Maker Pro
Maker Pro

My recent Tek, Agilent, and LeCroy oscilloscope fun

M

Mr.CRC

Jan 1, 1970
0
Hi:

Recently at work I reviewed Tek MSO4000, Tek DPO7000, Agilent 9000, and
LeCroy MXi-A series oscilloscopes.

I have purchased recently an Agilent MSO7054B, Tek MSO4054, and LeCroy
104MXi-A.

The only one I regret purchasing is the Tek. I would have preferred to
have gotten their new MSO5000 model, but anyway...

The MSO4000 is my least favorite instrument, and will likely be a loaner
scope for giving to labs for temporary use, or for a new tech. that we
hope to hire soon. The reason is that it doesn't have the in-depth data
analysis of the LeCroy, and unlike the Agilent 6000 and 7000 series,
it's waveform update rate slows to a crawl whenever deep memory is used,
or measurements and digital channels are turned on.

The Agilent seems to be the best bang for the buck. I've had a MSO6054
for a few years, and now several of their 7000 series. They maintain
high update rates with the always-on deep memory (not quite as deep as
Tek, only 4MB when running Normal, 8MB with a single acquisition). The
Tek also tends to display persistence artifacts that make you think
something is there, but it's not, just persistence from the last acq.
So now the MSO7054 is my primary mixed signal design scope. The MSO6054
will be there for a student intern this summer.

You are probably wondering, well why the heck did I buy the Tek? I
needed a histogram measurement of some PLL jitter, and the Agilent can't
do that. I also mistakenly was still discounting LeCroy as not worthy
of consideration.

My feelings on the LeCroy changed due to the following endeavor: First,
another lab rat bought a LeCroy because it had more segments to it's
segmented memory acquisition than the Agilent. Also, the Agilent can't
maintain full sampling rate when more than a few 100 segments are
selected. Wierd. So he started selling me on the LeCroy, after defying
me and buying one after I tried to steer him toward Agilent.

Then I had a new measurement problem: Measure and optimize in real time
the shot-shot energy stability (relative standard deviation) of an
optical parametric oscillator, as a function of various parameters--you
don't want to know. OPOs are a bitch.

Well the Agilent and Tek can spit out numbers for standard deviation and
mean of an amplitude or area measure. Then I have to type them into my
calculator to get rel. std. dev. The Agilent accumulates N
acquisitions. The Tek has a moving average with configurable count, but
an unspecified weighting function that seems to lead to discrepancies
vs. the accumulating stats. on it's histograms.

Either of them can get the number with the assistance of a calculator,
but neither can give it directly in real time, nor provide a trend plot.

Enter the LeCroy. LeCroy loaned me a 1GHz 104MXi-A for over two weeks.
On the third day, I figured out how to make my measurement. The thing
can do trend plots of any measurement parameter, such as amplitude. You
can set how many past measurements to plot, so that controls the rate at
which it rolls across the screen, in conjunction with the acquisition
rate.

The you can compute mean and standard deviation (or any other maths) on
the trend plots, not just waveforms! That produces another numeric
measurement parameter output. That in turn can be put into another
trend plot! Magnificent. Then I can do a mean on that, with an
indicator bar. So I get three displays of my relative standard
deviation measurement: a number, a trend plot, and a graphical bar
indicating the average of the trend, sort of like a meter movement.

Sweet!

There is more than one way to set this up as well. It has this amazing
"web math" editor that is like LabView. It lets you graphically pipe
waveforms, parameters, etc. to measurements, math operations, and trend
plots, then pipe those to more, then output the results to measurement
parameter displays or math trace displays. That way, you can compute
operations on a virtual trend plot, without having to use a visible
trace, which are limited to only 8, I believe.

In short, the LeCroy is the best data analysis-oriented scope in the
under $20000 class.

The only drawback to the LeCroy is waveform update rate. It simply
can't come close to the Tek 4000, which also pales compared to the
Agilent (note: Tek DPO7000 is in another class altogether, with up to
250000 wfms/s vs. 100000 for Agilent, but only with 25k samples!).
Fortunately for my 10Hz laser pulses, this limitation of the LeCroy
won't be an issue.

The LeCroy also has these cool "track" plots which are different from
trend plots. Tracks are plots of the history of a measurement, time
correlated with the waveform being measured. So this can effectively
demodulate a PWM waveform, for ex., and plot the pulse width as a
function of time, correlated with the actual PWM waveform. Likewise for
other modulation schemes.

There are of course many other features in all of these instruments that
I haven't yet touched upon.

For general electronic design&troubleshooting, I will be turning first
to my Agilents. For tricky measurements with complicated math and
multiple stages of computation as well as trend plots, the LeCroy is
simply magical.

I don't think I'll be buying any more Teks unless there is something
that just can't be done with any other instrument. Unlikely.

I will have a look at Yokogawa next time I'm in the market though.

LeCroy is developed and made in USA, BTW.
 
J

John Devereux

Jan 1, 1970
0
Mr.CRC said:
Hi:

Recently at work I reviewed Tek MSO4000, Tek DPO7000, Agilent 9000, and
LeCroy MXi-A series oscilloscopes.

I have purchased recently an Agilent MSO7054B, Tek MSO4054, and LeCroy
104MXi-A.

[...]


Then I had a new measurement problem: Measure and optimize in real time
the shot-shot energy stability (relative standard deviation) of an
optical parametric oscillator, as a function of various parameters--you
don't want to know. OPOs are a bitch.

Well the Agilent and Tek can spit out numbers for standard deviation and
mean of an amplitude or area measure. Then I have to type them into my
calculator to get rel. std. dev. The Agilent accumulates N
acquisitions. The Tek has a moving average with configurable count, but
an unspecified weighting function that seems to lead to discrepancies
vs. the accumulating stats. on it's histograms.

Either of them can get the number with the assistance of a calculator,
but neither can give it directly in real time, nor provide a trend plot.

Enter the LeCroy. LeCroy loaned me a 1GHz 104MXi-A for over two weeks.
On the third day, I figured out how to make my measurement. The thing
can do trend plots of any measurement parameter, such as amplitude. You
can set how many past measurements to plot, so that controls the rate at
which it rolls across the screen, in conjunction with the acquisition
rate.

The you can compute mean and standard deviation (or any other maths) on
the trend plots, not just waveforms! That produces another numeric
measurement parameter output. That in turn can be put into another
trend plot! Magnificent. Then I can do a mean on that, with an
indicator bar. So I get three displays of my relative standard
deviation measurement: a number, a trend plot, and a graphical bar
indicating the average of the trend, sort of like a meter movement.

Sweet!

There is more than one way to set this up as well. It has this amazing
"web math" editor that is like LabView. It lets you graphically pipe
waveforms, parameters, etc. to measurements, math operations, and trend
plots, then pipe those to more, then output the results to measurement
parameter displays or math trace displays. That way, you can compute
operations on a virtual trend plot, without having to use a visible
trace, which are limited to only 8, I believe.

In short, the LeCroy is the best data analysis-oriented scope in the
under $20000 class.

The only drawback to the LeCroy is waveform update rate. It simply
can't come close to the Tek 4000, which also pales compared to the
Agilent (note: Tek DPO7000 is in another class altogether, with up to
250000 wfms/s vs. 100000 for Agilent, but only with 25k samples!).
Fortunately for my 10Hz laser pulses, this limitation of the LeCroy
won't be an issue.

The LeCroy also has these cool "track" plots which are different from
trend plots. Tracks are plots of the history of a measurement, time
correlated with the waveform being measured. So this can effectively
demodulate a PWM waveform, for ex., and plot the pulse width as a
function of time, correlated with the actual PWM waveform. Likewise for
other modulation schemes.

There are of course many other features in all of these instruments that
I haven't yet touched upon.

For general electronic design&troubleshooting, I will be turning first
to my Agilents. For tricky measurements with complicated math and
multiple stages of computation as well as trend plots, the LeCroy is
simply magical.

I don't think I'll be buying any more Teks unless there is something
that just can't be done with any other instrument. Unlikely.

I will have a look at Yokogawa next time I'm in the market though.

Thanks for all that, I am (still) looking for some similar features
(trend plotting and maths on measurements). I had not considered LeCroy.

Yokogawa do trend plotting too. Rohde & Shwartz have some scopes that
look really nice other than lack of trend plotting, but I have been told
recently that this is going to be in their new firmware. (All these are
~$20k class devices too).

Can the LeCroy do XY plots of measurements? That would open the door to
lots of things, you could reproduce onscreen many of the graphs you see
in component datasheets. For example gain vs temperature, PWM duty vs
circuit output voltage etc.
 
J

JW

Jan 1, 1970
0
[...]

Nice write-up.
LeCroy is developed and made in USA, BTW.

There's a switch. Used to be that they were made in Switzerland, but since
I don't have your budget, I've not used or worked on anything newer than
the LC series.
 
M

Mr.CRC

Jan 1, 1970
0
John said:
Can the LeCroy do XY plots of measurements? That would open the door to
lots of things, you could reproduce onscreen many of the graphs you see
in component datasheets. For example gain vs temperature, PWM duty vs
circuit output voltage etc.

I am curious about the same question. I have to return the loaner on
Monday. I will investigate when the new one arrives later in Jan.
 
M

Mr.CRC

Jan 1, 1970
0
JW said:
[...]

Nice write-up.
LeCroy is developed and made in USA, BTW.

There's a switch. Used to be that they were made in Switzerland, but since
I don't have your budget, I've not used or worked on anything newer than
the LC series.


Speaking of budgets, I have been wondering lately what proportion of the
market for these instrument makers is government. I wonder if we would
see so many offerings if it wasn't for this demand.

Surely the pressure on private business must be intense to minimize
costs. I see John Larkin for instance looking into inexpensive Chinese
instruments rather than the premium brands. Though there is an
increasing tendency for Tek, Agilent, and LeCroy these days to offer a
large range of economy instruments.

I understand of course that the very high end instruments are targeting
the cutting edge of ludicrous-speed chip design, bus interfaces, and
communications links. But the quantities shipped of these instruments
must be in the handfuls.

I suppose the Chinese instruments will improve with time. But I am
hesitant at this stage, even for my personal hobby use, to step down
from the Agilent, Tek, and LeCroy offerings.
 
M

Mr.CRC

Jan 1, 1970
0
Paul said:
clip ....

There's been quite a discussion lately about all the analysis various
scopes can do. But whatever they build into a scope, its going to be
nothing compared to what you can do by downloading the data into a
programming environment such as Matlab, IDL, Octave, R, whatever. Why
not let the scope be a scope, let a computer be a computer? So then the
question is: how fast can these scopes transfer files to a PC? How easy
is it for a user program on a PC to control the scope, get it's status,
and transfer the files? Does the scope manufacturer supply you with some
PC software to make this kind of operation easier?

Paul Probert


Thanks for the comment.

I suppose for the same reason that there are scopes with full front
panel controls and displays as well as digitizer-only boxes that can
only be accessed by a computer, so it is with advanced analysis features
vs. just a simple "scope" interface.

That is, there is a market for real time analysis, and there is a market
for post-processing analysis. I agree that post processing will always
be more powerful. But some users need real-time results.

A similar trend is developing with data acquisition hardware, where more
manufacturers are providing devices with built-in DSP processors.

For the same scope that my coworker and I are buying, he will collect
2000 or so 10ns laser pulses, and in post-processing analyze them to
normalize some other data. Whereas I will look at the same laser pulses
and use real-time processing to optimize and monitor an OPO.

I had considered taking one of the existing scopes and writing some
program to pull down data and post-process into the same result that I
want. I suppose it could be done on the cheap with a GUI development
language and avoiding the expensive LabView or MatLAB. But in any
event, if it took two weeks of my time to develop (and I don't
specialize in high level programming, but rather bare silicon, so this
would probably take me several weeks) well, that's $6000 at a minimum of
my overhead rate.

It's funny how sometimes we'd rather do a job in house because the labor
cost is already committed. Yet other times we'd rather just buy an
off-the shelf gadget because the uncertainty as well as lost opportunity
cost of having labor sunk in some tangential development project instead
of the immediate priority (in this case--make the OPO work!) is more
costly than the gadget.
 
N

Nico Coesel

Jan 1, 1970
0
Mr.CRC said:
JW said:
[...]

Nice write-up.
LeCroy is developed and made in USA, BTW.

There's a switch. Used to be that they were made in Switzerland, but since
I don't have your budget, I've not used or worked on anything newer than
the LC series.


Speaking of budgets, I have been wondering lately what proportion of the
market for these instrument makers is government. I wonder if we would
see so many offerings if it wasn't for this demand.

Surely the pressure on private business must be intense to minimize
costs. I see John Larkin for instance looking into inexpensive Chinese
instruments rather than the premium brands. Though there is an
increasing tendency for Tek, Agilent, and LeCroy these days to offer a
large range of economy instruments.

I understand of course that the very high end instruments are targeting
the cutting edge of ludicrous-speed chip design, bus interfaces, and
communications links. But the quantities shipped of these instruments
must be in the handfuls.

I agree. Although I feel there is a huge gap in the market. There are
tons of low end <$1000 scopes. If you want a higher screen resolution,
 
N

Nico Coesel

Jan 1, 1970
0
Paul Probert said:
clip ....

There's been quite a discussion lately about all the analysis various
scopes can do. But whatever they build into a scope, its going to be
nothing compared to what you can do by downloading the data into a
programming environment such as Matlab, IDL, Octave, R, whatever. Why
not let the scope be a scope, let a computer be a computer? So then the
question is: how fast can these scopes transfer files to a PC? How easy
is it for a user program on a PC to control the scope, get it's status,
and transfer the files? Does the scope manufacturer supply you with some
PC software to make this kind of operation easier?

I think this is a good point. I used this in the past to do some
prototyping where I used a measurement instrument as an analog
front-end to process some real input signals. Another nice to have
would be the ability to write plug-ins. For example: a few months ago
I wrote an SPI decoder plug-in for my TLA704 logic analyser.
 
On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

OK, he did not make a profit this year?

Once a liar, always a liar, Jan.
 
On a sunny day (Sat, 18 Dec 2010 09:35:50 -0600) it happened
<[email protected]>:

On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

OK, he did not make a profit this year?

Once a liar, always a liar, Jan.

I did not know he was that too, because that was HIS statement.

You're illiterate, too. ...but I knew that.
 
On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

We were asked, by Los Alamos, to design and build a 1 ns resolution
TDC module, as an alternate to some truly terrible LeCroy units (their
4208, our M680). The LeCroys were expensive, slow delivery, up to 50%
DOA, and took six months to get fixed. The next bid, LeCroy somehow
found out there would be competition and cut their price in half to
kill us. Our friend at Los Alamos disqualified them on technical
grounds!

That TDC was our first high-speed product. I'd never done anything
like that before.

I wanted the liar to tell the rest of the story. I knew the Europeon moron
couldn't tell the truth.
 
On a sunny day (Sat, 18 Dec 2010 10:14:14 -0600) it happened
<[email protected]>:

On a sunny day (Sat, 18 Dec 2010 09:35:50 -0600) it happened
<[email protected]>:

On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

OK, he did not make a profit this year?

Once a liar, always a liar, Jan.

I did not know he was that too, because that was HIS statement.

You're illiterate, too. ...but I knew that.

Sorry about your Alzheimer progressing,
I have read they found a cause though.
It won't help you now, but if you re-incarnate as a demonrat you may have a chance.

You even sound like Slowman. Kill yourself now.
 
On Sat, 18 Dec 2010 09:35:50 -0600, "[email protected]"

On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

We were asked, by Los Alamos, to design and build a 1 ns resolution
TDC module, as an alternate to some truly terrible LeCroy units (their
4208, our M680). The LeCroys were expensive, slow delivery, up to 50%
DOA, and took six months to get fixed. The next bid, LeCroy somehow
found out there would be competition and cut their price in half to
kill us. Our friend at Los Alamos disqualified them on technical
grounds!

That TDC was our first high-speed product. I'd never done anything
like that before.

I wanted the liar to tell the rest of the story. I knew the Europeon moron
couldn't tell the truth.

What in the world are you talking about? The story is true. The CAMAC
modules exist. You can even buy them, ours and LeCroy's, on ebay now
and then.

I was replying to Jan Panteltje, Slowman's clone. He's the liar.
 
N

Nico Coesel

Jan 1, 1970
0
John Larkin said:
JW said:
[...]

Nice write-up.

LeCroy is developed and made in USA, BTW.

There's a switch. Used to be that they were made in Switzerland, but since
I don't have your budget, I've not used or worked on anything newer than
the LC series.


Speaking of budgets, I have been wondering lately what proportion of the
market for these instrument makers is government. I wonder if we would
see so many offerings if it wasn't for this demand.

Surely the pressure on private business must be intense to minimize
costs. I see John Larkin for instance looking into inexpensive Chinese
instruments rather than the premium brands.

The premium brands are mostly asian too. Most Tek stuff is built in
China. The low-end Agilent scopes are rebranded Rigols. Agilent is
increasingly moving manufacturing to Maylasia.

If I'm going to buy a Rigol scope, I may as well buy it from Rigol,
instead of paying Agilent 2.5X the price for the same box.

Some of the Keithley benchtop DVMs are Chinese rebrands. Crap. I sent
three of them back to Keithley.

Though there is an
increasing tendency for Tek, Agilent, and LeCroy these days to offer a
large range of economy instruments.

I understand of course that the very high end instruments are targeting
the cutting edge of ludicrous-speed chip design, bus interfaces, and
communications links. But the quantities shipped of these instruments
must be in the handfuls.

I suppose the Chinese instruments will improve with time. But I am
hesitant at this stage, even for my personal hobby use, to step down
from the Agilent, Tek, and LeCroy offerings.

My 50 MHz Rigol is great. It has a lot more features than my Tek
TDS2012, which is over three times the price. Nicer probes, too.

I wonder if there are any good Chinese spectrum analyzers.

Digital or analog? A couple of years ago my employer bought an Atten
AT6011 (1GHz + tracking generator) for simple RF stuff and EMC problem
finding. It takes some time to get used to an analog instrument but it
does work.
 
N

Nico Coesel

Jan 1, 1970
0
John Larkin said:
On a sunny day (Fri, 17 Dec 2010 20:27:25 -0800) it happened "Mr.CRC"

John Larkin wrote:
On Thu, 16 Dec 2010 22:46:51 -0800, "Mr.CRC"
[this and that]
LeCroy is developed and made in USA, BTW.


All nice, except that LeCroy is evil.

John


??? Care to elaborate?

Don't listen to him.
LeCroy was his competition once and undercut his price.
That makes it 'evil' you see.

Now tell them the rest of the story.

We were asked, by Los Alamos, to design and build a 1 ns resolution
TDC module, as an alternate to some truly terrible LeCroy units (their
4208, our M680). The LeCroys were expensive, slow delivery, up to 50%
DOA, and took six months to get fixed. The next bid, LeCroy somehow
found out there would be competition and cut their price in half to
kill us. Our friend at Los Alamos disqualified them on technical
grounds!

Isn't that just how the market works? If there are more competitors
prices will drop.
 
M

Mr.CRC

Jan 1, 1970
0
John said:
My 50 MHz Rigol is great. It has a lot more features than my Tek
TDS2012, which is over three times the price. Nicer probes, too.
John


How's the waveform update rate on those things?

I bought a Tek TDS3014 about 7 years ago for home hobby use. That cost
me 2.5 years of my allowance. I don't regret it. At that time, the
TDS3000 series was where the price/performance breakthrough was.

I'm considering to get an Agilent DSO7034 350MHz in a few years, then
upgrade some time to the digital channels option. People spend a heck
of a lot more on Harleys and other useless toys.

I'd like to build a Q-switched YAG laser or N2 laser and do a time of
flight speed of light experiment for my daughter when she gets to a
grade where that will make sense. Maybe 3rd grade or so ;-) Might even
be able to get LEDs to do the job. I could probably measure it with my
100MHz scope, but 350MHz is a lot better.
 
T

Tim Williams

Jan 1, 1970
0
Mr.CRC said:
I'd like to build a Q-switched YAG laser or N2 laser and do a time of
flight speed of light experiment for my daughter when she gets to a
grade where that will make sense. Maybe 3rd grade or so ;-) Might even
be able to get LEDs to do the job. I could probably measure it with my
100MHz scope, but 350MHz is a lot better.

We did that in, Physics 2 or something, using a regular laser diode, a
rather slow photodiode, and the hallway. The mirror is moved to set
distance. Fifty feet or so makes it easy. The hardest part is aligning
the mirrors, and foot traffic making them bounce.

Tim
 
M

Mr.CRC

Jan 1, 1970
0
Tim said:
We did that in, Physics 2 or something, using a regular laser diode, a
rather slow photodiode, and the hallway. The mirror is moved to set
distance. Fifty feet or so makes it easy. The hardest part is aligning
the mirrors, and foot traffic making them bounce.

Tim

Yeah, we did the same thing, but with an N2 pumped dye laser and a
300-ish MHz LeCroy scope. That was my first encounter with LeCroy.

Trying it with a retro-reflector might be interesting.
 
Top