EEVblog #306 - Jim Williams Pulse Generator

[BravoV]

#2. 

Juan, it looks you're very good at finding article at the internet and you love to quote them on every posts, now, get your favorite "audio" transistor datasheet, and then look for 2N2369A datasheet and start to compare between them.

Yes, I know it looks like an old outdated transistor, but still, please do it. This thread grew into this giant just because of this particular transistor.

Again, please, do it, for your own good sake, cause you're becoming to look like an idiot and worst it looks like you're starting to be proud of it. 

DO IT NOW !!! 

[JuanPC]

#2. 

Juan, it looks you're very good at finding article at the internet and you love to quote them on every posts, now, get your favorite "audio" transistor datasheet, and then look for 2N2369A datasheet and start to compare between them.

Yes, I know it looks like an old outdated transistor, but still, please do it. This thread grew into this giant just because of this particular transistor.

Again, please, do it, for your own good sake, cause you're becoming to look like an idiot and worst it looks like you're starting to be proud of it. 

DO IT NOW !!! 

 

http://pdf.datasheetcatalog.com/datasheet/SGSThomsonMicroelectronics/mXtuvtv.pdf
http://www.nteinc.com/specs/100to199/pdf/nte123a.pdf
http://www.centralsemi.com/leadedpdf/2n2369a.pdf
2n2369A Rise Time: 9ns-25ns.
Ft Current Gain BW: 300-500Mhz

Rt=0.35/BW
1ns = 350MHz scope ideal.
 

[alank2]

I can't wait for    free_electron to have more kits!!!!!

[c4757p]

The 2N2369 will often perform much better than datasheet specs. Note that Jim Williams said that one may have to be selected. I've made them switch with 1-2ns rise time just in the traditional manner, with a very hard drive (100 ohm || 10nF on the base, 100 ohm collector load) and very low emitter inductance (2mm straight to the ground plane).

And it will blow right past that if avalanched.

Check the other specs. They're clearly not the same transistor.

[JuanPC]

The 2N2369 will often perform much better than datasheet specs. Note that Jim Williams said that one may have to be selected. I've made them switch with 1-2ns rise time just in the traditional manner, with a very hard drive (100 ohm || 10nF on the base, 100 ohm collector load) and very low emitter inductance (2mm straight to the ground plane).

And it will blow right past that if avalanched.

Check the other specs. They're clearly not the same transistor.

Still can´t believe a short circuit can´t perform better.

Maybe a Tesla coil or a TV Flyback connected to a serious voltage divider, why can´t perform better?
 

[BravoV]

Though English is not his native, definitely this is not a language barrier, so its confirmed now that he is not that "bright" isn't it ?   

I'm done and out of here, too bad, this very fine thread is now tainted by this troll.   

[JuanPC]

Though English is not his native, definitely this is not a language barrier, so its confirmed now that he is not that "bright" isn't it ?   

I'm done and out of here, too bad, this very fine thread is now tainted by this troll.   

you have: 1589 Posts.
i have: 97.
 
Troll Envy   

[c4757p]

Still can´t believe a short circuit can´t perform better.

Still can't believe I can't fly if I flap my arms really fast.

Oh wait. It doesn't work whether or not I believe it should.

[c4757p]

Maybe a Tesla coil or a TV Flyback connected to a serious voltage divider, why can´t perform better?

Spark gaps actually can have very fast rise times. I'm not sure how they compare to an avalanche transistor. Why don't you try it? You should only need a few hundred volts, not a Tesla coil.

They're also unpredictable, dangerous and noisy as hell... but each to his own...

[free_electron]

wrote a truckload of blabla in this and other topics about using audiocard to measure clock risetime
and 1ns = 350MHz scope ideal

What kind of medications are you on ? it's time to pump up the dosage ....

That NTE replacement transistor doesn't come close. the 2N2369 is designed for ultrafast saturated switching applications the NTe for medium speed switching apps... not even saturated switching.
the NTE barely grabs 40ns turn-on time while the 2N2369 is in the 10nS time range. Anyway, the point is irellevant since we are using the transistor in avalanche mode. There is no telling how the NTE will behave as it is not specified in that mode. Avalanching a transistor is non-standard usage and parts are not characterised for that. Jim williams tried many different transistors and found the 2N2369 to be the best out of all. He then hand picked the best out of a batch of those.

Having built and measured 70 of these pulse generators, one the fastest scope around (then, now the Q version doubles its speed)  and done a spread (see below) the design is proven.
All you have done is point at stuff other people have done, and then post your own misinterpretation of it. So shut up and go away.

[c4757p]

Though English is not his native, definitely this is not a language barrier, so its confirmed now that he is not that "bright" isn't it ?   

Yo hablo un poco español también, quizás él puede hablar conmigo si él tiene un problema.

Aunque, creo que su problema es del tipo médico. No soy un psiquiatra.

[BravoV]

Though English is not his native, definitely this is not a language barrier, so its confirmed now that he is not that "bright" isn't it ?   

I'm done and out of here, too bad, this very fine thread is now tainted by this troll.   

you have: 1589 Posts.
i have: 97.
 
Troll Envy   

Fully understand a typical post count chasers do when they joined a forum, but I just don't get it on why bother with those efforts and troubles posting all those tons of links, and typing all those long arguments at each post ? Cause only with a simple single line of junk words will also increase the post counter too.

This is it, the 2nd confirmation about the "bright" thingy that I mentioned and I'm guessing the 3rd one is also coming.

[GeoffS]

Several useless posts removed.
Can we get this thread back on topic please?

[dr.diesel]

I can't wait for    free_electron to have more kits!!!!!

Agreed.  f_e, if you've got a list going, please add me as well!

[Lukas]

Using relays for generating fast-risetime pulses isn't a new idea: http://w140.com/tekwiki/wiki/109

Using an external voltage source, pulses up to 300V are possible.

Try to replicate that using today's equipment...

[warp_foo]

I can't wait for    free_electron to have more kits!!!!!

Agreed.  f_e, if you've got a list going, please add me as well!

+1 more

[kvitekp]

+1 here

[Dave]

+1

[AlfBaz]

I have a scope that's 1GHz into 50ohm and 500MHz into 1Mohm. 50ohm only goes to 1V per division so I'm not quite getting the whole pulse to fit on screen. Do I need to buy an expensive attenuator or is there a cheaper alternative?

[fmaimon]

w2eaw has a solution for you:

[AlfBaz]

w2eaw has a solution for you:

What's going on with my memory? I've already seen that video!

[G7PSK]

w2eaw has a solution for you:

What's going on with my memory? I've already seen that video!

It needs re-seating in its socket!

[AlfBaz]

It needs re-seating in its socket!

Thanks, gave me a good chuckle
Although I fear contaminants on the pcb traces are causing clock skew

[BravoV]

It needs re-seating in its socket!

Thanks, gave me a good chuckle
Although I fear contaminants on the pcb traces are causing clock skew

Easy, get a sniff with a strong inhalation on a single burst of contact cleaner & lubricant, or just plain WD40. 

[van-c]

I can't wait for    free_electron to have more kits!!!!!

free_electron Please add me to the list too.

Speaking of w2eaw's videos, I followed this one: to build a Schmitt trigger pulse generator like his example.  I wanted to use his pulse generator do do a sanity check of my 150 MHz Tek 2445A scope's bandwidth. 

I used a BNC connector soldered to the board that directly connected the pulse generator to channel 1 input (no probe used).   I measured a rise time of about 4.4 nS on the 2445A.  The 74AC14 datasheet gives a propagation delay of 1.5nS minimum to 10 nS maximum and typical delay of 7 nS.   So, the rise time of the scope should equal the measured rise time minus the propagation delay of the pulse generator (right?).  Using an estimated delay of, say, 2 nS for the pulse generator itself, that would give a scope rise time of 4.4 - 2.0 = 2.4 nS, or a bandwidth of about 145 MHz, which confirms the sanity check, assuming I am doing this right, i.e., making no invalid assumptions.

Of course, this pulse generator is only useful for bandwidth measurement on scopes where the circuit's delay is small compared to the rise time of the scope.  I would really like to use the JW pulse generator when the kit is available again to get a more accurate measurement.

--Van