Hello Everyone,

I'd like to build a signal generator that is not too professional but good
enough for general use i.e. that can produce sine, sawtooth and square
waveforms in 1-50kHz range or so with reasonable reliability. Do you happen
to have any schematics, suggestions, assembly tips etc.?

Many thanks in advance,
Murat

--
NEU : GMX Internet.FreeDSL
Ab sofort DSL-Tarif ohne Grundgeb???hr:

On Fri, 16 Apr 2004 10:42:46 +0200 (MEST), Murat Ceylan <hector@...> wrote:

Hello Everyone,

I'd like to build a signal generator that is not too professional but good
enough for general use i.e. that can produce sine, sawtooth and square
waveforms in 1-50kHz range or so with reasonable reliability. Do you happen
to have any schematics, suggestions, assembly tips etc.?

Many thanks in advance,
Murat

I have two max038 lying around which wait only for me making a schematic for a generator.
these are sine/triangle/square wave generators that work up to 20Mhz.
there is a cheaper version that works to 1Mhz which is sufficient for what you need.

But you can get the max038 for free as maxim samples so there is no need to cut corners.

which output amplitude do you need?

i plan to set up the second max038 as second independent generator but als selectable
to modulate the primary generator in frequency or amplitude.

You will want to look at the maxim datasheets and app. notes for the max038 and you
might want to search google with "frequency generator schematic".

ST

On Friday 16 April 2004 05:55 am, Stefan Trethan wrote:

On Fri, 16 Apr 2004 10:42:46 +0200 (MEST), Murat Ceylan

<hector@...> wrote:

Hello Everyone,

I'd like to build a signal generator that is not too professional but
good
enough for general use i.e. that can produce sine, sawtooth and square
waveforms in 1-50kHz range or so with reasonable reliability. Do you
happen
to have any schematics, suggestions, assembly tips etc.?

Many thanks in advance,
Murat

I have two max038 lying around which wait only for me making a schematic
for a generator.
these are sine/triangle/square wave generators that work up to 20Mhz.
there is a cheaper version that works to 1Mhz which is sufficient for what
you need.

Is this the same sort of part as the Intersil (?) 8038? I built a generator
around one of those some years back, it was fairly simple, and it's very
flexible.

I have two max038 lying around which wait only for me making a schematic
for a generator.
these are sine/triangle/square wave generators that work up to 20Mhz.
there is a cheaper version that works to 1Mhz which is sufficient for what
you need.

Is this the same sort of part as the Intersil (?) 8038? I built a generator
around one of those some years back, it was fairly simple, and it's very
flexible.

It's similar but not the same.
The intersil part is only 300kHz max.

I doubt you get the intersil part for free so why bother?

I dunno when i have time for designing the generator and making a pcb,
could be a long time...
I want to allow for modulation (although i most likely never ever need it ;-) ).

The biggest problem now is the output stage, i have not designed such fast circuits before,
and i don't know how exactly to build it. i guess a opamp would be easiest but i don't know
if i can use it at 20Mhz just as i would use it at audio frequency.
need to think about it, and just now i have no time..
I only ordered the part to fill up the 8 item list having in the back of my mind i need a generator.


ST

On Friday 16 April 2004 01:13 pm, Stefan Trethan wrote:

I have two max038 lying around which wait only for me making a schematic
for a generator.

<...>

Is this the same sort of part as the Intersil (?) 8038?

<...>

It's similar but not the same. The intersil part is only 300kHz max.

Well, since I was primarily only interested in audio, that wasn't any big
deal...

I doubt you get the intersil part for free so why bother?

Because this was a bunch of years ago, when the parts you refer to weren't
available, and I doubt that I could've gotten it for free anyhow. Not that
it was all that expensive, it was sold at Radio Shack for a while and
could't have been more than a couple of dollars.

I dunno when i have time for designing the generator and making a pcb,
could be a long time...

The design was fun, I used an op amp and provided for a variable DC offset in
there. Also a logic-level output for sending slower pulses into TTL. The
board was one of those plug-in boards (44-pin edge connector) sold for
prototyping work. I've been messing with this stuff for a *long* time and
STILL haven't seen the need to make boards for stuff, unless I was going to
do a lot of some particular circuit, and it hasn't worked out that way.

I want to allow for modulation (although i most likely never ever need it
;-) ).

I didn't see that as being useful in what I intended to be an item of bench
test equipment. For a musical instrument, yeah, but that would be a
different project then.

The biggest problem now is the output stage, i have not designed such fast
circuits before, and i don't know how exactly to build it. i guess a opamp
would be easiest but i don't know if i can use it at 20Mhz just as i would
use it at audio frequency. need to think about it, and just now i have no
time..

That would take a very fast and expensive op amp, to be sure. Why go up that
high in terms of frequency? When I get up there I usually want some fixed
frequency and will likely use one of those crystal oscillator modules I have
a pile of here.

I only ordered the part to fill up the 8 item list having in the back of
my mind i need a generator.

What are you going to use it for that you want to go all the way up to 20 MHz?

I doubt you get the intersil part for free so why bother?

Because this was a bunch of years ago, when the parts you refer to weren't
available, and I doubt that I could've gotten it for free anyhow. Not that
it was all that expensive, it was sold at Radio Shack for a while and
could't have been more than a couple of dollars.

sure, i only want to say the max038 costs some eur30 to 50 or so if you buy it,
and the interesil much less, but if you can get the maxim for nothing this is no
problem.

The design was fun, I used an op amp and provided for a variable DC offset in
there. Also a logic-level output for sending slower pulses into TTL. The
board was one of those plug-in boards (44-pin edge connector) sold for
prototyping work. I've been messing with this stuff for a *long* time and
STILL haven't seen the need to make boards for stuff, unless I was going to
do a lot of some particular circuit, and it hasn't worked out that way.

That only depends on how well your process is set up.
I tended to make circuits without prined boards in the past, but now it is not
really more time demanding to make a board, and it is really much more convenient.

also you can hardly use smd on those proto boards.
(And i can make boards cheaper than buying the proto boards)

If you are facing working with a rather unknown pcb software, making a layout,
sending it out and paying a lot, waiting weeks... i know this makes you use protoboards.
But if you are experienced with the layout software and have a good pcb setup you just
print it out, fuse it, peel it, etch it, drill it, ready. no more than an hour.

And you have the benefit that it looks really good.

I didn't see that as being useful in what I intended to be an item of bench
test equipment. For a musical instrument, yeah, but that would be a
different project then.

I want it as a test bench item.
I used the frequency sweep in the past to make quick and easy bode plots (amplitude).
Of course i can live without but if i can include it.. well..
most of the commercial generators have f sweep so it might be a good idea.

(I agree the amp. modulation is maybe less useful but i can imagine uses if i try hard)

That would take a very fast and expensive op amp, to be sure. Why go up that
high in terms of frequency? When I get up there I usually want some fixed
frequency and will likely use one of those crystal oscillator modules I have
a pile of here.

I don't have a pile of crystals just now. If it is possible without too much investment
i will try to preserve the whole frequency range. maybe i just add another direct
output from the 038 with 50Ohms, you won't need high amplitude there, or dc offset.
Seems like a good idea to do that, and much easier if i can keep the amp. slow.

What are you going to use it for that you want to go all the way up to 20 MHz?

I need a generator, i have none.. all i have is a 1Hz to 5kHz generator in a multimeter, TTL.
I'm on a tight budget so all is moving between homebrew and ebay....

You see i don't need the 20Mhz now, since i obviously don't know how to design 20Mhz circuits ;-).
But who knows what exactly the future will bring, things are definitely going faster and faster.

I'm thinking about adding a ttl out the the counter i am just programming, i have a fast
crystal in there already, a thumbwheel encoder and a 8 digit display so i reckon it is not
hard to allow for using the counter the other way around.
This will satisfy all need for high frequency clock signals, precise and TTL.
So if it works out i only need the generator for sine and higher amplitude things.

thanks for making me think about what makes sense to build and what not, sometimes i overshoot the target.

ST

--- In Electronics_101@..., "Murat Ceylan" <hector@g...>
wrote:

Hello Everyone,

I'd like to build a signal generator that is not too professional

but good

enough for general use i.e. that can produce sine, sawtooth and square
waveforms in 1-50kHz range or so with reasonable reliability. Do

you happen

to have any schematics, suggestions, assembly tips etc.?

Many thanks in advance,
Murat

--
NEU : GMX Internet.FreeDSL
Ab sofort DSL-Tarif ohne Grundgebühr:

I have a few XR2206 chips around. Dirt simple. The signal generator I
use the most is based on an XR2206. It's an ugly thing but it always
does what I want and can sweep over a couple of decades. It has AM and
FM inputs.

On Friday 16 April 2004 01:50 pm, Stefan Trethan wrote:

<...>

The design was fun, I used an op amp and provided for a variable DC
offset in there. Also a logic-level output for sending slower pulses into
TTL. The board was one of those plug-in boards (44-pin edge connector)
sold for prototyping work. I've been messing with this stuff for a *long*
time and STILL haven't seen the need to make boards for stuff, unless I
was going to do a lot of some particular circuit, and it hasn't worked
out that way.

That only depends on how well your process is set up. I tended to make
circuits without prined boards in the past, but now it is not really more
time demanding to make a board, and it is really much more convenient.

I don't tend to see it that way. And making a circuit-specific board is fine
if you have a design that's nailed down, and you know it's going to work as
it is. I was fiddling with things and changing things as I went along...

also you can hardly use smd on those proto boards.

I have no intention of doing anything with that stuff if I can help it. Sure,
the old parts are reportedly getting harder to find, I hear, but I have
plenty of parts to mess with, and while some of the newer stuff may be
exceedingly nifty, I don't see a _need_ for it. I'm not after minimum size,
and feel that overall the major thrust of SMD is convenience in manufacturing
more than anything else -- manufacturing in large quantity, which I'm not
doing.

(And i can make boards cheaper than buying the proto boards)

Perhaps. I don't use all that many of them, and in fact have found some in a
box that I've had for years, and never opened.

If you are facing working with a rather unknown pcb software, making a
layout, sending it out and paying a lot, waiting weeks... i know this makes
you use protoboards. But if you are experienced with the layout software and
have a good pcb setup you just print it out,

This assumes the use of a laser printer?

fuse it,

Are you talking about toner transfer here?

peel it,

And here?

etch it,

Yeah, not time-consuming if you happen to have everything already set up to
use, and don't have to go through all of that process...

drill it,

Another time-consuming aspect of it, unless you have a CNC setup or similar
(something that's on my list to do).

ready. no more than an hour.

I suspect that it would take me considerably longer than that.

And you have the benefit that it looks really good.

That's not a particularly relevant factor for me.

I didn't see that as being useful in what I intended to be an item of
bench test equipment. For a musical instrument, yeah, but that would be
a different project then.

I want it as a test bench item. I used the frequency sweep in the past to
make quick and easy bode plots (amplitude).

Ok, I can see where that would be useful. Maybe.

Of course i can live without but if i can include it.. well..
most of the commercial generators have f sweep so it might be a good idea.

(I agree the amp. modulation is maybe less useful but i can imagine uses
if i try hard)

That would take a very fast and expensive op amp, to be sure. Why go
up that high in terms of frequency? When I get up there I usually want
some fixed frequency and will likely use one of those crystal oscillator
modules I have a pile of here.

I don't have a pile of crystals just now.

I've been salvaging / scrapping stuff for a lot of years. :-)

If it is possible without too much investment i will try to preserve the
whole frequency range. maybe i just add another direct output from the 038
with 50Ohms, you won't need high amplitude there, or dc offset.

What part would you use to do that?

Seems like a good idea to do that, and much easier if i can keep the amp.
slow.

Yes.

What are you going to use it for that you want to go all the way up to
20 MHz?

I need a generator, i have none.. all i have is a 1Hz to 5kHz generator in
a multimeter, TTL.

In a meter? First I've heard of that.

I'm on a tight budget so all is moving between homebrew and ebay....

I've also had little to spend on this sort of thing most of the time.

You see i don't need the 20Mhz now, since i obviously don't know how to
design 20Mhz circuits ;-).
But who knows what exactly the future will bring, things are definitely
going faster and faster.

Yeah, well, just because computer speeds are getting up into the UHF range
doesn't mean that I'm going to take my hobby endeavors there, I'm happy
enough to work with audio...

I'm thinking about adding a ttl out the the counter i am just programming,
i have a fast crystal in there already, a thumbwheel encoder and a 8 digit
display so i reckon it is not hard to allow for using the counter the other
way around. This will satisfy all need for high frequency clock signals,
precise and TTL.

Tell me more about this?

So if it works out i only need the generator for sine and higher amplitude
things.

That's a key, figure out what you want to use it for first! :-)

thanks for making me think about what makes sense to build and what not,
sometimes i overshoot the target.

It's easy to do when you're sitting there playing with ideas, and lose track
of what you would or wouldn't do with most of those features.

I have a few XR2206 chips around. Dirt simple. The signal generator I
use the most is based on an XR2206. It's an ugly thing but it always
does what I want and can sweep over a couple of decades. It has AM and
FM inputs.

this was the other chip i meant... i think this one is good to 1Mhz?

Do you have a schematic of the thing and want to provide it?


ST

I have no intention of doing anything with that stuff if I can help it. Sure,
the old parts are reportedly getting harder to find, I hear, but I have
plenty of parts to mess with, and while some of the newer stuff may be
exceedingly nifty, I don't see a _need_ for it. I'm not after minimum size,
and feel that overall the major thrust of SMD is convenience in manufacturing
more than anything else -- manufacturing in large quantity, which I'm not
doing.

you may be right there, one advantage is manufacturing.. but not only in large quantity,
also in small quantity it is nice if you don't need to drill holes, bend wires, cut wires.

I find it faster to use and throughhole parts are definitely going to be hard to find in the
near future. (and smd is cheaper too.. but i was told that this is no "how to do it cheap" list)

Perhaps. I don't use all that many of them, and in fact have found some in a
box that I've had for years, and never opened.

I have no such box ;-)

If you are facing working with a rather unknown pcb software, making a
layout, sending it out and paying a lot, waiting weeks... i know this makes
you use protoboards. But if you are experienced with the layout software and
have a good pcb setup you just print it out,

This assumes the use of a laser printer?

yes.

fuse it,

Are you talking about toner transfer here?

yes.

peel it,

And here?

still TT.

etch it,

Yeah, not time-consuming if you happen to have everything already set up to
use, and don't have to go through all of that process...

correct, if you have to get everything ready you might just choose not to make a board.

drill it,

Another time-consuming aspect of it, unless you have a CNC setup or similar
(something that's on my list to do).

I try to avoid CNC by heavily going into SMD, limiting the drilling to connectors
and such. if you only convert all the 2-legged ones to smd you can still use your stock
of ics and special parts - but greatly reduce drill effort.

ready. no more than an hour.

I suspect that it would take me considerably longer than that.

Only becaus you haven't set it up and are used to it.

I've been salvaging / scrapping stuff for a lot of years. :-)

me too.. need more storage space. definitely.

If it is possible without too much investment i will try to preserve the
whole frequency range. maybe i just add another direct output from the 038
with 50Ohms, you won't need high amplitude there, or dc offset.

What part would you use to do that?

a resistor. just like in the datasheet. and a bnc connector of course.
the max 038 has 3V amplitude or so on a low impedance output, and a seperate TTL out.

In a meter? First I've heard of that.

Metex has it all, but you can only adjust in steps.
they have crammed in all from Cap. meter, inductance meter, transistor, temp.,
diode, frequency, sig out, logic test,
of course all the RMS amp and volt, ohm stuff and a beeper into a standard size handheld meter...

But there are of course limitations... especially 3,5digit..
I only use it for these special functions, like the desktop meters (with LED display) more.
(fluke 8000). i also highly recommend norma meters, 4 1/2 digit, really fast and high quality..

Tell me more about this?

Counter, using a CPLD, you have a oscillator (i hope i find a more precise one some day)
and you then program your counter in there.
of course you can interchange gate and input so you get period display (you have different
gate times inside by dividers). then you can also add a counter for the thumbwheel (out of
a microwave ;-) ) and let it count n gate times (selectable) before it inverts the output.
you simply buffer the output to protect the cpld and that should be it then.

really not much difference between a counter and a programmable generator.

I have coded the counter function already, and there is still space for a generator i think.
Now i am doing the hardware (when i have time) because i want to do some tests
before continuing the programming.
It's a damn small package these cplds did came in (ebay so i couldn't select plcc).
The toner transfer works fine down to 6,66 mil (2 dots at 300dpi) but the metric spacing
of the tqfp100 leads to errors (uneven spacing).

I'm also not sure if the package can take the heating of the led current, will see.
If it doesn't work i need to scan the display and do the hardware again.

That's a key, figure out what you want to use it for first! :-)

Of course i know what i want to do with it, because otherwise i wouldn't notice a particular tool is missing.
I just don't want to cut corners and then have to make a new unit next week when i need other
functions.


ST

--- In Electronics_101@..., Stefan Trethan
<stefan_trethan@g...> wrote:
...

I try to avoid CNC by heavily going into SMD, limiting the drilling

to

connectors
and such. if you only convert all the 2-legged ones to smd you can

still

use your stock
of ics and special parts - but greatly reduce drill effort.

<much discussion of making PCBs and such deleted>

I was very reluctant to get into SMDs. It seemed like a bad fit for
my aging eyes and deteriorating steadiness of hand. But I made the
jump and now find the thought of using through hole stuff dismaying.
I will do anything to avoid drilling and love the simplicity of
SMDs. They are very easy to use - I feel they are easier than
through hole stuff once you get the technique down. There is lot
less mass to heat up so soldering is less damaging to the parts. I
can put a board together a lot faster as there is no bending or
cutting. Positioning the parts takes a little care but with the well
described techniques, its no problem at all.

You can also prototype with SMDs by using surfboards or making your
own equivalent.

On the subject of making PCBs or not. Its really worth the effort to
make my own. My perfboard creations were a disasster. I'd
inevitably wire it up wrong no matter how carefull I was and then
have a complete mess fixing it. With making my own boards, I still
make design errors but assembly errors are pretty much a thing of the
past for me. Plus, you get a professional looking product out of it.

Anyway, each to his own.

Phil

I toyed with a design not too long ago that would have generated a
logarithmic sweep voltage and swept the MAX038 while applying horizontal
drive to an oscilloscope, and thereby produce 2-D frequency response graphs.
I don't know if that is practical or not, but it sounded cool.

i do not think you can actually get usable numbers out of it.

i only did it once in a school lab this way, was a nice summer day and
we wanted to go home early. amazing what you can do with a digital scope...
well, we did go home early after all...
But i used a professional generator then, no max038.

ST

-DM

开云体育

My 2 cents: ICL 8038 or MAX038. The ICL 8038 is discontinued, but last
I checked Mouser still has a few. The MAX038 is a bit more expensive
but as I recall takes less external circuitry. And goes up quite a bit
higher.

The VCO capabilities of it mean you could also use it as a sweep
generator. Send a ramp wave into the VCO input of either of those
chips, the same ramp into the Horizontal (X) input of your scope,
output of the sign gen into the amp under test, and the output of the
amp into the vertical input of your scope.

Alien Steve

The VCO capabilities of it mean you could also use it as a sweep
generator. Send a ramp wave into the VCO input of either of those
chips, the same ramp into the Horizontal (X) input of your scope,
output of the sign gen into the amp under test, and the output of the
amp into the vertical input of your scope.

Alien Steve

That's how i did it.. but do you have any ideas on how one can get the phase
shift on the screen? (If the scope can't calculate it of course)

ST

--- In Electronics_101@..., Stefan Trethan
<stefan_trethan@g...> wrote:

The VCO capabilities of it mean you could also use it as a sweep
generator. Send a ramp wave into the VCO input of either of those
chips, the same ramp into the Horizontal (X) input of your scope,
output of the sign gen into the amp under test, and the output of the
amp into the vertical input of your scope.

Alien Steve

That's how i did it.. but do you have any ideas on how one can get the
phase
shift on the screen? (If the scope can't calculate it of course)

ST

What do you mean? If you want to see the phase shift of an amp from
input to output, you need to compare input to output and look at the
zero crossing.

Hmm... how about a dual zero crossing detector, one from the output of
the sig gen, other from the output of the amp. The two zero crossing
signals into a phase shift to voltage converter. Sweep the frequency
(not too fast) with the sweep voltage on the X (horiz) input of the
Oscope. Output of the PS2V into the Y (vert) input of the scope. tada!
A rough graph of phase shift vs frequency.

So why am I not a millionaire? Oh, that's right, it's the salesmen and
management who make the big bucks.

Alien Steve

What do you mean? If you want to see the phase shift of an amp from
input to output, you need to compare input to output and look at the
zero crossing.

Hmm... how about a dual zero crossing detector, one from the output of
the sig gen, other from the output of the amp. The two zero crossing
signals into a phase shift to voltage converter. Sweep the frequency
(not too fast) with the sweep voltage on the X (horiz) input of the
Oscope. Output of the PS2V into the Y (vert) input of the scope. tada!
A rough graph of phase shift vs frequency.

So why am I not a millionaire? Oh, that's right, it's the salesmen and
management who make the big bucks.

Alien Steve

thought about that... but was looking for a really simple thing using existing components.
building a PS2V circuit takes me at least as long as measuring the phase shift point by point.
of course if you measure 100 units things are different.

If the scope is able to do add/subtract, what would one get?
i mean the signals are obviously in sync and same frequency.
You would of course need to normalize the amplitude on each channel, i think easy enough to do in any
lab with a series resistor and 2 diodes (or a shorted rectifier).
now switch to add on your scope and get 0 to 2*dide clamp voltage for 180 to 0 degree phase shift.
No opamp, no complicated circuit, no thinking required.

Most scopes can add/subtract (even my tek 7633 which i think is older than myself.)

Now, why am i no millionaire? guess you are right, simple solutions are no good for making
money nowadays... You need bad ideas, and loads of people annoying other people to make money.

ST

--- In Electronics_101@..., "Murat Ceylan" <hector@g...>
wrote:

Hello Everyone,

I'd like to build a signal generator that is not too professional

but good

enough for general use i.e. that can produce sine, sawtooth and

square

waveforms in 1-50kHz range or so with reasonable reliability. Do

you happen

to have any schematics, suggestions, assembly tips etc.?

Neatest way is to use a software DDS:



Leon