thompson]

[snip]

I'm very comfortable with pictures. Words, less so. And math, only
when there's no other way. Starting with a picture, a schematic drawing, suits
my way of thinking best. I think I must have an editic memory for pictures.

I can tell when a circuit "looks right" in schematic form and then use
spice to do the grunt work of telling me the component values necessary to make
it "work right".

Jim

Same here. I do believe I think in schematics (as my second language
;-)

...Jim Thompson

 

[Paul Burridge]

You can make things easier by using a speadsheet to replace pencil,
paper and calulator. It's better in some ways because you can play with
different values and see what happens. Then check the design with SPICE.
I've got several spreadsheets I use for doing things like designing bias
circuits, inductors, etc.

I'm sure you're right, but I'm equally sure I'm clueless about
spreadsheets. What can a spreadsheet do that a programmable calculator
can't? AFAIK, they're just an application that's useful for
book-keeping/accountancy. But I'm sure I'm about to be put straight...

 

[Spehro Pefhany]

I'm sure you're right, but I'm equally sure I'm clueless about
spreadsheets. What can a spreadsheet do that a programmable calculator
can't? AFAIK, they're just an application that's useful for
book-keeping/accountancy. But I'm sure I'm about to be put straight...

Spreadsheets are incredibly useful in engineering. If you need
heavy-duty math you can use something like MATLAB, but spreadsheets
are great for all kinds of mundane (calculate the pad positions in
Imperial units for a metric part rotated by 33°) calculations or for
more complex stuff such as solving nonlinear equations numerically. In
theory you can do much of the same thing with a sophisticated
calculator, but the spreadsheet allows you to make changes and observe
the results without re-entering everything and without writing a
single line of code.

Best regards,
Spehro Pefhany

 

[James Meyer]

[snip]

I'm very comfortable with pictures. Words, less so. And math, only
when there's no other way. Starting with a picture, a schematic drawing, suits
my way of thinking best. I think I must have an editic memory for pictures.

I can tell when a circuit "looks right" in schematic form and then use
spice to do the grunt work of telling me the component values necessary to make
it "work right".

Jim

Same here. I do believe I think in schematics (as my second language
;-)

...Jim Thompson

Most good engineers score above average on spacial perception problems
in standard tests. Those questions like "From the following shapes, pick two
that can be fitted together to make a rectangle." That type of skill is
generally more extant for males in general and is one reason (IMHO) why there
are more male than female engineers.

Jim

 

[John Woodgate]

I read in sci.electronics.design that Spehro Pefhany <[email protected]
DOTyou.knowwhat> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sat, 24
Jan 2004:

the
spreadsheet allows you to make changes and observe the results without
re-entering everything and without writing a single line of code.

You have to write lines of code in a horrible format (IMHO) to define
the data to be used for 'charts'. In Excel, anyway. And the examples
given in magazine pages by resident 'experts' of ways of doing unusual
sorting operations etc. also require what is undoubtedly code writing.

 

[Terry Pinnell]

I read in sci.electronics.design that Spehro Pefhany <[email protected]
DOTyou.knowwhat> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sat, 24
Jan 2004:

You have to write lines of code in a horrible format (IMHO) to define
the data to be used for 'charts'. In Excel, anyway. And the examples
given in magazine pages by resident 'experts' of ways of doing unusual
sorting operations etc. also require what is undoubtedly code writing.

No, that's rarely necessary. Using the Chart Wizard, from the data in
http://www.terrypin.dial.pipex.com/Images/DataJW.gif
to the basic chart in
http://www.terrypin.dial.pipex.com/Images/ChartJW.gif
took only a few seconds.

 

[John Woodgate]

I read in sci.electronics.design that Terry Pinnell <[email protected]
..pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:

No, that's rarely necessary. Using the Chart Wizard, from the data in
http://www.terrypin.dial.pipex.com/Images/DataJW.gif
to the basic chart in
http://www.terrypin.dial.pipex.com/Images/ChartJW.gif
took only a few seconds.

I can't tell exactly what you did, but it looks to me as though you have
plotted column A and column B against one of Excel's built-in 'time
scales', 12 months I suppose.

Normally, in engineering, one wants to plot columns B, C, D,... against
column A. You have to choose 'scatter' as the chart type and then enter
the 'source data' in forms like:

=Sheet1!$A$5:$A$28

....and woe betide you if you get it wrong!

If that isn't code, it must be something else. (;-)

 

[Terry Pinnell]

I read in sci.electronics.design that Terry Pinnell <[email protected]
.pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:
I can't tell exactly what you did, but it looks to me as though you have
plotted column A and column B against one of Excel's built-in 'time
scales', 12 months I suppose.

No! X = N, and Y = N squared, as shown. I just happened to choose a
max N of 12.

Normally, in engineering, one wants to plot columns B, C, D,... against
column A. You have to choose 'scatter' as the chart type and then enter
the 'source data' in forms like:

=Sheet1!$A$5:$A$28

...and woe betide you if you get it wrong!

If that isn't code, it must be something else. (;-)

Still don't see the problem. Do you mean like this second example?
http://www.terrypin.dial.pipex.com/Images/Chart2JW.gif
which again involved no code.

 

[John Woodgate]

I read in sci.electronics.design that Terry Pinnell <[email protected]
..pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:

No! X = N, and Y = N squared, as shown. I just happened to choose a max
N of 12.

No. You look again. You have both N and N^2 graphed against the '1 to
12' horizontal axis. 'Line' graphs can't use data as the x-axis See the
Help.

[snip]

Still don't see the problem. Do you mean like this second example?
http://www.terrypin.dial.pipex.com/Images/Chart2JW.gif
which again involved no code.

By sending GIFs, you prevent me from seeing how you arrived at your
results. You could have sent the Excel files; they are unlikely to be
too large.

 

[Active8]

I read in sci.electronics.design that Terry Pinnell <[email protected]
.pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:
I can't tell exactly what you did, but it looks to me as though you have
plotted column A and column B against one of Excel's built-in 'time
scales', 12 months I suppose.

Normally, in engineering, one wants to plot columns B, C, D,... against
column A. You have to choose 'scatter' as the chart type and then enter
the 'source data' in forms like:

=Sheet1!$A$5:$A$28

...and woe betide you if you get it wrong!

If that isn't code, it must be something else. (;-)

More like scirpt, no? The VB modules you can write for Excell
*could* be called code if you consider VB a *real* language.

But you're all right. Spreadsheets *are* a handy way to do
repetetive calculations. I like those features like dragging a cell
handle from a cell containing say 1 (or 2 cells with 1 and 5) and
getting an auto fill of incremented numbers or 1, 5, 9, ...

 

[Keith R. Williams]

I read in sci.electronics.design that Terry Pinnell <[email protected]
.pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:
I can't tell exactly what you did, but it looks to me as though you have
plotted column A and column B against one of Excel's built-in 'time
scales', 12 months I suppose.

Normally, in engineering, one wants to plot columns B, C, D,... against
column A. You have to choose 'scatter' as the chart type and then enter
the 'source data' in forms like:

=Sheet1!$A$5:$A$28

...and woe betide you if you get it wrong!

If that isn't code, it must be something else. (;-)

I'll agree that it's code, and perhaps ugly code at that!
However, spreadsheets *are* useful in all sorts of strange ways.
....not all have much to do with formulas, as you are used to,
anyway.

I used Excel to map pins on an FPGA (Xilinx in a FG680 package).
I first started with the outline of the balls, then on another
sheet added the fixed supply pins. ...on another the design
specific supplies and I/O specifications/restrictions. On still
other sheets I put in the generic pins and then the specific
signal names. As in most hierarchical designs the errors were
minimal, and *all* caught well before I went to the board.

In another case, I've been given a task to isolate various "black
boxes" within a rather complicated design (58M transistors, give
or take 2-3x, or so we're told) for circuit analysis. Since the
design is hierarchical, it's not easy to find the "real" signal
names of the lower-level elements. Spreadsheets are a huge help
in tracing the signals back to the "owner" and formatting the
results for the extraction tools *and* documentation (same data
sorta, but different format).

Spreadsheets are simply another tool. ...and a very valuable
one. If an engineer doesn't see the value of a spreadsheet, it's
not the spreadsheet's problem. ;-)

Finance is the only one that can use spreadsheets? Pshaw!

 

[YD]

I'm sure you're right, but I'm equally sure I'm clueless about
spreadsheets. What can a spreadsheet do that a programmable calculator
can't? AFAIK, they're just an application that's useful for
book-keeping/accountancy. But I'm sure I'm about to be put straight...

Much more than that. For example, you might set it up for a bias
calculator. An input column for Vcc, beta, Ic, desired gain, Ve, etc.
Another column for calculated values, Rb1, Rb2, Re and so on. Just
input your values and see all the outputs at the same time. If they
seem OK try them in spice, if not try other values. Sometimes you'll
see where a trade-off or compromise is needed. Much easier than
puttering about with repetitve inputs in a calculator.

- YD.

 

[Terry Pinnell]

I read in sci.electronics.design that Terry Pinnell <[email protected]
.pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Sun, 25
Jan 2004:

No! X = N, and Y = N squared, as shown. I just happened to choose a max
N of 12.

No. You look again. You have both N and N^2 graphed against the '1 to
12' horizontal axis. 'Line' graphs can't use data as the x-axis See the
Help.

[snip]

Still don't see the problem. Do you mean like this second example?
http://www.terrypin.dial.pipex.com/Images/Chart2JW.gif
which again involved no code.

By sending GIFs, you prevent me from seeing how you arrived at your
results. You could have sent the Excel files; they are unlikely to be
too large.

I thought you'd be able to reconstruct those examples quickly from the
steps I included.

I don't follow what you mean by " 'Line' graphs can't use data as the
x-axis." Here's the same example again, using 'x' instead of 'N'
http://www.terrypin.dial.pipex.com/Images/Chart3JW.gif

'x' simply represents one independent variable. It could equally well
be 't' for time. FWIW I've posted the Excel file too
http://www.terrypin.dial.pipex.com/Images/JWExample3.xls

But I'm by now pretty sure we're at cross purposes. Of course, the
type of chart you have in mind will be more complex than my
deliberately simple examples. But I'd hoped the point I made (that you
can use the Chart Wizard to avoid manual coding) would still be
applicable. Why don't you post an example of what you mean?

 

[John Woodgate]

I read in sci.electronics.design that Terry Pinnell <[email protected]
..pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Mon, 26
Jan 2004:

I don't follow what you mean by " 'Line' graphs can't use data as the
x-axis." Here's the same example again, using 'x' instead of 'N'
http://www.terrypin.dial.pipex.com/Images/Chart3JW.gif

Your 'line' graph plots x, a, b and d against the 'categories' 1 to 9. I
can't find a definition of 'category'; I think even the Help writers
balked at it. But in the Troubleshoot part of Help is:

"My chart shows the wrong values on the horizontal axis" (paraphrased,
because the title is a link and won't copy.)

"Make sure you didn't select a line chart. If your category axis
displays numbers like 1, 2, 3, 4 instead of the values you want, you may
have selected a line chart instead of an xy (scatter) chart when you
created the chart."

What this result in is that the values of a, b and d for x = 10 appear
on the chart in line with the 'category' number 9. Try using x values
from, say, 5 to 15 and you will find that the horizontal axis still goes
from 1 to 9, counting the *items*, not the x-values.

'x' simply represents one independent variable. It could equally well
be 't' for time. FWIW I've posted the Excel file too
http://www.terrypin.dial.pipex.com/Images/JWExample3.xls

That's what I'm looking at.

But I'm by now pretty sure we're at cross purposes. Of course, the
type of chart you have in mind will be more complex than my
deliberately simple examples.

Yes, I hope you now see the point about the line chart horizontal axis.
But the simplicity has hidden the second of the very points we are
discussing.

You can us the chart wizard if you want to plot all the columns. If you
don't, and I usually don't, the chart wizard either balks at being asked
to plot too many columns or gives me loads of stuff I don't want and I
have to delete it all. (BTW, I'm talking about Excel 97: later versions
may differ somewhat.) Sometimes I make a scatter chart (and there is an
option to add either straight or 'best fit' lines between the data
points) with just one curve on it and then want to add others. That
means writing the new data in the ! and $-ridden code in the Values box
of the Series pane of Chart->Source data.

But I'd hoped the point I made (that you
can use the Chart Wizard to avoid manual coding) would still be
applicable. Why don't you post an example of what you mean?

I didn't previously because I wasn't sure what the problem was, and thus
how to demonstrate it. I hope I don't need to now.

 

[Terry Pinnell]

I read in sci.electronics.design that Terry Pinnell <[email protected]
.pipexTHIS.com> wrote (in <[email protected]>)
about 'That legendary Field-Strength Meter: the latest...', on Mon, 26
Jan 2004:

Your 'line' graph plots x, a, b and d against the 'categories' 1 to 9. I
can't find a definition of 'category'; I think even the Help writers
balked at it. But in the Troubleshoot part of Help is:

"My chart shows the wrong values on the horizontal axis" (paraphrased,
because the title is a link and won't copy.)

"Make sure you didn't select a line chart. If your category axis
displays numbers like 1, 2, 3, 4 instead of the values you want, you may
have selected a line chart instead of an xy (scatter) chart when you
created the chart."

What this result in is that the values of a, b and d for x = 10 appear
on the chart in line with the 'category' number 9. Try using x values
from, say, 5 to 15 and you will find that the horizontal axis still goes
from 1 to 9, counting the *items*, not the x-values.

That's what I'm looking at.

Yes, I hope you now see the point about the line chart horizontal axis.
But the simplicity has hidden the second of the very points we are
discussing.

Indeed - line graphs are clearly inappropriate here as you say.

But I'm not clear what the problem is with using a Scatter chart?
Changing the range of x from 5 to 15 in that case gives a correct
graph, yes?

You can us the chart wizard if you want to plot all the columns. If you
don't, and I usually don't, the chart wizard either balks at being asked
to plot too many columns or gives me loads of stuff I don't want and I
have to delete it all. (BTW, I'm talking about Excel 97: later versions
may differ somewhat.) Sometimes I make a scatter chart (and there is an
option to add either straight or 'best fit' lines between the data
points) with just one curve on it and then want to add others. That
means writing the new data in the ! and $-ridden code in the Values box
of the Series pane of Chart->Source data.


I didn't previously because I wasn't sure what the problem was, and thus
how to demonstrate it. I hope I don't need to now.

Anyway, plainly using the wizard for your sort of data analysis still
means resorting to code. I haven't had that problem.

I have Excel 2000, but I doubt that its Chart Wizard is radically
cleverer than that in Excel '97.