Need VCO for 14 MHz PLL

Joerg · Jul 17, 2005

Hello Jim,

Ordinary 74HC... is optimized at +5V, will work reasonably well down
to +2.5V, but becomes like molasses at +1.5V

Same with the CD series. It is technically ok down to 3V but there it's
like molasses at 110F.

The I/C stuff I'm working on right now has 5V, 3.3V, 2.5V and 1.8V
devices on a single chip.

The only chip where I did the architecture and some of the analog parts
had 45V and 5V. It was one of the automotive processes. Our budget
allowed exactly one shot and lo and behold it worked. Whew...

It also had to be much smaller than possible so we had to throw
"ballast". It was the electromigration versus lifetime thing. The chip
engineer wasn't too enthused because that isn't the way they usually
work. But in our case it was for a disposable device.

Your chips are probably also required to run well into the picosecond
range. Ours didn't have to, nothing was above 40MHz. So we could go easy
on the routing.

Regards, Joerg

http://www.analogconsultants.com

Jim Thompson · Jul 17, 2005

On Mon, 18 Jul 2005 00:43:37 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Jim,

Ordinary 74HC... is optimized at +5V, will work reasonably well down
to +2.5V, but becomes like molasses at +1.5V

Same with the CD series. It is technically ok down to 3V but there it's
like molasses at 110F.

The I/C stuff I'm working on right now has 5V, 3.3V, 2.5V and 1.8V
devices on a single chip.

The only chip where I did the architecture and some of the analog parts
had 45V and 5V. It was one of the automotive processes. Our budget
allowed exactly one shot and lo and behold it worked. Whew...

It also had to be much smaller than possible so we had to throw
"ballast". It was the electromigration versus lifetime thing. The chip
engineer wasn't too enthused because that isn't the way they usually
work. But in our case it was for a disposable device.

Your chips are probably also required to run well into the picosecond
range. Ours didn't have to, nothing was above 40MHz. So we could go easy
on the routing.

Regards, Joerg

http://www.analogconsultants.com

480)460-2350 | |
| E-mail Address at Website Fax

Joerg · Jul 17, 2005

Hi Winfield,

These days the multiplicity of supply-voltages is getting to be a
real pain for the designer. For example, modern 2.5 and 1.8V cPLDs
have a rather substantial power draw, and generally their supply is
not well-made using a linear regulator from +5V. One ends up with
multiple supply sources, often with special sequencing rules, and
perhaps with surprise interactions, like the difference frequency
between two switching regulators. A design I'm finishing now, a
five-channel DDS frequency generator, has seven supply voltages...
And it's meant to be well-behaved to the 10ppm level. What a pain!

Year ago I used to do that with flyback or other transformer-based
switchers and multiple taps. But when voltages go down to 1.8V and the
tolerances are tight that becomes impractical because you can't regulate
individually. Unless there are LDO regulators in tow but these have
their own stability issues.

Regards, Joerg

http://www.analogconsultants.com

Joerg · Jul 18, 2005

Hello Jim,

I'm working a project at 2.2GHz right now.

Mine right now is a puny 400kHz :-(

Regards, Joerg

http://www.analogconsultants.com

Jim Thompson · Jul 18, 2005

On Mon, 18 Jul 2005 01:02:37 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Jim,

I'm working a project at 2.2GHz right now.

Mine right now is a puny 400kHz :-(

Regards, Joerg

http://www.analogconsultants.com

480)460-2350 | |
| E-mail Address at Website Fax

Joerg · Jul 18, 2005

Hello Jim,

I'm also working on a hearing aid (analog) with MEMS microphone
on-chip.

Similar to an on-chip pressure sensor? If yes and it isn't confidential,
what is your favorite fab to do that with?

Regards, Joerg

http://www.analogconsultants.com

Jim Thompson · Jul 18, 2005

On Mon, 18 Jul 2005 01:23:06 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Jim,

I'm also working on a hearing aid (analog) with MEMS microphone
on-chip.

Similar to an on-chip pressure sensor? If yes and it isn't confidential,
what is your favorite fab to do that with?

Regards, Joerg

http://www.analogconsultants.com

480)460-2350 | |
| E-mail Address at Website Fax

John Larkin · Jul 18, 2005

On 17 Jul 2005 17:32:42 -0700, Winfield Hill
<Winfield_member@newsguy.com> wrote:

Jim Thompson wrote...

The I/C stuff I'm working on right now has 5V, 3.3V, 2.5V and 1.8V
devices on a single chip.

These days the multiplicity of supply-voltages is getting to be a
real pain for the designer. For example, modern 2.5 and 1.8V cPLDs
have a rather substantial power draw, and generally their supply is
not well-made using a linear regulator from +5V. One ends up with
multiple supply sources, often with special sequencing rules, and
perhaps with surprise interactions, like the difference frequency
between two switching regulators. A design I'm finishing now, a
five-channel DDS frequency generator, has seven supply voltages...
And it's meant to be well-behaved to the 10ppm level. What a pain!

How about this: +5, +3.5, +2.5, and +1.25... just for the digital
stuff. Then 16 isolated channels, each with +-18, +5, and +-2.5.
Arguably 84 power supplies on one board!

John

Rene Tschaggelar · Jul 18, 2005

W Letendre wrote:

Need to generate in phase and quadrature digital signals, locked to
laser detector output at ~ 3.6 MHz (quite stable). Need reasonably high
precision 90deg shift and 50% duty fator +/- a percent or so. Thought
of building 4X multiplying PLL, followed by 2 bit quadrature divider.
So, looking for VCO that will run at 14.4 MHz.

Seem to find lots of chips for cell phone and other VHF/UHF/uWave apps,
and old standbys from 556 and 4046 family, but nothing that would work
at 14.4 MHz. Suppose I could make a Colpitts or Hartley circuit using
varactor diode for variable capacitance, but, hoping there is a chip
that will do this. Checked Maxim, National, Motorola, Linear
Technologies, Burr Brown, with no joy. Any other suggestions?

Yes, have a look at the Linear Technology
Oscillators 69xx Series. There a linear
voltage can be used.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

Tim Shoppa · Jul 18, 2005

old standbys from 556 and 4046 family, but nothing
that would work at 14.4 MHz

74HC4046's will typically go up to 18MHz if you don't get too far away
from 5V Vcc and room temperature.

If you have to operate at the corners of temp/Vcc then they may be
ruled out.

A much more stable VCO oscillator is an LC oscillator with a varactor.

Tim.

Tim Shoppa · Jul 18, 2005

Win wrotes:

These days the multiplicity of supply-voltages
is getting to be a real pain for the designer.

Reminds me of mixing PMOS with NMOS with TTL in the 70's. Remember how
EPROM's and DRAM's and dynamic shift registers used to need all sorts
of funny voltages like -22.5, -18, -12, +7, etc? Don't forget those
WOM chips needing 6.3VAC either

Joerg replies:

Unless there are LDO regulators in tow

I like the latest generation LDO's for my
one-off-will-never-be-used-off-my-bench gadgets. If a chip needs 1.8V,
then a 1.8V LDO goes on the board.

I occasionally get out my precision voltmeters and many of the
low-noise LDO's have remarkably good outputs. Always way better than
the specs for DC tolerance and always within the published noise specs.

Never had a problem following the spec-sheet charts. I contrast this
with my usage of 3-terminal 780x style regulators, where occasional
batches seemed to always oscillate, resulting in us banning that brand.
You're in trouble when you've banned all the brands!

But for someone looking to put a bunch of mixed-voltage chips on the
same board and mass produce it, I certainly do appreciate that it's a
pain.

Tim.

W Letendre · Jul 18, 2005

Tim Wescott wrote:

W Letendre wrote:

Need to generate in phase and quadrature digital signals, locked to
laser detector output at ~ 3.6 MHz (quite stable). Need reasonably high
precision 90deg shift and 50% duty fator +/- a percent or so. Thought
of building 4X multiplying PLL, followed by 2 bit quadrature divider.
So, looking for VCO that will run at 14.4 MHz.

Seem to find lots of chips for cell phone and other VHF/UHF/uWave apps,
and old standbys from 556 and 4046 family, but nothing that would work
at 14.4 MHz. Suppose I could make a Colpitts or Hartley circuit using
varactor diode for variable capacitance, but, hoping there is a chip
that will do this. Checked Maxim, National, Motorola, Linear
Technologies, Burr Brown, with no joy. Any other suggestions?

Thanks
W Letendre

If "quite stable" means "crystal stable", and you don't need a large
range, and you have the time for things to settle down, then take a look
at VXCO's -- they're available from many, if not most, of the crystal
oscillator folks. And they're nearly crystal stable themselves.

--
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Sad to say, laser vendor specs their detector output as being stable to
~ 0.1 ppm, but, with a frequency accuracy of +/- 10%. Not even sure how
this odd combination of specs is arrived at by design, but, it's what
I'm working with. Means that VCXO is ruled out.....

W Letendre

Joerg · Jul 18, 2005

Hello Jim,

X-Fab

Cool. So you might visit Erfurt some day where they do their MEMS stuff.
Nice historic town.

Regards, Joerg

http://www.analogconsultants.com

Joerg · Jul 18, 2005

Hello Tim,

Never had a problem following the spec-sheet charts. I contrast this
with my usage of 3-terminal 780x style regulators, where occasional
batches seemed to always oscillate, resulting in us banning that brand.
You're in trouble when you've banned all the brands!

We once got a bloody nose because one of the LDOs (early 90's) had a
habit that wasn't on the data sheet. It became unstable when the source
impedance exceeded a certain value. Not good.

I never use 78xx but stay with Old Faithful, the LM317 series. Keeps the
inventory levels down because there is only one adjustable type. But
most of the time I design in switchers.

Regards, Joerg

http://www.analogconsultants.com

Tim Shoppa · Jul 18, 2005

If "quite stable" means "crystal stable", and you don't
need a large range, and you have the time for
things to settle down, then take a look at
VXCO's -- they're available from many, if not
most, of the crystal oscillator folks. And they're
nearly crystal stable themselves.

The VCTXO's are remarkably good AND available off-the-shelf for many
frequencies.

This guy (who does have the requisite measurement equipment!) shows how
many of them are good to the part per billion in stability if treated
nicely:

http://www.leapsecond.com/museum/htv2/

But when the original poster mentioned 555's and 4046's, maybe we've
hit the overkill stage

.

Tim.

Joerg · Jul 18, 2005

Hello Tim,

74HC4046's will typically go up to 18MHz if you don't get too far away
from 5V Vcc and room temperature.

That also makes sure you don't need to worry about de-icing them in
winter ;-)

A much more stable VCO oscillator is an LC oscillator with a varactor.

That is probably also the lowest cost method in this case.

Regards, Joerg

http://www.analogconsultants.com

Need VCO for 14 MHz PLL

Joerg

Guest

Jim Thompson

Guest

Joerg

Guest

Joerg

Guest

Jim Thompson

Guest

Joerg

Guest

Jim Thompson

Guest

John Larkin

Guest

Rene Tschaggelar

Guest

Tim Shoppa

Guest

Tim Shoppa

Guest

W Letendre

Guest

Joerg

Guest

Joerg

Guest

Tim Shoppa

Guest

Joerg

Guest

Welcome to EDABoard.com

Sponsor

Online statistics

Forum statistics

Need VCO for 14 MHz PLL

Joerg

Guest

Jim Thompson

Guest

Joerg

Guest

Joerg

Guest

Jim Thompson

Guest

Joerg

Guest

Jim Thompson

Guest

John Larkin

Guest

Rene Tschaggelar

Guest

Tim Shoppa

Guest

Tim Shoppa

Guest

W Letendre

Guest

Joerg

Guest

Joerg

Guest

Tim Shoppa

Guest

Joerg

Guest

Log in

Welcome to EDABoard.com

Sponsor