Pulse Generator KIts or Ready to Use

[max-bit]

Pulse Generator (based Jim Williams PG)

Ready to Use 48$

Payment via PayPal (please contact via personal message)

Shipping cost Europe - 5 Euro
                    World  -  7 $

Precise measurements
used oscilloscope:
LeCroy LC584AL 1 GHz 8GS/s (rise time ~330-350ps)
Pulse the load 50Om


Pulse the load 1MOm

jitter trigger pulse

[max-bit]

price adjustment

[free_electron]

Your pulse does not have a plateau , so you don't really know what the rise time is as the pulse decays before you hit full conduction.
I do like the enclosure though

[tggzzz]

Your pulse does not have a plateau , so you don't really know what the rise time is as the pulse decays before you hit full conduction.
I do like the enclosure though

While I understand why you say that, does it matter?

Provided the test setup isn't the limiting factor, you can measure delta-v and delta-t to get the risetime of this particular pulse.

[macboy]

Your pulse does not have a plateau , so you don't really know what the rise time is as the pulse decays before you hit full conduction.
I do like the enclosure though

While I understand why you say that, does it matter?

Provided the test setup isn't the limiting factor, you can measure delta-v and delta-t to get the risetime of this particular pulse.

No, you can't. Not unless you can see the top of the pulse. The problem with a short "hill" type pulse that has no plateau is that you can't know if your scope is displaying the peak value or not. Without the peak value, you can't tell the risetime. Keep in mind that the slew rate (V/second) of the display on the scope will be linked to the magnitude of the pulse. The risetime however is not. That is the fundamental difference.

[free_electron]

correct. you need to find the 10 and 90% points . to do that you need two plateaus. a lower and an upper.

Then you can establish the time needed to traverse the 10-90 point , look at the voltage delta between the points   ,and convert it into volts/second. that is the meaningful number.

risetime is expressed volts per second.  just saying rise time is 260picoseconds means nothing. Is it 260 picoseconds for 1 volt or 260picoseconds for 1000 volts ? ( the latter is a much faster pulse ... )  the slope is much steeper.

[tggzzz]

Your pulse does not have a plateau , so you don't really know what the rise time is as the pulse decays before you hit full conduction.
I do like the enclosure though

While I understand why you say that, does it matter?

Provided the test setup isn't the limiting factor, you can measure delta-v and delta-t to get the risetime of this particular pulse.

No, you can't. Not unless you can see the top of the pulse. The problem with a short "hill" type pulse that has no plateau is that you can't know if your scope is displaying the peak value or not. Without the peak value, you can't tell the risetime. Keep in mind that the slew rate (V/second) of the display on the scope will be linked to the magnitude of the pulse. The risetime however is not. That is the fundamental difference.

I think much of that is covered by my caveat/precondition "Provided the test setup isn't the limiting factor".

However, I'm irked about the distinction between the risetime and the slewrate - of course you're right. Coming at it from the direction of considering the relative power vs frequency spectrum of the pulse caused me to forget that fundamental, doh!

[macboy]

correct. you need to find the 10 and 90% points . to do that you need two plateaus. a lower and an upper.

Then you can establish the time needed to traverse the 10-90 point , look at the voltage delta between the points   ,and convert it into volts/second. that is the meaningful number.

risetime is expressed volts per second.  just saying rise time is 260picoseconds means nothing. Is it 260 picoseconds for 1 volt or 260picoseconds for 1000 volts ? ( the latter is a much faster pulse ... )  the slope is much steeper.

No, rise time is expressed solely in seconds, or most commonly ns or ps. If you see V/s then it is slew rate, which is a very different thing.

[free_electron]

You still need to tell me how many volts you traversed to get that risetime. Risetime is meaningless unless you know the voltage swing.
For example : they specify on logic gates always what is the power supply voltage at which a particular risetime was measured.
Take a simple 4011 for example: Risetime 180nS @ 5v , 90nS @ 10v  65 nS @15v ...

Risetime is 10% to 90% of the output swing. But, without knowing what the output swing is, risetime is a useless number.

A 1 hz triangle can have a risetime of 1 picosecond... if its vpp is 1.2 picovolt ...

That's why, for opamps we dont talk about rise time but slew rate. faster slewrate = higher bandwidth ...

so if you are after testing a scope's bandwidth ... you need SLEWRATE ! . risetime is meaningless.
to establish slewrate you need the plateau's,

[macboy]

You still need to tell me how many volts you traversed to get that risetime. Risetime is meaningless unless you know the voltage swing.
For example : they specify on logic gates always what is the power supply voltage at which a particular risetime was measured.
Take a simple 4011 for example: Risetime 180nS @ 5v , 90nS @ 10v  65 nS @15v ...

Risetime is 10% to 90% of the output swing. But, without knowing what the output swing is, risetime is a useless number.

A 1 hz triangle can have a risetime of 1 picosecond... if its vpp is 1.2 picovolt ...

That's why, for opamps we dont talk about rise time but slew rate. faster slewrate = higher bandwidth ...

so if you are after testing a scope's bandwidth ... you need SLEWRATE ! . risetime is meaningless.
to establish slewrate you need the plateau's,

I am afraid that you are wrong here.

On an oscilloscope, the rise time and bandwidth are directly related (though less closely now with digital scopes than with old analog scopes). It is generally accepted that this relationship is BW = 0.35 / risetime. So a 200 MHz scope will have a rise time of approximately 1.75 ns. This rise time will be observed regardless of the magnitude of the pulse (until the slew rate limit is encountered).

For opamps, the slew rate is a prime specification, yes. But slew rate for op-amps is not directly related to bandwidth. The GBP is a more useful specification for that. Only when the output swing is very large does slew rate limiting come into play. For high voltage measurements with an oscilloscope, it is quite possible to come up against the slew rate limit before the rise time becomes the limiting factor, just as for an op-amp circuit you can be limited by slew rate before exceeding the bandwidth of the system, when outputting a very large voltage swing.

Rise time for an oscilloscope is not useless without a magnitude as you contend. The magnitude is, of course proportional to the input voltage. When thinking in the domain of digital logic levels, slew rate and rise time are related, by the logic levels themselves (the rise time is essentially how long the logic gate takes to slew at full speed from one level to the other). For an oscilloscope, you need to think differently. There is no fixed voltage. It is proportional to the input. And for that reason, the rise time is constant regardless of the input magnitude. If the scope has a rise time of 1 ns to a 10 mV pulse, it will have a rise time of 1 ns to a 10 V pulse too. This is despite the fact that the effective slew rate is three orders of magnitude higher in the latter case.

Google "oscilloscope rise time and bandwidth". Read.

[tggzzz]

You still need to tell me how many volts you traversed to get that risetime. Risetime is meaningless unless you know the voltage swing.
For example : they specify on logic gates always what is the power supply voltage at which a particular risetime was measured.
Take a simple 4011 for example: Risetime 180nS @ 5v , 90nS @ 10v  65 nS @15v ...

Risetime is 10% to 90% of the output swing. But, without knowing what the output swing is, risetime is a useless number.

A 1 hz triangle can have a risetime of 1 picosecond... if its vpp is 1.2 picovolt ...

That's why, for opamps we dont talk about rise time but slew rate. faster slewrate = higher bandwidth ...

so if you are after testing a scope's bandwidth ... you need SLEWRATE ! . risetime is meaningless.
to establish slewrate you need the plateau's,

Um no. The clue is in the name: riseTIME. (BTW, time is measured in seconds (SI abbreviation "s"), not Siemens (SI abbreviation "S").)

It is true that digital logic signals are typically measured in terms of the time taken to traverse 10%-90% of the final value, but the pulse generator's output is not a digital logic signal. Even so, a digital logic signal's risetime is measured in seconds, not volts per second.

It is also true that a primary characteristic of opamps is the slewrate, measured in volts per second - but that's not relevant here. Similarly EMI/EMC measurements are based on slewrate.

As for your triangle wave, please draw an annotated picture showing the voltages and times (in seconds). Then add the slewrate (in volts per second).

[macboy]

...
A 1 hz triangle can have a risetime of 1 picosecond... if its vpp is 1.2 picovolt ...
...

Think about that again. A 1 Hz triangle (assuming equal rise and fall slopes) has a 10-90% rise time of 0.4 seconds. This is completely independent of voltage.

[Electro Fan]

Possibly some related discussions in this thread:
https://www.eevblog.com/forum/blog/eevblog-306-jim-williams-pulse-generator/msg775096/#msg775096

From practical experience it seems there are at least a couple types of pulse generators:  those that make "peaky" pulses and those that make a wider flat top pulse.  From further practical experience it seems that this statement might be applicable when trying to get an accurate measurement:  "Provided the test setup isn't the limiting factor".

[max-bit]

OK
To dispel doubts
used oscilloscope:
LeCroy LC584AL 1 GHz 8GS/s (rise time ~330-350ps)
Pulse the load 50Om


Pulse the load 1MOm

jitter trigger pulse

[Electro Fan]

max-bit, I'm confident that your pulse gens are solid - I think there is a minimum scope performance needed to correctly observe your pulse gen's full performance capability. 

Any chance you can put your pulse gen on a scope that is 200MHz or 400MHz or 500 MHz (or all three if possible) and see what it shows for the pulse shape and the rise time?  These results could be contrasted with the results shown for the 1 GHz scope used in your post above.

Thanks, EF

[free_electron]

...
A 1 hz triangle can have a risetime of 1 picosecond... if its vpp is 1.2 picovolt ...
...

Think about that again. A 1 Hz triangle (assuming equal rise and fall slopes) has a 10-90% rise time of 0.4 seconds. This is completely independent of voltage.

You're right. brain fart on my part. note to self : drink TWO cups of coffee before posting early morning ...

[max-bit]

The rise time is dependent on the rise time of the oscilloscope.
oscilloscope 300MHz (Rigol DS2302) RT~ 1-1,1nS

[tecman]

can you post some pix of what you are selling ?

paul

[cs.dk]

can you post some pix of what you are selling ?

paul

Pictures is in the ebay link, in first post: http://www.ebay.com/itm/Pulse-Generator-RiseTime-270ps-PulseWidth-600ps-TDR-/262105510628?

[max-bit]

current Auction:
eBay auction: #262128187720  PRICE 55 $

[max-bit]

price adjustment low price only 50 $
For details, see first post

[max-bit]

Price 48 $