Maker Pro
Maker Pro

That Notorious Field-Strength Meter: the pictures at long last

P

Paul Burridge

Jan 1, 1970
0
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.
Just a note on the trace side: the black interconnecting wires are all
the ground points.
Follow these links and let me know what you think of my constructional
abilities...
http://freespace.virgin.net/orion.osiris/trace_side.gif
http://www.burridge8333.fsbusiness.co.uk/top_side.gif

The schematic to which this circuit relates can be found at this page:
http://www.burridge8333.fsbusiness.co.uk/fsm3.gif
 
B

Bill Garber

Jan 1, 1970
0
: Hello chaps,
:
: After 24 hours of struggling to get these pictures uploaded, I
believe
: I've finally succeeded although I've had to use two different
sites,
: one for each picture(!)
: I can imagine this will probably be a good exercise in "how not
to
: layout a board" and "how many possible causes of parasitic
oscillation
: can you spot here?" but nevertheless, the public needs to see
it.
: Just a note on the trace side: the black interconnecting wires
are all
: the ground points.
: Follow these links and let me know what you think of my
constructional
: abilities...
: http://freespace.virgin.net/orion.osiris/trace_side.gif
: http://www.burridge8333.fsbusiness.co.uk/top_side.gif
:
: The schematic to which this circuit relates can be found at
this page:
: http://www.burridge8333.fsbusiness.co.uk/fsm3.gif

I'd say that's an ugly board there.
Looks like it was done before they
made rulers.

Bill @ GarberStreet Enterprizez };-)
Web Site - http://garberstreet.netfirms.com
Email - [email protected]
Remove - SPAM and X to contact me
 
F

Fred Bloggs

Jan 1, 1970
0
Bill said:
I'd say that's an ugly board there.
Looks like it was done before they
made rulers.

And would you look at the way the damned fool juxtaposed those
inductors, and square feet of circuit loops to pick up anything in the
air- no antenna needed!
 
J

John

Jan 1, 1970
0
About the neatest way I've seen of doing it without etching a pcb is with
the use of an "island pad cutter" http://www.gqrp.com/ click on Club Sales
and look a little way down the page.

I have no affiliation with the seller.
 
J

John Fields

Jan 1, 1970
0
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.

---
"Needs" to see it? Why?
---
Just a note on the trace side: the black interconnecting wires are all
the ground points.
Follow these links and let me know what you think of my constructional
abilities...
http://freespace.virgin.net/orion.osiris/trace_side.gif
http://www.burridge8333.fsbusiness.co.uk/top_side.gif

---
Horrible, simply horrible.

Treating the grounds with care is arguably the single most important
part of doing a PCB layout, and yet you seem to have laid them out as an
afterthought.

Additionally, there is _no_ reason why the components should be
scattered about on the board with so much distance between them. Doing
it that way invites disaster because you're connecting everything with
antennas, and it seems disaster took the invitation.
 
A

Active8

Jan 1, 1970
0
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.
Just a note on the trace side: the black interconnecting wires are all
the ground points.
Follow these links and let me know what you think of my constructional
abilities...

You do know how to leave yourself open for abuse, don't you :)
Don't ask for personal comments, ask about the circuit.

I'm screwed over here trying to get single sided copper clad (none
at radio shack) without eating a shipping charge, and there you are
wasting that which I need, cock-knocker ;)

I wouldn't have a helter-skelter layout like this. Eagle, if you
have nothing better (not that it's bad), will help you line things
up better, eliminate jumpers, move thing together, etc. Even if you
don't have a LASER printer, you can use a printout to mark your
holes and connect the dots with your Sharpie. Please don't tell me
you buy those high dollar resist-ink pens. You'll hurt my ears.

+/- power rails on the top and a nice ground plane on both sides
wouldn't hurt. Keep the signal path down the middle of the solder
side.
Uh. two torroids gang banging each other *might* have less coupling
than those 2 coils. At least they're not end to end. If you tried a
shield between them to no avail, maybe they still contribute a
little to the prob. Can't put them perpendicular to each other,
though. Didn't you get cans for them?

I think the input stages are farthest away from the power supply
leads so I don't think RF currents are coupling back to the input
but who knows what might happen in the other direction. The track
look pretty beefy, so eh?

Think of your power rails as a possible a high impedance path
(inductor) at 40MHz where a potential diff can occur. You have a
scope, so you can lift component leads an shoot the thing from
input to output.

Long component leads, not good. There have been all kinds of HF
circuits build dead-bug style that *do* work, but I figure every
little advantage you give yourself helps.

I don't like those cross-country ground runs.

I see your meter out comes off the side next door to the coils. I
can't really tell at quick glance but maybe you've created a
current loop that links to part of the input net.

Slap some parasitic inductors and mutual inductance into your sim
and see if it oscillates - no! I was only kidding. Don't waste your
time.

You could Muntz it. Selectively remove caps and see if anything
changes.
 
H

Helmut Sennewald

Jan 1, 1970
0
Paul Burridge said:
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.
Just a note on the trace side: the black interconnecting wires are all
the ground points.

Hello Paul,
forget everything about star-grounding when it comes to high fequency.
Every of your black star-ground wires has a lot of "bad" inductance.
I estimate 20nH to 50nH. Put that into your simulation.

Here is a good example for prototypes:
http://diz.faithweb.com/cgi-bin/i/minipig/minipig.jpg
It is thousand times better and it needs no drill hole.

A complete ground plane is the best what you can spend a HF-circuit.
If you spend a double sided PCB, then you could wire the parts on the
other side, but still leave one side complete with copper for ground.

Best Regards,
Helmut
 
P

Paul Burridge

Jan 1, 1970
0
Horrible, simply horrible.

Not a pretty sight, I freely admit...
Treating the grounds with care is arguably the single most important
part of doing a PCB layout, and yet you seem to have laid them out as an
afterthought.

It does *look* that way, I agree. However, I was originally going to
use double-sided board with one side for ground and use vias for all
the ground points. But I couldn't find any DS board so did it with SS
and wires. I wonder if it really makes that much difference. I mean, a
lot of people talk of 40Mhz here in disparaging terms as if it's near
enough DC.
Additionally, there is _no_ reason why the components should be
scattered about on the board with so much distance between them. Doing
it that way invites disaster because you're connecting everything with
antennas, and it seems disaster took the invitation.

Interesting way of looking at it. I was hoping the seperation would
have the opposite effect.
 
R

Rich Webb

Jan 1, 1970
0
Follow these links and let me know what you think of my constructional
abilities...
[snip...snip...]

Good lord, man! There's a ton of inexpensive (and even free; see
http://pcb.sourceforge.net/) layout software. With moderate trace widths
and spacing, the "iron-on" technique (and a resist pen for touch-up)
works pretty well for one-off boards and can (reportedly) use plain
coated ink jet paper (printed on a laser printer).
 
J

Jim Thompson

Jan 1, 1970
0
On Sat, 17 Jan 2004 19:13:05 +0100, "Helmut Sennewald"

[snip]
Hello Paul,
forget everything about star-grounding when it comes to high fequency.
Every of your black star-ground wires has a lot of "bad" inductance.
I estimate 20nH to 50nH. Put that into your simulation.

Here is a good example for prototypes:
http://diz.faithweb.com/cgi-bin/i/minipig/minipig.jpg
It is thousand times better and it needs no drill hole.

A complete ground plane is the best what you can spend a HF-circuit.
If you spend a double sided PCB, then you could wire the parts on the
other side, but still leave one side complete with copper for ground.

Best Regards,
Helmut

Now that's *my kind* of breadboard ;-)

...Jim Thompson
 
J

John Fields

Jan 1, 1970
0
Not a pretty sight, I freely admit...


It does *look* that way, I agree. However, I was originally going to
use double-sided board with one side for ground and use vias for all
the ground points. But I couldn't find any DS board so did it with SS
and wires. I wonder if it really makes that much difference.

---
It makes a _huge_ difference. If you've only got single-sided board, a
very effective technique is to consider the entire copper surface a
ground plane and to carve from that the traces you need, leaving
everything else behind. That way, what you don't need for wiring will
be, by default, the ground plane. Also, the positive rail and _all_ the
power wiring should be as wide as you can stand and generously bypassed
to the ground plane with low-inductance caps.
---
I mean, a
lot of people talk of 40Mhz here in disparaging terms as if it's near
enough DC.

---
Be that as it may, it isn't. Even a board populated with active devices
intended for DC only operation can easily break into oscillation if due
care isn't given to the layout.
---
Interesting way of looking at it. I was hoping the seperation would
have the opposite effect.

---
It doesn't. Consider what happens if you have a trace with AC on it
which is, say, 3" long compared to another trace which is an inch long
with the same signal on it. Roughly speaking, the three inch long trace
will have a magnetic field building up and collapsing about it which is
three times longer than that which would be around the one inch long
trace, so the field around the long trace is going to be much more
likely to cause trouble than the short trace just because the long field
has a chance to cut through (and induce voltages in) more adjacent
conductors than the short field.
 
L

Leon Heller

Jan 1, 1970
0
Paul said:
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.
Just a note on the trace side: the black interconnecting wires are all
the ground points.
Follow these links and let me know what you think of my constructional
abilities...
http://freespace.virgin.net/orion.osiris/trace_side.gif
http://www.burridge8333.fsbusiness.co.uk/top_side.gif

The schematic to which this circuit relates can be found at this page:
http://www.burridge8333.fsbusiness.co.uk/fsm3.gif

There's no way that board will work at 40 MHz!

Leon
 
L

Leon Heller

Jan 1, 1970
0
Paul said:
Hello chaps,

After 24 hours of struggling to get these pictures uploaded, I believe
I've finally succeeded although I've had to use two different sites,
one for each picture(!)
I can imagine this will probably be a good exercise in "how not to
layout a board" and "how many possible causes of parasitic oscillation
can you spot here?" but nevertheless, the public needs to see it.
Just a note on the trace side: the black interconnecting wires are all
the ground points.
Follow these links and let me know what you think of my constructional
abilities...
http://freespace.virgin.net/orion.osiris/trace_side.gif
http://www.burridge8333.fsbusiness.co.uk/top_side.gif

The schematic to which this circuit relates can be found at this page:
http://www.burridge8333.fsbusiness.co.uk/fsm3.gif

Have alook at this design of mine if you want to see how to build an RF
circuit on a PCB:

http://www.geocities.com/leon_heller/dc_rx.html

I don't claim that it is perfect, but it works and looks quite neat.

Leon
 
J

Jim Thompson

Jan 1, 1970
0
About the neatest way I've seen of doing it without etching a pcb is with
the use of an "island pad cutter" http://www.gqrp.com/ click on Club Sales
and look a little way down the page.

I have no affiliation with the seller.

NEATO! But clear as mud as to how to obtain in the US. Is there a US
source that doesn't require sending "bank notes" in the mail?

...Jim Thompson
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Jim Thompson
4ax.com>) about 'That Notorious Field-Strength Meter: the pictures at
long last', on Sat, 17 Jan 2004:
NEATO! But clear as mud as to how to obtain in the US. Is there a US
source that doesn't require sending "bank notes" in the mail?

These cutters are known engineering tools, but are not widely used. A
GOOD engineering tool supplier should have them. If not, a local
engineering company can easily make them for you.

One problem is that the names of tools vary widely. My local one didn't
know what I meant by 'spot-face cutter' and, no, it's not an acne
remedy. It cuts circular recesses in thick panels so that you can mount
spindle controls with short bushes.

Possibly, the thing to do is to simulate a pad on a board with a Dremel
drill and router and take that to show what you want to do.
 
P

Paul Burridge

Jan 1, 1970
0
It makes a _huge_ difference. If you've only got single-sided board, a
very effective technique is to consider the entire copper surface a
ground plane and to carve from that the traces you need, leaving
everything else behind. That way, what you don't need for wiring will
be, by default, the ground plane. Also, the positive rail and _all_ the
power wiring should be as wide as you can stand and generously bypassed
to the ground plane with low-inductance caps.

Thanks, John. That *does* make a lot of sense.
Be that as it may, it isn't. Even a board populated with active devices
intended for DC only operation can easily break into oscillation if due
care isn't given to the layout.

Cripes! Noted.
---
It doesn't. Consider what happens if you have a trace with AC on it
which is, say, 3" long compared to another trace which is an inch long
with the same signal on it. Roughly speaking, the three inch long trace
will have a magnetic field building up and collapsing about it which is
three times longer than that which would be around the one inch long
trace, so the field around the long trace is going to be much more
likely to cause trouble than the short trace just because the long field
has a chance to cut through (and induce voltages in) more adjacent
conductors than the short field.

I assume it's also got a lot to do with the frequency involved, but
the general principle is certainly interesting.
 
P

Paul Burridge

Jan 1, 1970
0
About the neatest way I've seen of doing it without etching a pcb is with
the use of an "island pad cutter" http://www.gqrp.com/ click on Club Sales
and look a little way down the page.

Interesting. But isn't this way of cutting traces supposedly
carcinogenic?
 
P

Paul Burridge

Jan 1, 1970
0
Have alook at this design of mine if you want to see how to build an RF
circuit on a PCB:

http://www.geocities.com/leon_heller/dc_rx.html

I don't claim that it is perfect, but it works and looks quite neat.

Looks very 'microwavy' with all that copper GP-dominated appearance.
Would I be right in supposing that this 'minimalist-trace' type of
layout is the preferred method of producing boards for just about
*any* frequency - just to be on the safe side? I mean, it can't do any
harm at DC or LF can it?
 
L

Leon Heller

Jan 1, 1970
0
Paul said:
Looks very 'microwavy' with all that copper GP-dominated appearance.
Would I be right in supposing that this 'minimalist-trace' type of
layout is the preferred method of producing boards for just about
*any* frequency - just to be on the safe side? I mean, it can't do any
harm at DC or LF can it?


A lot of that board is at AF. You don't really need a ground plane like
that at DC or LF, unless you have RF floating around in the same system.

It actually makes the board easier to design, you don't have to worry
about routing the grounds.

I've built a high-gain 9 MHz IF amplifier using three MOSFET amplifier
stages using these techniques, it worked first time and was completely
stable.

Leon
 
R

Rene Tschaggelar

Jan 1, 1970
0
Paul said:
Hello chaps,

[ snip ]
Follow these links and let me know what you think of my constructional
abilities...
http://freespace.virgin.net/orion.osiris/trace_side.gif
http://www.burridge8333.fsbusiness.co.uk/top_side.gif

The schematic to which this circuit relates can be found at this page:
http://www.burridge8333.fsbusiness.co.uk/fsm3.gif

Paul,
squeeze it and add a GND plane. Examples :

http://www.ibrtses.com/projects/diodenamp.jpg
http://www.ibrtses.com/products/SyncRF350MHz.jpg

Rene
 
Top