Recommendation on o'scope for working with ADI 's DSPs

Hi, 

Maybe this should be crossposted to other newsgroups, but I don't know
which ones.

I need your recommendation for buying a digital oscilloscope mainly
for use with ADI DSPs (I mention that because of the SPORTs and link
ports), and also general digital design. All of these instruments are
to be bougth used, but functional.

The choices are:

- HP/Agilent 54100D
*1 GHz bandwidth/10 ps resolution
*40 megasample/second digitizing rate
*10 bits vertical resolution with averaging
*1k memory per channel
*Two vertical input channels and two trigger inputs, each selectable
for high, low, don't care.
*In addition to triggering on entering or exiting the specified
pattern, it can be set to trigger when a specified pattern is present
for less than or greater than a specified duration, from 10 nS to 5
Seconds.
*Can be set to trigger on an edge of either polarity on any one of the
four inputs.
*Four HP 54002A pods included

- Tektronix TDS350
Signal Acquisition System
* Bandwidth: 200 MHz
* Sample Rate: 1 GS/s on each channel
* Channels: two identical channels, each with invert function
* Sensitivity: 2 mV to 10 V/div
* Vertical Resolution: 8 bits (256 levels over 10.24 vertical
divisions)
Time Base System (Main and Delayed)
* Time/Division Range: 2.5 ns - 5 s/div
* Record Length: 1000 sample points per channel 
Triggering System (Main Only)
* Triggering Types: Edge, Video 
Automatic Measurements 
Option 14 (I/O Interface): Computer Interface and Hardcopy Capability 
* GPIB (IEEE-488.2) Programmability
* RS-232 Interface Programmability
* Video Output
* Hardcopy Port - Centronics-type parallel or RS-232
* Graphics Interface File Formats - Interleaf (.img), TIFF, PCX (PC
Paintbrush), BMP (Microsoft Windows), and Encapsulated Postscript
(EPS)
* Printer Formats - ThinkJet, DeskJet, LaserJet, Epson (9 & 24 Pin),
Seiko DPU 411/II, DPU 412.  DC power provided for Seiko printers

- Tektronix TDS2002
* 60 MHz Bandwidth * Sample Rate 1 GS/s * 2 Channels * Color Display *
CompactFlash Mass Storage Option with TDS2MEM Module * Autoset Menu
with Waveform Selection * Probe Check Wizard To Ensure Correct Probe
Usage * Context-sensitive Help * Dual Time Base * Advanced Triggers,
including Pulse Width Trigger and Line-Selectable Video Trigger * 11
Automatic Measurements * Multi-language User Interface * Waveform and
Setup Memories * FFT Standard on All Models

My main doubt is about the "1 GHz bandwidth/10 ps resolutio, 40
megasample/second digitizing rate" of the HP 54100D against the
"Bandwidth: 200 MHz, Sample Rate: 1 GS/s on each channel" of the Tek
TDS350. What is better?

What would you recommend me, guys??

Regards,

JaaC

I have a TDS3054B that is real neat and has served my needs so far. The fact
that it is so small and portable is very attractive.
I think this one had 5GS/s and 500 MHz bandwidth and 4 channels.
While you only use 2 channels for the most part, a 4 channel can come in
real handy at times.

Here is how I interpret the specs on these scopes...(the Agilent specs are a
little unclear to me)
Bandwidth refers to the analog bandwidth of the front-end. So if you have
100MHz clocks and you use a scope with a bandwidth of 75MHz, you are
probably distorting your clock. In this case, I'd feel comfortable with a
125MHz bandwidth or higher scope.
It's nice to have atleast a 10x factor oversampling at the highest bandwidth
(clock) you plan to use. So if my clocks won't exceed 100MHz, I'd be ok with
a 500MHz bandwidth with a 1GS/s spec (or 125MHz bandwidth and 1GS/s).
Clearly, more oversampling the better.
I'm not exactly a digital designer  - so treat some of my statements with
caution.

Cheers
Bhaskar


"Jaime Andres Aranguren Cardona" <jaime.aranguren@ieee.org> wrote in message
news:14a86f87.0403101441.726f2aa8@posting.google.com...
> Hi,
>
> Maybe this should be crossposted to other newsgroups, but I don't know
> which ones.
>
> I need your recommendation for buying a digital oscilloscope mainly
> for use with ADI DSPs (I mention that because of the SPORTs and link
> ports), and also general digital design. All of these instruments are
> to be bougth used, but functional.
>
> The choices are:
>
> - HP/Agilent 54100D
> *1 GHz bandwidth/10 ps resolution
> *40 megasample/second digitizing rate
> *10 bits vertical resolution with averaging
> *1k memory per channel
> *Two vertical input channels and two trigger inputs, each selectable
> for high, low, don't care.
> *In addition to triggering on entering or exiting the specified
> pattern, it can be set to trigger when a specified pattern is present
> for less than or greater than a specified duration, from 10 nS to 5
> Seconds.
> *Can be set to trigger on an edge of either polarity on any one of the
> four inputs.
> *Four HP 54002A pods included
>
> - Tektronix TDS350
> Signal Acquisition System
> * Bandwidth: 200 MHz
> * Sample Rate: 1 GS/s on each channel
> * Channels: two identical channels, each with invert function
> * Sensitivity: 2 mV to 10 V/div
> * Vertical Resolution: 8 bits (256 levels over 10.24 vertical
> divisions)
> Time Base System (Main and Delayed)
> * Time/Division Range: 2.5 ns - 5 s/div
> * Record Length: 1000 sample points per channel
> Triggering System (Main Only)
> * Triggering Types: Edge, Video
> Automatic Measurements
> Option 14 (I/O Interface): Computer Interface and Hardcopy Capability
> * GPIB (IEEE-488.2) Programmability
> * RS-232 Interface Programmability
> * Video Output
> * Hardcopy Port - Centronics-type parallel or RS-232
> * Graphics Interface File Formats - Interleaf (.img), TIFF, PCX (PC
> Paintbrush), BMP (Microsoft Windows), and Encapsulated Postscript
> (EPS)
> * Printer Formats - ThinkJet, DeskJet, LaserJet, Epson (9 & 24 Pin),
> Seiko DPU 411/II, DPU 412.  DC power provided for Seiko printers
>
> - Tektronix TDS2002
> * 60 MHz Bandwidth * Sample Rate 1 GS/s * 2 Channels * Color Display *
> CompactFlash Mass Storage Option with TDS2MEM Module * Autoset Menu
> with Waveform Selection * Probe Check Wizard To Ensure Correct Probe
> Usage * Context-sensitive Help * Dual Time Base * Advanced Triggers,
> including Pulse Width Trigger and Line-Selectable Video Trigger * 11
> Automatic Measurements * Multi-language User Interface * Waveform and
> Setup Memories * FFT Standard on All Models
>
> My main doubt is about the "1 GHz bandwidth/10 ps resolutio, 40
> megasample/second digitizing rate" of the HP 54100D against the
> "Bandwidth: 200 MHz, Sample Rate: 1 GS/s on each channel" of the Tek
> TDS350. What is better?
>
> What would you recommend me, guys??
>
> Regards,
>
> JaaC

"Bhaskar Thiagarajan" <bhaskart@deja.com> wrote in message news:<c2o90l$1v3dip$1@ID-82263.news.uni-berlin.de>...
> I have a TDS3054B that is real neat and has served my needs so far. The fact
> that it is so small and portable is very attractive.
> I think this one had 5GS/s and 500 MHz bandwidth and 4 channels.
> While you only use 2 channels for the most part, a 4 channel can come in
> real handy at times.
> 
> Here is how I interpret the specs on these scopes...(the Agilent specs are a
> little unclear to me)
> Bandwidth refers to the analog bandwidth of the front-end. So if you have
> 100MHz clocks and you use a scope with a bandwidth of 75MHz, you are
> probably distorting your clock. In this case, I'd feel comfortable with a
> 125MHz bandwidth or higher scope.
> It's nice to have atleast a 10x factor oversampling at the highest bandwidth
> (clock) you plan to use. So if my clocks won't exceed 100MHz, I'd be ok with
> a 500MHz bandwidth with a 1GS/s spec (or 125MHz bandwidth and 1GS/s).
> Clearly, more oversampling the better.
> I'm not exactly a digital designer  - so treat some of my statements with
> caution.
> 
> Cheers
> Bhaskar

Hello everybody. Bhaskar, I really appreciate your comments. I've been
reading more documentation abou the three (HP/Agilent 54100D, Tek
TDS350 and Tek TDS2002). Something to discuss is the following:

TDS2002's analog bandwidth is up to 60 MHz (drops to 6Mhz with 1x
probe), TDS350's is >= 200MHz, and 54100's is the most strange: 1GHz
for repetitive signals, but ridiculous 4MHz for single shot signals.

TDS350 has pulse response, peak detect and envelope modes of
operation. TDS2002 has sampling, peak detect and average modes of
operation. HP54100D has variable and infinite persistance, averaging
and magnify functions,  plus instant hardcopy and disk support. What
do you normally find useful? Are the two trigger inputs of the 54100D
really interesting for such an application?

The 54100D's digitizing rate is 40MS/s, whereas TDS350 and TDS2002 is
1GS/s. Taking this into account, and the bandwidths mentioned in the a
paragraph above, what is more useful/recommended for digital design,
mainly based in DSPs, ranging from 218/9x thru Blackfin thru SHARC
thru TigerSHARC?

One document I found in the web about the HP54100D says: "The HP54100
unites a powerful 1-GHz bandwidth with a random repetitive sampling
technique for viewing rarely ocurring narrow waveforms. This
oscilloscope has 0.002 percent time base accuracy and 10 ps resolution
for confident measurements of critical timing parameters in high speed
circuitry".

So, is it "ideal" for a purpose like mine? What seems strange to me,
and a bit confusing, is it's low sampling rate (40MS/s vs. 1GS/s), but
huge analog bandwidth (1GHz vs. 200MHz vs. 60MHz)

How does the 2.5k record length of the TDS2002 affects the
instrument's "usability" (for digital design) against the 1k of the
TDS350 and HP54100D?

Also to take into account are the prices (remember: buying used)

- TDS2002: around US$290
- TDS350: around US$380
- HP54100D: around US$200

I need your advice, guys, coming from the experience. What do you
recommend me?

Regards,

JaaC

> 
> 
> "Jaime Andres Aranguren Cardona" <jaime.aranguren@ieee.org> wrote in message
> news:14a86f87.0403101441.726f2aa8@posting.google.com...
> > Hi,
> >
> > Maybe this should be crossposted to other newsgroups, but I don't know
> > which ones.
> >
> > I need your recommendation for buying a digital oscilloscope mainly
> > for use with ADI DSPs (I mention that because of the SPORTs and link
> > ports), and also general digital design. All of these instruments are
> > to be bougth used, but functional.
> >
> > The choices are:
> >
> > - HP/Agilent 54100D
> > *1 GHz bandwidth/10 ps resolution
> > *40 megasample/second digitizing rate
> > *10 bits vertical resolution with averaging
> > *1k memory per channel
> > *Two vertical input channels and two trigger inputs, each selectable
> > for high, low, don't care.
> > *In addition to triggering on entering or exiting the specified
> > pattern, it can be set to trigger when a specified pattern is present
> > for less than or greater than a specified duration, from 10 nS to 5
> > Seconds.
> > *Can be set to trigger on an edge of either polarity on any one of the
> > four inputs.
> > *Four HP 54002A pods included
> >
> > - Tektronix TDS350
> > Signal Acquisition System
> > * Bandwidth: 200 MHz
> > * Sample Rate: 1 GS/s on each channel
> > * Channels: two identical channels, each with invert function
> > * Sensitivity: 2 mV to 10 V/div
> > * Vertical Resolution: 8 bits (256 levels over 10.24 vertical
> > divisions)
> > Time Base System (Main and Delayed)
> > * Time/Division Range: 2.5 ns - 5 s/div
> > * Record Length: 1000 sample points per channel
> > Triggering System (Main Only)
> > * Triggering Types: Edge, Video
> > Automatic Measurements
> > Option 14 (I/O Interface): Computer Interface and Hardcopy Capability
> > * GPIB (IEEE-488.2) Programmability
> > * RS-232 Interface Programmability
> > * Video Output
> > * Hardcopy Port - Centronics-type parallel or RS-232
> > * Graphics Interface File Formats - Interleaf (.img), TIFF, PCX (PC
> > Paintbrush), BMP (Microsoft Windows), and Encapsulated Postscript
> > (EPS)
> > * Printer Formats - ThinkJet, DeskJet, LaserJet, Epson (9 & 24 Pin),
> > Seiko DPU 411/II, DPU 412.  DC power provided for Seiko printers
> >
> > - Tektronix TDS2002
> > * 60 MHz Bandwidth * Sample Rate 1 GS/s * 2 Channels * Color Display *
> > CompactFlash Mass Storage Option with TDS2MEM Module * Autoset Menu
> > with Waveform Selection * Probe Check Wizard To Ensure Correct Probe
> > Usage * Context-sensitive Help * Dual Time Base * Advanced Triggers,
> > including Pulse Width Trigger and Line-Selectable Video Trigger * 11
> > Automatic Measurements * Multi-language User Interface * Waveform and
> > Setup Memories * FFT Standard on All Models
> >
> > My main doubt is about the "1 GHz bandwidth/10 ps resolutio, 40
> > megasample/second digitizing rate" of the HP 54100D against the
> > "Bandwidth: 200 MHz, Sample Rate: 1 GS/s on each channel" of the Tek
> > TDS350. What is better?
> >
> > What would you recommend me, guys??
> >
> > Regards,
> >
> > JaaC

> 
> Here is how I interpret the specs on these scopes...(the Agilent specs are a
> little unclear to me)
> Bandwidth refers to the analog bandwidth of the front-end. So if you have
> 100MHz clocks and you use a scope with a bandwidth of 75MHz, you are
> probably distorting your clock. In this case, I'd feel comfortable with a
> 125MHz bandwidth or higher scope.

That's clear. So... maybe the TDS2002 is not the _best_ option.

> It's nice to have atleast a 10x factor oversampling at the highest bandwidth
> (clock) you plan to use. 

This is certainly the point where I need some more explanation. The
TDS2002 has a great oversampling factor, or Sampling Rate/Analog
Bandwidth, so to speak (1GS_s/60MHz). That factor for the TDS350 is
also good (1GS_s/200MHz). But for the HP54100D it is completely the
inverse: 40MS_s/1GHz !!! Could someone explain me why? It seems to me
that they are targeted at different apps. Maybe the TDS2002 and TDS350
are more targeted to digital design, and the HP54100D is more targeted
to analog-domain characterization of digital circuits. Am I wrong? I
need your help!!

And then, what to say about these two other (new) choices:

- LeCroy 9310AM: 2ch, 400MHz BW, 100MS/s, 10GS/s equivalent sampling,
storage 200k points/ch.

- LeCroy 9420: 2ch, 350MHz BW, 100MS/s, 10GS/s equivalent sampling,
storage 50k points/ch.

What whould those two scopes (LeCroy) be targeted at, then? What to
choose?

> So if my clocks won't exceed 100MHz, I'd be ok with
> a 500MHz bandwidth with a 1GS/s spec (or 125MHz bandwidth and 1GS/s).
> Clearly, more oversampling the better.
> I'm not exactly a digital designer  - so treat some of my statements with
> caution.
> 
> Cheers
> Bhaskar
> 
>

Hi JaaC,

I use a Tektronix TDS754D, which is a "Digital Phosphor 
Oscilloscope". For those who don't know: its DPO mode provides a 
colorized display, where signals are accumulated over finite or 
infinite time.

I like this mode very much and wouldn't want to miss it.
If it comes to evaluate things like jitter or dynamic effects on a 
signal, this is very helpful.

An example: I trigger a peripheral bit on entry and exit of an 
interrupt service routine. With the scope on that pin, I can 
monitor the interrupt and see any deviation of the normal. 

And I appreciate the different sampling modes where I can setup 
either for high dynamic range or for very fast signals.

Both features have been useful.

When I had to choose, I preferred scopes which have knobs like the 
old analog scopes had, especially to switch vertical and horizontal 
scales. Their handling is simply easier and faster.

Bernhard

"Jaime Andres Aranguren Cardona" <jaime.aranguren@ieee.org> wrote in
message news:14a86f87.0403101930.4e092c24@posting.google.com...
>
> TDS2002's analog bandwidth is up to 60 MHz (drops to 6Mhz with 1x
> probe), TDS350's is >= 200MHz, and 54100's is the most strange: 1GHz
> for repetitive signals, but ridiculous 4MHz for single shot signals.
>
> The 54100D's digitizing rate is 40MS/s, whereas TDS350 and TDS2002
is
> 1GS/s. Taking this into account, and the bandwidths mentioned in the
a
> paragraph above, what is more useful/recommended for digital design,
> mainly based in DSPs, ranging from 218/9x thru Blackfin thru SHARC
> thru TigerSHARC?
>
> One document I found in the web about the HP54100D says: "The
HP54100
> unites a powerful 1-GHz bandwidth with a random repetitive sampling
> technique for viewing rarely ocurring narrow waveforms. This
> oscilloscope has 0.002 percent time base accuracy and 10 ps
resolution
> for confident measurements of critical timing parameters in high
speed
> circuitry".
>
> So, is it "ideal" for a purpose like mine? What seems strange to me,
> and a bit confusing, is it's low sampling rate (40MS/s vs. 1GS/s),
but
> huge analog bandwidth (1GHz vs. 200MHz vs. 60MHz)
>
<snip>
> I need your advice, guys, coming from the experience. What do you
> recommend me?
>
> Regards,
>
> JaaC
>
I'll try and explain the HP54100D specs.

The front end of the 54100 has a sample and hold that is _very_ fast,
together with trigger circuitry that is good for up to 1GHz signals.
The following ADC only samples at 40MHz rates.

For non-repetitive signals, Nyquist dictates that anything above 20MHz
will be aliased by the ADC, so the spec says for that mode _usable_
bandwidth of 4MHz (a substantial factor below Nyquist, so the
waveforms
are "clean"). There's lots more bandwidth there, just not usable.

If the signals are repetitive, i.e. you can get a stable trigger, then
the scope
can locate the input samples in time to 10psec with respect to the
trigger
point. That allows the scope to slowly (at 40MHz rate) build up the
waveform on the screen. In this mode the full S/H bandwidth can be
used
without aliasing, and you get 1GHz bandwidth.

Regards
    Ian

Jaime Andres Aranguren Cardona wrote:

   ...

> This is certainly the point where I need some more explanation. The
> TDS2002 has a great oversampling factor, or Sampling Rate/Analog
> Bandwidth, so to speak (1GS_s/60MHz). That factor for the TDS350 is
> also good (1GS_s/200MHz). But for the HP54100D it is completely the
> inverse: 40MS_s/1GHz !!! Could someone explain me why? It seems to me
> that they are targeted at different apps. Maybe the TDS2002 and TDS350
> are more targeted to digital design, and the HP54100D is more targeted
> to analog-domain characterization of digital circuits. Am I wrong? I
> need your help!!

It sounds to me like a sampling scope. (Don't jump. Of course, all
digital scopes sample, but I mean a technique used back in the 60s with
analog scopes.) Google on "sampling scope". If that doesn't help, the
inventor was Juan Amodei. Maybe you can find the patent, assigned to RCA
(and sold to Tektronix). At worst, I'll try to explain it without
pictures.

   ...

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Ian Buckner wrote:

... an excellent description of a sampling scope. Knowing how it works
with an analog scope without a storage tube, I was tongue tied when
wanting to explain it. Gosh, those ancients were clever!

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Ian Buckner wrote:
> I'll try and explain the HP54100D specs.
> 
> The front end of the 54100 has a sample and hold that is _very_ fast,
> together with trigger circuitry that is good for up to 1GHz signals.
> The following ADC only samples at 40MHz rates.
> 
> For non-repetitive signals, Nyquist dictates that anything above 20MHz
> will be aliased by the ADC, so the spec says for that mode _usable_
> bandwidth of 4MHz (a substantial factor below Nyquist, so the
> waveforms
> are "clean"). There's lots more bandwidth there, just not usable.
> 
> If the signals are repetitive, i.e. you can get a stable trigger, then
> the scope
> can locate the input samples in time to 10psec with respect to the
> trigger
> point. That allows the scope to slowly (at 40MHz rate) build up the
> waveform on the screen. In this mode the full S/H bandwidth can be
> used
> without aliasing, and you get 1GHz bandwidth.


Excellent explanation.

For this reason, the scope is great for analog work, and not to my 
liking for digital.  That is because digital involves lots of 
single-shot events that you want to be able to capture.  So I'd much 
rather have a sampling rate > bandwidth scope for digital.

I'd go for the TDS350 to get the maximum bandwidth of the fast sampling 
scopes (200MHz > 60MHz) and live with the limited memory until I could 
afford an upgrade to a deeper memory scope.

Good day!


-- 
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov

Hi Jerry,

Jerry Avins <jya@ieee.org> wrote in message news:<40506f7e$0$3103$61fed72c@news.rcn.com>...
> Ian Buckner wrote:
> 
> ... an excellent description of a sampling scope. Knowing how it works
> with an analog scope without a storage tube, I was tongue tied when
> wanting to explain it. Gosh, those ancients were clever!

Its easy to explain in two words - temporal dithering.

Most of the dithering people talk about here is amplitude dithering.
In image processing spacial dithering is more common. In fast capture,
temporal dithering has seen widespread use for many years. I think
HP's first digital scope was of this type, before they made the more
common continuous capture type. HP also introduced timing analysers in
the 80's, using a similar underlying technique. A number of defence
applications have also used this - in fact, when I first saw HP's
presentation about their dithering timing analyser, I thought it
looked a lot like a piece of defence kit I knew about. :-)

Regards,
Steve