On Sat, 21 Feb 2004 03:20:07 -0800, Roy Lewallen
wrote:

Tom Bruhns wrote:
. . .
. . . Seems like step
recovery diodes are not in as great favor as they once were, since
there are generally better ways to generate higher order harmonics.
. . .

Getting a bit off-topic here, but as of a few years ago, we were using
step recovery diodes to generate the step in high speed TDR systems, and
to generate the strobe for the sampling gate in high speed sampling
scopes. Rise times were on the order of 7 - 15 ps (bandwidth up to 50
GHz or so), limited primarily by circuitry external to the diodes. SRDs
replaced tunnel diodes in earlier generations of instruments. I've been
out of touch with that class of instruments for a few years now -- do
you know if something has replaced the SRD for generating fast steps, or
just for harmonic generation?

What's a doubler based on the good old 1N4148 good for, top end
frequency-wise?
--

The BBC: Licensed at public expense to spread lies.

On Sat, 21 Feb 2004 12:28:11 +0000, Paul Burridge
wrote:

On Sat, 21 Feb 2004 03:20:07 -0800, Roy Lewallen
wrote:

Tom Bruhns wrote:
. . .
. . . Seems like step
recovery diodes are not in as great favor as they once were, since
there are generally better ways to generate higher order harmonics.
. . .

Getting a bit off-topic here, but as of a few years ago, we were using
step recovery diodes to generate the step in high speed TDR systems, and
to generate the strobe for the sampling gate in high speed sampling
scopes. Rise times were on the order of 7 - 15 ps (bandwidth up to 50
GHz or so), limited primarily by circuitry external to the diodes. SRDs
replaced tunnel diodes in earlier generations of instruments. I've been
out of touch with that class of instruments for a few years now -- do
you know if something has replaced the SRD for generating fast steps, or
just for harmonic generation?

What's a doubler based on the good old 1N4148 good for, top end
frequency-wise?

made some experiments 10-15 years ago with doublers to 144Mc/s, and
they probably would work on at least 200Mc. Best experience with a
BFR90 amplifier following the 'rectifier', see
http://home.online.no/~la8ak/c13.htm
It was important with certain dc load following the diodes and some
bias current

Another interesting multiplier used for 100kc calibrator - on vhf -
described in UKW Berichte uses quad nand schmidt trigger, where the
input signal is splitted - one part to a nand input and the other to
3x nand gates connected as inverters and connected to the second input
of the nand-gate such that the truth table said constant logic high
output, but a very thin spike occured because of the transition time
delay

73
JM
----
Jan-Martin, LA8AK, N-4623 Kristiansand
http://home.online.no/~la8ak/

On Sat, 21 Feb 2004 12:28:11 +0000, Paul Burridge
wrote:

On Sat, 21 Feb 2004 03:20:07 -0800, Roy Lewallen
wrote:

Tom Bruhns wrote:
. . .
. . . Seems like step
recovery diodes are not in as great favor as they once were, since
there are generally better ways to generate higher order harmonics.
. . .

Getting a bit off-topic here, but as of a few years ago, we were using
step recovery diodes to generate the step in high speed TDR systems, and
to generate the strobe for the sampling gate in high speed sampling
scopes. Rise times were on the order of 7 - 15 ps (bandwidth up to 50
GHz or so), limited primarily by circuitry external to the diodes. SRDs
replaced tunnel diodes in earlier generations of instruments. I've been
out of touch with that class of instruments for a few years now -- do
you know if something has replaced the SRD for generating fast steps, or
just for harmonic generation?

What's a doubler based on the good old 1N4148 good for, top end
frequency-wise?

made some experiments 10-15 years ago with doublers to 144Mc/s, and
they probably would work on at least 200Mc. Best experience with a
BFR90 amplifier following the 'rectifier', see
http://home.online.no/~la8ak/c13.htm
It was important with certain dc load following the diodes and some
bias current

Another interesting multiplier used for 100kc calibrator - on vhf -
described in UKW Berichte uses quad nand schmidt trigger, where the
input signal is splitted - one part to a nand input and the other to
3x nand gates connected as inverters and connected to the second input
of the nand-gate such that the truth table said constant logic high
output, but a very thin spike occured because of the transition time
delay

73
JM
----
Jan-Martin, LA8AK, N-4623 Kristiansand
http://home.online.no/~la8ak/

Paul Burridge wrote in message . ..

What's a doubler based on the good old 1N4148 good for, top end
frequency-wise?

Thanks to Jan-Martin for his reference to some actual experiments.

But in reply to Paul, I'd ask: Do you understand how the "full-wave
rectifier doubler" works, basically? (Ideal waveforms and all that.)
Do you have a data sheet for the 1N4148? What items from the data
sheet do you suppose might limit the useful frequency? Can you make
an estimate, based on the data sheet numbers? What would you do in a
design to extend the frequency range for a given diode characteristic?
For example, what does diode capacitance do to circuit operation?
What does reverse recovery do? In the full-wave frequency doubler
circuit, what does the input impedance look like, assuming an ideal
transformer, when one diode is forward biased and the other is
reverse-recovering? Can you think of parts to add to cause that to
not be so much of a problem (assuming it is a problem)?

Thinking about this sort of thing is useful not only in figuring out
what to expect, at least ball-park, but also in getting better
performance out of someone else's circuit and/or understanding its
limitations.

Cheers,
Tom

On 22 Feb 2004 15:50:35 -0800, (Tom Bruhns) wrote:

But in reply to Paul, I'd ask: Do you understand how the "full-wave
rectifier doubler" works, basically? (Ideal waveforms and all that.)

IFAIA, pretty much the same as a full-wave rectifier in a PSU. All the
negative going half-cycles are converted to postitive going
half-cycles that 'slot in' between the unmodified positive half
cycles. The resulting waveform is twice the original frequency and
subsequent signal processing can restore this half-wave signal back to
a full sine wave of the desired output frequency.

Do you have a data sheet for the 1N4148?

I could get one off the good old 'net. Aw, **** it; I'll get one. Hang
on a minute..... Got it!

What items from the data
sheet do you suppose might limit the useful frequency?

Primarily junction capacitance, I guess. In this instance, 4pF.
There's also something called transit time, IIRC, which is relevant,
but for some reason it's not mentioned on the DS.

Can you make
an estimate, based on the data sheet numbers?

Nope. Maybe from the graphs, though...

What would you do in a
design to extend the frequency range for a given diode characteristic?
For example, what does diode capacitance do to circuit operation?

More slows it down.

What does reverse recovery do?

Slower recovery the same, I guess. I'm not totally familiar with this
parameter but it would make sense.

In the full-wave frequency doubler
circuit, what does the input impedance look like, assuming an ideal
transformer, when one diode is forward biased and the other is
reverse-recovering? Can you think of parts to add to cause that to
not be so much of a problem (assuming it is a problem)?

Totally beyond me at this stage, I'm afraid!

--

The BBC: Licensed at public expense to spread lies.

On 22 Feb 2004 15:50:35 -0800, (Tom Bruhns) wrote:

But in reply to Paul, I'd ask: Do you understand how the "full-wave
rectifier doubler" works, basically? (Ideal waveforms and all that.)

IFAIA, pretty much the same as a full-wave rectifier in a PSU. All the
negative going half-cycles are converted to postitive going
half-cycles that 'slot in' between the unmodified positive half
cycles. The resulting waveform is twice the original frequency and
subsequent signal processing can restore this half-wave signal back to
a full sine wave of the desired output frequency.

Do you have a data sheet for the 1N4148?

I could get one off the good old 'net. Aw, **** it; I'll get one. Hang
on a minute..... Got it!

What items from the data
sheet do you suppose might limit the useful frequency?

Primarily junction capacitance, I guess. In this instance, 4pF.
There's also something called transit time, IIRC, which is relevant,
but for some reason it's not mentioned on the DS.

Can you make
an estimate, based on the data sheet numbers?

Nope. Maybe from the graphs, though...

What would you do in a
design to extend the frequency range for a given diode characteristic?
For example, what does diode capacitance do to circuit operation?

More slows it down.

What does reverse recovery do?

Slower recovery the same, I guess. I'm not totally familiar with this
parameter but it would make sense.

In the full-wave frequency doubler
circuit, what does the input impedance look like, assuming an ideal
transformer, when one diode is forward biased and the other is
reverse-recovering? Can you think of parts to add to cause that to
not be so much of a problem (assuming it is a problem)?

Totally beyond me at this stage, I'm afraid!

--

The BBC: Licensed at public expense to spread lies.

Paul Burridge wrote in message . ..

What's a doubler based on the good old 1N4148 good for, top end
frequency-wise?

Thanks to Jan-Martin for his reference to some actual experiments.

But in reply to Paul, I'd ask: Do you understand how the "full-wave
rectifier doubler" works, basically? (Ideal waveforms and all that.)
Do you have a data sheet for the 1N4148? What items from the data
sheet do you suppose might limit the useful frequency? Can you make
an estimate, based on the data sheet numbers? What would you do in a
design to extend the frequency range for a given diode characteristic?
For example, what does diode capacitance do to circuit operation?
What does reverse recovery do? In the full-wave frequency doubler
circuit, what does the input impedance look like, assuming an ideal
transformer, when one diode is forward biased and the other is
reverse-recovering? Can you think of parts to add to cause that to
not be so much of a problem (assuming it is a problem)?

Thinking about this sort of thing is useful not only in figuring out
what to expect, at least ball-park, but also in getting better
performance out of someone else's circuit and/or understanding its
limitations.

Cheers,
Tom