Why no delayed trigger on newer DSOs?

[Joerg]

Hello Folks,

Question: It seems that most portable DSOs in the 200MHz/1GSPS class
don't offer delayed trigger anymore which makes pulse-echo measurements
a real pain. Why don't they?

After all it's just a simple counter that is needed. Same for hold-off
which typically ain't there either :-(

Beats me but IIRC I used delayed trigger on a lower end TDS220 although
its data sheet said that it could not do it. However, that scope had
some other serious issues (noise) so I asked the client to obtain a used
2265. Old Faithful, of course, always has delayed trigger but sure
enough the plastic clutch for that function fell apart right away.
Designing a trigger delay into an analog scope is no small feat but for
a DSO it should be really easy.

 

[Tom Bruhns]

Hello Folks,

Question: It seems that most portable DSOs in the 200MHz/1GSPS class
don't offer delayed trigger anymore which makes pulse-echo measurements
a real pain. Why don't they?

After all it's just a simple counter that is needed. Same for hold-off
which typically ain't there either :-(

Beats me but IIRC I used delayed trigger on a lower end TDS220 although
its data sheet said that it could not do it. However, that scope had
some other serious issues (noise) so I asked the client to obtain a used
2265. Old Faithful, of course, always has delayed trigger but sure
enough the plastic clutch for that function fell apart right away.
Designing a trigger delay into an analog scope is no small feat but for
a DSO it should be really easy.

Is this a "when are you going to stop beating your mother" sort of
question? Check out the Agilent 5000 series of scopes, listed as
"new" on the Agilent web site today. Trigger delay range: pre-
trigger, the greater of 1 screen or 125usec; post-trigger 1 second
range minimum, up to 500 seconds for slower timebase settings.

On bottom-end scopes, especially ones with deep memory, trigger delay
may well be dropped for a combination of reasons: the desire to
differentiate the product line, and the lower need given lots of
memory--acquire once, inspect in detail wherever you want. One of my
DSOs doesn't have specific trigger delay, but does let you set up the
display of any portion of the memory, as a "delayed trace".

 

[Joerg]

Is this a "when are you going to stop beating your mother" sort of
question? Check out the Agilent 5000 series of scopes, listed as
"new" on the Agilent web site today. Trigger delay range: pre-
trigger, the greater of 1 screen or 125usec; post-trigger 1 second
range minimum, up to 500 seconds for slower timebase settings.

Ok, but now we are talking $5k class. I meant the lower end, around $2K
such as the TDS2024.

On bottom-end scopes, especially ones with deep memory, trigger delay
may well be dropped for a combination of reasons: the desire to
differentiate the product line, and the lower need given lots of
memory--acquire once, inspect in detail wherever you want. One of my
DSOs doesn't have specific trigger delay, but does let you set up the
display of any portion of the memory, as a "delayed trace".

Memory is a problem on some. The Tek mid-range offers 2.5K, Agilent
mid-range is 4K and mine (Instek) is 25K. This is only for single
channel use. So a pulse-echo measurement at 1nsec slivers becomes
impossible for anything above a few ten usec.

Leaving it out for differentiation reasons would be rather foolish
because it causes (and has caused in my past) potential users to be
hesitant and stick to their old analog scopes. Those offer delayed
trigger in rather basic models which can and are bought via EBay. There
has rarely been a client of mine where I didn't ask them to bid on a
2465, and none has ever regretted it. On the contrary, sometimes we
discovered that we needed more of these "real" scopes because everyone
suddenly wanted to have it.

 

[Mike Harrison]

Hello Folks,

Question: It seems that most portable DSOs in the 200MHz/1GSPS class
don't offer delayed trigger anymore which makes pulse-echo measurements
a real pain. Why don't they?

After all it's just a simple counter that is needed. Same for hold-off
which typically ain't there either :-(

Beats me but IIRC I used delayed trigger on a lower end TDS220 although
its data sheet said that it could not do it. However, that scope had
some other serious issues (noise) so I asked the client to obtain a used
2265. Old Faithful, of course, always has delayed trigger but sure
enough the plastic clutch for that function fell apart right away.
Designing a trigger delay into an analog scope is no small feat but for
a DSO it should be really easy.

Is it that it isn't there, or is no longer an explicit feature - you just scroll the display..?.
I find it hard to believe that any digital scope doesn't have holdoff.... can you name specific
models?

 

[Joerg]

Is it that it isn't there, or is no longer an explicit feature - you just scroll the display..?.
I find it hard to believe that any digital scope doesn't have holdoff.... can you name specific
models?

In my case it's the Instek GDS-2204 but when looking at competitor
models such as the Tektronix TDS-2024 it seems that it is a similar
situation there. John Larkin would know for sure because he's got one.

No hold-off. Here is the real puzzler: I've got the 4-ch model and the
less expensive 2-ch model does contain delayed trigger (!). Snooping
around in the firmware I saw that it's still in the control set but
somehow not engaged, IOW I can access delayed trigger but not tie it
into any of the channels which renders it useless. Plus it isn't
available via the scope's buttons, only if I enter via the doggie door
with HyperTerminal Wrote to Instek about it but no answer yet. Hopefully
they'll realize that disconnecting delayed trigger on their most
expensive flagship was not such a great idea, and maybe they'll fix that
in the firmware.

Scrolling the display gets you a few usec on Tek, and a few tens of usec
on the Instek. But even the 25K on mine will eventually run out,
precisely at 25usec when using the highest sample rate. Ok, on the Hameg
2008 with its Megabyte buffer I could understand this strategy but not
on the others. Even 25usec is nothing on pulse-echo jobs.

I do have to say that the advanced trigger features of these scopes are
really great. For example, this Instek can trigger on stuff like "only
when pulse exceeds x usec" and this feature goes to 10 seconds. But of
course then it waits for the slope so I can't hitch a delay-trigger ride
here. So they've got a long counter in there already and implementing a
delay feature should almost be a no-brainer.

 

[Winfield Hill]

In my case it's the Instek GDS-2204 but when looking at competitor
models such as the Tektronix TDS-2024 it seems that it is a similar
situation there. John Larkin would know for sure because he's got one.

No hold-off. Here is the real puzzler: I've got the 4-ch model and the
less expensive 2-ch model does contain delayed trigger (!).

Maybe you need a firmware update. Our TDS-2024B has holdoff.
What it doesn't have is A and B horizontal sweeping, with an
adjustable time delay before the B trigger is armed, like our
older TDS-3024B and TDS-3034B scopes. It's too bad, because
the new TDS-2024B is very sexy with its two USB plus ethernet
ports, etc. The older TDS-3034B does have the ethernet port.

 

[Joerg]

Maybe you need a firmware update. Our TDS-2024B has holdoff.
What it doesn't have is A and B horizontal sweeping, with an
adjustable time delay before the B trigger is armed, like our
older TDS-3024B and TDS-3034B scopes. It's too bad, because
the new TDS-2024B is very sexy with its two USB plus ethernet
ports, etc. The older TDS-3034B does have the ethernet port.

So the 2024B still doesn't have a delayed trigger then? I've got the
Instek GDS-2204, not the Tek, because I need the 25K of memory. Even
better would have been the analog/digital Hameg 2008 comboscope but it's
too large for carry-on if you have to fit some clothes in for overnight
stays, plus I am not sure how that classic CRT in there would fare
during really bad turbulence.

I am quite sure Instek will do firmware updates in the future and hope
they'll throw in a delayed trigger. It's just that I don't get it why
they took it out. Just like I don't understand why some CAD systems fail
to provide hierarchical sheet structures. Well, at least I brought it to
their attention that their marketing brochure promised delayed trigger
and the product then didn't have it. It can't be rocket science to put
it back in since the next lower model has it.

Other than that it's a nice scope, especially regarding USB
connectivity. One thing I want to try out when I find the time is
operating it out of Excel via VBA since they published the complete
command set for it. That would be cool. Their custom SW is ok for
catching CSV files and the like but otherwise it is very spartan.
Freezes a lot, too, but then again it is a beta version.

 

[Joel Kolstad]

It's just that I don't get it why they took it out. Just like I don't
understand why some CAD systems fail to provide hierarchical sheet
structures.

The usual reason is because the people implementing the tools don't actually
end up *using* the tools combined with management that doesn't have a strong
technical background. Hence, the only feedback path of what features are
desired is customers and the competitions!

 

[Joerg]

The usual reason is because the people implementing the tools don't actually
end up *using* the tools combined with management that doesn't have a strong
technical background. Hence, the only feedback path of what features are
desired is customers and the competitions!

^^^^^^^^^^^

Yep, and I did let off a couple remarks to that effect, hoping that this
might "accelerate" the firmware upgrade. If they are smart they'll turn
around and use people like me as testers. That's what we do in medical
devices all the time but somehow that concept catches on really slowly
in electronics. It's taken over half a century by now ...

 

[Rich Grise]

Hello Folks,

Question: It seems that most portable DSOs in the 200MHz/1GSPS class
don't offer delayed trigger anymore which makes pulse-echo measurements
a real pain. Why don't they?

After all it's just a simple counter that is needed. Same for hold-off
which typically ain't there either :-(

Beats me but IIRC I used delayed trigger on a lower end TDS220 although
its data sheet said that it could not do it. However, that scope had
some other serious issues (noise) so I asked the client to obtain a used
2265. Old Faithful, of course, always has delayed trigger but sure
enough the plastic clutch for that function fell apart right away.
Designing a trigger delay into an analog scope is no small feat but for
a DSO it should be really easy.

But, if it's a digital storage scope, why can't you just trigger it at
the beginning of the test and "zoom in" to the part you're interested
in seeing delayed?

Good Luck!
Rich

 

[Joerg]

But, if it's a digital storage scope, why can't you just trigger it at
the beginning of the test and "zoom in" to the part you're interested
in seeing delayed?

Because the Instek runs out of memory after 25usec. Similarly priced HPs
run out in under 5usec and the Teks run out even earlier :-(

The 2.5K of memory on a Tek-2000 series is IMHO really pathetic. I
wanted to buy one but when I saw that I didn't. If you do pulse-echo
stuff you can't use that.

 

[Tony Williams]

Because the Instek runs out of memory after 25usec. Similarly
priced HPs run out in under 5usec and the Teks run out even
earlier :-(

The 2.5K of memory on a Tek-2000 series is IMHO really pathetic.
I wanted to buy one but when I saw that I didn't. If you do
pulse-echo stuff you can't use that.

The Agilent DSO5000-series can have deep memory (1M)
and delayed trigger.

 

[Joerg]

The Agilent DSO5000-series can have deep memory (1M)
and delayed trigger.

I know, but then you are talking $5k and up. I mean, heck, even the old
Tek 2465 has delayed trigger. Bidding ist pretty active on EBay and Tek
is cutting into their own sales by not providing delayed trigger in
"modern" scopes. Because people like me tell clients to bid on a 2465.
And then they do. From a marketing point of view it makes absolutely no
sense.

 

[Fred Bartoli]

Le Wed, 15 Aug 2007 07:13:09 +0000, Joerg a écrit:

Because the Instek runs out of memory after 25usec. Similarly priced HPs
run out in under 5usec and the Teks run out even earlier :-(

Right now, I have to do some really accurate compensation to a system's
step response. From the 10ns scale, down to the 100ms scale. What would
really be handy is a logarithmic sweep. That could easily be implemented
with say a sampling rate scaling down by a two factor while time rises by
a two factor.
Well, with GPIB and some post-processing it works, but it's much longer
and a bit... bulky.

 

[David L. Jones]

In my case it's the Instek GDS-2204 but when looking at competitor
models such as the Tektronix TDS-2024 it seems that it is a similar
situation there. John Larkin would know for sure because he's got one.

No hold-off. Here is the real puzzler: I've got the 4-ch model and the
less expensive 2-ch model does contain delayed trigger (!). Snooping
around in the firmware I saw that it's still in the control set but
somehow not engaged, IOW I can access delayed trigger but not tie it
into any of the channels which renders it useless. Plus it isn't
available via the scope's buttons, only if I enter via the doggie door
with HyperTerminal Wrote to Instek about it but no answer yet. Hopefully
they'll realize that disconnecting delayed trigger on their most
expensive flagship was not such a great idea, and maybe they'll fix that
in the firmware.

Scrolling the display gets you a few usec on Tek, and a few tens of usec
on the Instek. But even the 25K on mine will eventually run out,
precisely at 25usec when using the highest sample rate. Ok, on the Hameg
2008 with its Megabyte buffer I could understand this strategy but not
on the others. Even 25usec is nothing on pulse-echo jobs.

I do have to say that the advanced trigger features of these scopes are
really great. For example, this Instek can trigger on stuff like "only
when pulse exceeds x usec" and this feature goes to 10 seconds. But of
course then it waits for the slope so I can't hitch a delay-trigger ride
here. So they've got a long counter in there already and implementing a
delay feature should almost be a no-brainer.

The Instek GDS-820 has delayed trigger, but only operates on the
external trigger input!

Dave.

 

[[email protected]]

Question: It seems that most portable DSOs in the 200MHz/1GSPS class
don't offer delayed trigger anymore which makes pulse-echo measurements
a real pain. Why don't they?

Maybe the real question to ask: when is someone (almost certainly a
new party) going to come out with a customizeable DSO "kit".

Basically, I'm thinking of something that is workable with the
shipping "reference design" but gives you / "the community" lots of
options to rewrite the firmware, FPGA code, etc. It's possible that
this would limit performance (1 GSPS into an FPGA requires quite a bit
of care), but there'd probably still be a market for a fully
customizeable fractional-GSPS instrument.

 

[Joerg]

The Instek GDS-820 has delayed trigger, but only operates on the
external trigger input!

That's how it is with the GDS-2202, the less expensive 2-ch sibling of
mine. Beats me why they did that but I could live with external trigger
in most cases. However, what really puzzles me is that they completely
took it away for the 4-ch version.

 

[Joerg]

Le Wed, 15 Aug 2007 07:13:09 +0000, Joerg a écrit:




Right now, I have to do some really accurate compensation to a system's
step response. From the 10ns scale, down to the 100ms scale. What would
really be handy is a logarithmic sweep. That could easily be implemented
with say a sampling rate scaling down by a two factor while time rises by
a two factor.
Well, with GPIB and some post-processing it works, but it's much longer
and a bit... bulky.

The newer scopes such as this Instek allow data dumps via USB into the
PC at an amazing clip. Not realtime but close. When I want to do
something unusual with the data I use the CSV import into Excel.
However, I am not an expert (yet) in VBA so I can't do any processing
on-the-fly. But I'll learn it. Was hoping it was C and not Basic but, oh
well, I guess that was Bill Gates' decision.

 

[Joerg]

Maybe the real question to ask: when is someone (almost certainly a
new party) going to come out with a customizeable DSO "kit".

Basically, I'm thinking of something that is workable with the
shipping "reference design" but gives you / "the community" lots of
options to rewrite the firmware, FPGA code, etc. It's possible that
this would limit performance (1 GSPS into an FPGA requires quite a bit
of care), but there'd probably still be a market for a fully
customizeable fractional-GSPS instrument.

That's what I was thinking when the GDS-2204 arrived and I took it for a
spin. Endless possibilities. But while Instek is quite open in accessing
the data and the control codes there is no service manual with
schematics and no firmware access. I bet that what you described could
already be done with many new DSO, it's just that the manufacturers
don't want us to.

They must have some blazing processor in there, the speed of the FFT
blew me away. You can experiment with, say, a snubber network while
watching the spectrum change. However, with this scope you have to get
used to the bonbon-colored traces ;-)

 

[Joel Kolstad]

However, I am not an expert (yet) in VBA so I can't do any processing
on-the-fly. But I'll learn it. Was hoping it was C and not Basic but, oh
well, I guess that was Bill Gates' decision.

If you really want to get ambitious, note that VBA (as provided in
Word/Excel/etc.) will let you access COM objects which -- for the purpose of
this discussion -- boils down to nothing more than a function call interface
with a prescribed set of data types and calling conventions. COM objects can
be programmed (implemented) using C++ (and plenty of them are), although it's
much easier to write them using one of the ".Net" languages such as C#.

Realistically I think Gates' choice of VBA rather than C for Office makes
sense... I know office workers who have absolutely zero formal programming
background, and they can manage to modify and write various little utilities
and macros in VBA, whereas I think they'd find the initial learning curve a
lot steeper with C and wouldn't even bother with it.

---Joel