Maker Pro
Maker Pro

Tek 2400 series digital scope usability question

H

Hawker

Jan 1, 1970
0
I spent this morning trying to track down a clock problem only to find
out that it was the scope and not the clock. Wasted morning.

I was looking at a 24MHz clock. The good scope (a TDS520B) was being
used so I figured for only 24MHz I could use the Tek 2430A (a 150Mhz,
100M/s scope. The clock looked like a sine wave (not square) had bad -
visible, jitter and seemed to be amplitude modulating at some lowish
frequency. Three issues. Looking at it with my trusty 2430A (also a
150Mhz but analog scope) and TDS520B it looked as it should.

I'm wondering if the issue was solely to low a sample rate (only 4
samples per clock). I had hoped to upgrade the TDS scope and get rid of
the 2430A but I can't afford what I want now. So I was thinking of
getting a 2440 (300Mhz, 500m/s ) but if I can't look at a 50MHz square
wave accurately with it then it won't do me much good. That would be 10
samples per clock at 50Mhz. Since there were three issues with looking
at it with the 2430A I wonder if all three can be explained and will not
be issues with the 2440 at 2x the frequency and 5x the sample rate.

Is this going to work or will I be disappointed?

Thanx
Hawker
 
J

Joerg

Jan 1, 1970
0
Hawker said:
I spent this morning trying to track down a clock problem only to find
out that it was the scope and not the clock. Wasted morning.

I was looking at a 24MHz clock. The good scope (a TDS520B) was being
used so I figured for only 24MHz I could use the Tek 2430A (a 150Mhz,
100M/s scope. The clock looked like a sine wave (not square) had bad -
visible, jitter and seemed to be amplitude modulating at some lowish
frequency. Three issues. Looking at it with my trusty 2430A (also a
150Mhz but analog scope) and TDS520B it looked as it should.

I'm wondering if the issue was solely to low a sample rate (only 4
samples per clock). I had hoped to upgrade the TDS scope and get rid of
the 2430A but I can't afford what I want now. So I was thinking of
getting a 2440 (300Mhz, 500m/s ) but if I can't look at a 50MHz square
wave accurately with it then it won't do me much good. That would be 10
samples per clock at 50Mhz. Since there were three issues with looking
at it with the 2430A I wonder if all three can be explained and will not
be issues with the 2440 at 2x the frequency and 5x the sample rate.

Is this going to work or will I be disappointed?

To gauge the quality of a 50MHZ clock you'll need 1GSPS, at least.
Better 2GSPS. Repetitive scans at any lower rate plus piecing together
isn't going to work when you suspect a noisy clock. About a month ago I
had to look at laser diode noise, became frustrated with their digital
scopes and convinced that client to buy a used Tek 2465. One week and
about $350 later that allowed us to get the job done ;-)

Another downside of digital scopes is their rather limited vertical
resolution or effective number of bits.
 
P

PeterD

Jan 1, 1970
0
I spent this morning trying to track down a clock problem only to find
out that it was the scope and not the clock. Wasted morning.

I was looking at a 24MHz clock. The good scope (a TDS520B) was being
used so I figured for only 24MHz I could use the Tek 2430A (a 150Mhz,
100M/s scope.

IMHO, a 100 M/s scope won't give you a very good indication of a 25
Mhz waveform, especially a square wave. In fact, I'd expect it to look
almost exactly like the way you described it! <g>

Remember, a square wave consists of two frequencies, the slopes
(leading edge, and trailing edge are very high frequency) while the
middle part between the slopes constitutes a low frequency.
 
D

David L. Jones

Jan 1, 1970
0
I spent this morning trying to track down a clock problem only to find
out that it was the scope and not the clock. Wasted morning.

I was looking at a 24MHz clock. The good scope (a TDS520B) was being
used so I figured for only 24MHz I could use the Tek 2430A (a 150Mhz,
100M/s scope. The clock looked like a sine wave (not square) had bad -
visible, jitter and seemed to be amplitude modulating at some lowish
frequency. Three issues. Looking at it with my trusty 2430A (also a
150Mhz but analog scope) and TDS520B it looked as it should.

I'm wondering if the issue was solely to low a sample rate (only 4
samples per clock). I had hoped to upgrade the TDS scope and get rid of
the 2430A but I can't afford what I want now. So I was thinking of
getting a 2440 (300Mhz, 500m/s ) but if I can't look at a 50MHz square
wave accurately with it then it won't do me much good. That would be 10
samples per clock at 50Mhz. Since there were three issues with looking
at it with the 2430A I wonder if all three can be explained and will not
be issues with the 2440 at 2x the frequency and 5x the sample rate.

Is this going to work or will I be disappointed?

Thanx
Hawker

Digital scopes with low sample rates like this have to use repetitive
sampling techniques and not real-time sampling for high frequencies.
This can be very troublesome and cause lots of issues, and is why
almost all new DSO's are "real-time" sampling that have a sample rate
at least 10 times the analog bandwidth.

If your trigger point is stable then your 100MS/s 150MHz DSO *should*
be able to display a stable *repetitive* signal at up to the 150MHz
bandwidth. But once you single shot capture it, it is using real-time
mode and you only get your 4 samples per cyle for a 25MHz signal.

A 500MS/s scope will be *just* ok with a 50MHz signal, as 10 times
minimum is the general rule of thumb. But you will most likley be a
bit dissappointed with the 500MS/s, I would be aiming at a 1GS/s scope
fro this kind of work. Even the cheapest sub $1000 entry level scope
can do 1GS/s, albeit with limited sample memory size (a few KB) and
analog bandwidth.

Dave.
 
D

David L. Jones

Jan 1, 1970
0
IMHO, a 100 M/s scope won't give you a very good indication of a 25
Mhz waveform, especially a square wave. In fact, I'd expect it to look
almost exactly like the way you described it! <g>

Not if the trigger point is stable and you aren't using single shot
capture, it will work fine in repetitive mode. That is why
manufacturers actually make 150MHz bandwidth scopes with 100MS/s
sampling, if it didn't work they would make the bandwith only 50MHz or
so at best. They also make 20GHz bandwidth scopes with 100MS/s
sampling etc.

Dave.
 
J

Joerg

Jan 1, 1970
0
PeterD said:
IMHO, a 100 M/s scope won't give you a very good indication of a 25
Mhz waveform, especially a square wave. In fact, I'd expect it to look
almost exactly like the way you described it! <g>

Remember, a square wave consists of two frequencies, the slopes
(leading edge, and trailing edge are very high frequency) while the
middle part between the slopes constitutes a low frequency.


It consists of a whole lot more frequencies. If you can't acquire up to
the 7th-9th harmonic you won't get much of a display to write home about.
 
J

Joerg

Jan 1, 1970
0
David said:
Digital scopes with low sample rates like this have to use repetitive
sampling techniques and not real-time sampling for high frequencies.
This can be very troublesome and cause lots of issues, and is why
almost all new DSO's are "real-time" sampling that have a sample rate
at least 10 times the analog bandwidth.

If your trigger point is stable then your 100MS/s 150MHz DSO *should*
be able to display a stable *repetitive* signal at up to the 150MHz
bandwidth. But once you single shot capture it, it is using real-time
mode and you only get your 4 samples per cyle for a 25MHz signal.

"Troublesome" and "stable repetitive" are the right words. If there was
indeed a clock problem that can easily turn the displayed trace of a
repetitive scan into pulp.
 
J

john jardine

Jan 1, 1970
0
Hawker said:
I spent this morning trying to track down a clock problem only to find
out that it was the scope and not the clock. Wasted morning.

I was looking at a 24MHz clock. The good scope (a TDS520B) was being
used so I figured for only 24MHz I could use the Tek 2430A (a 150Mhz,
100M/s scope. The clock looked like a sine wave (not square) had bad -
visible, jitter and seemed to be amplitude modulating at some lowish
frequency. Three issues. Looking at it with my trusty 2430A (also a
150Mhz but analog scope) and TDS520B it looked as it should.

I'm wondering if the issue was solely to low a sample rate (only 4
samples per clock). I had hoped to upgrade the TDS scope and get rid of
the 2430A but I can't afford what I want now. So I was thinking of
getting a 2440 (300Mhz, 500m/s ) but if I can't look at a 50MHz square
wave accurately with it then it won't do me much good. That would be 10
samples per clock at 50Mhz. Since there were three issues with looking
at it with the 2430A I wonder if all three can be explained and will not
be issues with the 2440 at 2x the frequency and 5x the sample rate.

Is this going to work or will I be disappointed?

Thanx
Hawker
Wouldn't trust 'em as even a door stop.
Last week I used a customer's expensive 100's of MHz TDS to look at a slow
synching pulse alongside a 50Hz 5V sawtooth. At 100ms/div noticed bad
distortion on the sawtooth. Did not waste time investigating as previous
experience said "suspect the digital scope first". At 5secs/div my sawtooth
magically reappeared in all it's glory but now apparently .05Hz. Damned
waveform had been aliasing all the way down and was worthless. (I won't
mention the hit and miss display of the synching pulse)
Nice coloured traces.
Nice printouts for those important management presentations and emails.
Nice FFT display.
Small and light.

Tiny, lo-res, pixelly display. May as well be looking at a ZX81 screen. Kind
of "is that my waveform or a display artefact"?.
Unable to trace low duty cycles.
Continual "am I seeing the real waveform or some alias"? and am I really
generating all that noise?.
As a technology demonstrator marvellous, as a 'scope f****** worthless.
 
D

David L. Jones

Jan 1, 1970
0
"Troublesome" and "stable repetitive" are the right words. If there was
indeed a clock problem that can easily turn the displayed trace of a
repetitive scan into pulp.

For sure.
You get lots of nice random dots on the screen though, and it's kinda
pretty to look at :->

Dave.
 
J

Joerg

Jan 1, 1970
0
David said:
For sure.
You get lots of nice random dots on the screen though, and it's kinda
pretty to look at :->

Younger users may like to push the obfuscation button which collects
lots of those random dots and makes it all look like graffiti :)))
 
H

Hawker

Jan 1, 1970
0
Digital scopes with low sample rates like this have to use repetitive
sampling techniques and not real-time sampling for high frequencies.
This can be very troublesome and cause lots of issues, and is why
almost all new DSO's are "real-time" sampling that have a sample rate
at least 10 times the analog bandwidth.

If your trigger point is stable then your 100MS/s 150MHz DSO *should*
be able to display a stable *repetitive* signal at up to the 150MHz
bandwidth. But once you single shot capture it, it is using real-time
mode and you only get your 4 samples per cyle for a 25MHz signal.

A 500MS/s scope will be *just* ok with a 50MHz signal, as 10 times
minimum is the general rule of thumb. But you will most likley be a
bit dissappointed with the 500MS/s, I would be aiming at a 1GS/s scope
fro this kind of work. Even the cheapest sub $1000 entry level scope
can do 1GS/s, albeit with limited sample memory size (a few KB) and
analog bandwidth.

Thanx.
That is the information I thought, but didn't want to hear.
The reality is my 500Mhz 1 G/S TDS520B (I think) is starting to show
it's age. There are signals I need to see but it can't show very well
(esp outputs from CCDs in one sweep). I had hoped to make it my second
scope, get rid of the 2430A and get a new scope (TDS 600 or TDS700
series?) but looks like I need to shell out $4k-$10k for one of those
and I don't have it now - probably not till the end of the year. I may
get a 2440 just to get me by for now. I assume it will hold it's value a
bit. I looked at TDS300 and TDS400 stuff but other than active FET
probes they actually look like a step down from a 2440.

Anyone have a suggestion for a basic feature, active FET capable scope
with 500Mhz Bandwidth and 2 G/S (perhaps 4-5GS?) resolution? No fancy
features, although FFT would be handy. Step up from a TDS520B. I assume
I'm looking at used but if the price is right I could consider new.

Hawker
 
J

Joerg

Jan 1, 1970
0
Hawker said:
Thanx.
That is the information I thought, but didn't want to hear.
The reality is my 500Mhz 1 G/S TDS520B (I think) is starting to show
it's age. There are signals I need to see but it can't show very well
(esp outputs from CCDs in one sweep). I had hoped to make it my second
scope, get rid of the 2430A and get a new scope (TDS 600 or TDS700
series?) but looks like I need to shell out $4k-$10k for one of those
and I don't have it now - probably not till the end of the year. I may
get a 2440 just to get me by for now. I assume it will hold it's value a
bit. I looked at TDS300 and TDS400 stuff but other than active FET
probes they actually look like a step down from a 2440.

Anyone have a suggestion for a basic feature, active FET capable scope
with 500Mhz Bandwidth and 2 G/S (perhaps 4-5GS?) resolution? No fancy
features, although FFT would be handy. Step up from a TDS520B. I assume
I'm looking at used but if the price is right I could consider new.

Quite frankly I have never found the FFT feature on "modern" scopes that
useful. If it was me I'd go for a used 2465 plus a nice 1GHz+ spectrum
analyzer for any spectral investigation. Should be able to obtain that
for under $2k total if you don't mind boat anchor size. That, plus a
simple digital scope if you need to be able to see pre-trigger.
 
D

David L. Jones

Jan 1, 1970
0
Thanx.
That is the information I thought, but didn't want to hear.
The reality is my 500Mhz 1 G/S TDS520B (I think) is starting to show
it's age. There are signals I need to see but it can't show very well
(esp outputs from CCDs in one sweep). I had hoped to make it my second
scope, get rid of the 2430A and get a new scope (TDS 600 or TDS700
series?) but looks like I need to shell out $4k-$10k for one of those
and I don't have it now - probably not till the end of the year. I may
get a 2440 just to get me by for now. I assume it will hold it's value a
bit. I looked at TDS300 and TDS400 stuff but other than active FET
probes they actually look like a step down from a 2440.

Anyone have a suggestion for a basic feature, active FET capable scope
with 500Mhz Bandwidth and 2 G/S (perhaps 4-5GS?) resolution? No fancy
features, although FFT would be handy. Step up from a TDS520B. I assume
I'm looking at used but if the price is right I could consider new.

Hawker

For this performance you won't get new scope for under around $4K or
so.
But there are plenty of used Lecroy's on Ebay with 500-1GHz bandwidth,
2GS/s+, and massively deep memories.
They are a dog to drive, but the performance is there, and the price
is right at under $2K
This is but one example:
http://cgi.ebay.com.au/LeCroy-9374L...89494673QQcategoryZ104247QQrdZ1QQcmdZViewItem

Dave.
 
J

Jim Yanik

Jan 1, 1970
0
I have a 2445 and access to a 2465b but honestly don't use it much.
When I need high speed analog I need FET probes which the 2400 series
stuff does not have. But more often then not I need to do fancy
triggering and holding (trigger on an event and hold the capture) that
can only be done with a DSO. The analog stuff gets some usage but not
much anymore. I hate to admit it because for what it can do it usually
is better than any digital scope.

you can use a FET probe with a 2400 scope;you just have to power the probe
externally.TEK used to make an external supply expressly for this.
 
A

Anthony Fremont

Jan 1, 1970
0
john said:
Wouldn't trust 'em as even a door stop.
Last week I used a customer's expensive 100's of MHz TDS to look at a
slow synching pulse alongside a 50Hz 5V sawtooth. At 100ms/div
noticed bad distortion on the sawtooth. Did not waste time
investigating as previous experience said "suspect the digital scope
first". At 5secs/div my sawtooth magically reappeared in all it's
glory but now apparently .05Hz. Damned waveform had been aliasing
all the way down and was worthless. (I won't mention the hit and miss
display of the synching pulse)
Nice coloured traces.
Nice printouts for those important management presentations and
emails. Nice FFT display.
Small and light.

Tiny, lo-res, pixelly display. May as well be looking at a ZX81
screen. Kind of "is that my waveform or a display artefact"?.
Unable to trace low duty cycles.
Continual "am I seeing the real waveform or some alias"? and am I
really generating all that noise?.
As a technology demonstrator marvellous, as a 'scope f******
worthless.

I just got my first DSO. Honestly so far I'm really impressed compared to
my old Hitach V650-F. It's certainly allot easier to see what happened
before the trigger event. ;-)

I also got my first taste of aliasing too, kinda scarey. I was watching a
PWM that updates ~80 times/second on channel 1 and put the CLKOUT from the
PIC onto channel 2. CLKOUT was 2MHz. Instead of seeing the nice thick
"bar" across the screen as I would expect on an analog scope, I could see
the clock signal and it looked just like the clock signal, only running
wayyyyyyyyyy tooooo slowwww. That's the part that really surprised me: It
looked EXACTLY like the CLKOUT signal including the little voltage spikes.
:-O

OTOH, it's the most amazing thing I've ever used. :)
 
H

Hawker

Jan 1, 1970
0
Quite frankly I have never found the FFT feature on "modern" scopes that
useful. If it was me I'd go for a used 2465 plus a nice 1GHz+ spectrum
analyzer for any spectral investigation. Should be able to obtain that
for under $2k total if you don't mind boat anchor size. That, plus a
simple digital scope if you need to be able to see pre-trigger.

I have a 2445 and access to a 2465b but honestly don't use it much.
When I need high speed analog I need FET probes which the 2400 series
stuff does not have. But more often then not I need to do fancy
triggering and holding (trigger on an event and hold the capture) that
can only be done with a DSO. The analog stuff gets some usage but not
much anymore. I hate to admit it because for what it can do it usually
is better than any digital scope.
 
J

Joerg

Jan 1, 1970
0
Hawker said:
I have a 2445 and access to a 2465b but honestly don't use it much.
When I need high speed analog I need FET probes which the 2400 series
stuff does not have. But more often then not I need to do fancy
triggering and holding (trigger on an event and hold the capture) that
can only be done with a DSO. The analog stuff gets some usage but not
much anymore. I hate to admit it because for what it can do it usually
is better than any digital scope.


Ok, depending on your typical design a DSO might indeed be the better
choice. Definitely when chasing digital state changes where you don't
want to crack out the logic analzyer.

As Jim said a FET probe can be powered externally. I use a Philips FET
probe on a Tek scope and they never even growled at each other.
 
J

john jardine

Jan 1, 1970
0
Anthony Fremont said:
I just got my first DSO. Honestly so far I'm really impressed compared to
my old Hitach V650-F. It's certainly allot easier to see what happened
before the trigger event. ;-)

I also got my first taste of aliasing too, kinda scarey. I was watching a
PWM that updates ~80 times/second on channel 1 and put the CLKOUT from the
PIC onto channel 2. CLKOUT was 2MHz. Instead of seeing the nice thick
"bar" across the screen as I would expect on an analog scope, I could see
the clock signal and it looked just like the clock signal, only running
wayyyyyyyyyy tooooo slowwww. That's the part that really surprised me: It
looked EXACTLY like the CLKOUT signal including the little voltage spikes.
:-O

OTOH, it's the most amazing thing I've ever used. :)
Yes. Other than good sex, there's nothing finer than a new toy turning up in
the mail :)
My main scope is a straightforward 100MHz Hitachi V-1065A but look forward
to the day I can dump it in favour of some digital model. Probably be a wait
as they'll need to offer something realistic such as a 1THz sample rate at
10bits and a display of 1024x768 with screen update at 70-100HZ
john
 
L

Lionel

Jan 1, 1970
0
I'm wondering if the issue was solely to low a sample rate (only 4
samples per clock). I had hoped to upgrade the TDS scope and get rid of
the 2430A but I can't afford what I want now. So I was thinking of
getting a 2440 (300Mhz, 500m/s ) but if I can't look at a 50MHz square
wave accurately with it then it won't do me much good. That would be 10
samples per clock at 50Mhz.

Three samples per cycle is inevitably going to give you a horribly
distorted waveform. Ten samples per cycle won't be pretty, but should
be usable.
If you're pushing the sampling speed that way, an analog scope is
probably more useful.
 
Top