Cheap Fast Rise Time Oscillator

[jonpaul]

Bonjour, topic seems off a bit this is a fast rise pulse generator, an oscillator is normaly a sinne wave output.

See this related EEVBLOG thread re probes, many tests and shots of fast rise pulsers, DIY, Leo B with various scopes and probes.

https://www.eevblog.com/forum/testgear/probe-into-probes-whats-up/50/

The Leo Bodnar 40 pS BNC pulse gen is low cost, high developed and very fine performance. Was ~~ L 50
http://www.leobodnar.com/shop/index.php?main_page=product_info&products_id=295

Our Leo B has give fine service for many years.

Here are my  2019 Tektronix  7104/&A29/LeoB 40 pS shots with Mini Ckts 20 db 50 Ohm atten.

After selection of 7A29 and transient response adjustment both  7A29 and 7104, perhaps the best I can do.

Your thoughts appreciated!

Lower pix   BNC inline atten and 50 Ohm  term collection.

Vive l'ANALOG! Vive TEKTRONIX! !

Jon

[BillyO]

So I maybe not following you 100%, but what I'm getting is that the LB pulser gave you a 10ns rise time when directly connected, but a .79 ns when connected via a 20 db attenuator.  Is that correct?

Just a bit of confusion.

In my experience, the rise time improves with increased current in the load.

[jonpaul]

Rebonjour, BillyO:

1/ Leo Bodnar 40 pS  is designed to drive a 50 Ohm termination or 50 Ohm scope input.

There is no spec for other Zo,  so current is the same with and without the attenuator,  and load not the issue.

See the spec sheet of the LB 40 pS.

2/ A constant impedance  atten is a resistive T or Pi network designed to attenuate with same Zo input/output eg 50 Ohms.

 See spec sheet of Mini Circuits HAT-20 wideband 50 Ohm attenuator.

See Tee /Pi attenuator

Thus the load on the LeoB pulser is 50 Ohms with or sans the HAT-20 attenuator.

3/ Purpose of the attenuator is to the reduce reflected energy from scope t pulser due to VSWR or Zo mismatch.

Thus pulse aberrations due to reflections from scope to pulser are reduced and the test aberration is reduced.

Pix: Leo B USB utility , allows Vo adjust 50 mV>>1.5 V set at 1.0V P-P into 50 Ohm.

Yokogawa DL7440 500 MHz BW, FS = equv 100 G/sec, sweep  at 1 nS/div

Upper trace LeoB40 direct to scope 50 Ohm input
Lower trace LeoB40>>Mini Ckts HAT-20 atten>scope 50 Ohm input

Many previous old EEVBLOG threads on similar like "show us your squarewave" TD pulser, Avalanche pulser, etc.

Hope all is now clear!


Jon




 

[David Hess]

One way to get a cleaner pulse is to use the output of the logic gate to drive the emitter/source of a common base/gate transistor with then has a pull-up or pull-down on its collector/drain to produce the output.  Then on the rising or falling edge, the transistor "disconnects" the gate from the output and only the pull-up or pull-down in parallel with a very small capacitance is connected to the transmission line.  This essentially duplicates how an ECL output from a differential pair works.

Some designs may use a microwave diode in place of the transistor.

[magic]

Do you expect that ordinary 3904 could be good enough to compete with "bare" LVC logic in terms of rise/fall time?
Or are we talking some special RF transistors?

[mawyatt]

The 2N3904 would be too slow for this app, a better solution would be a high Ft SiGe HBT like the BFR840, BFP620, BFU725, and others. These inexpensive SiGe devices aren't designed nor specified for pulse use (they are RF devices), however should yield very fast pulse response if applied properly.

Best,

[BillyO]

The 2N3904 would be too slow for this app, a better solution would be a high Ft SiGe HBT like the BFR840, BFP620, BFU725, and others. These inexpensive SiGe devices aren't designed nor specified for pulse use (they are RF devices), however should yield very fast pulse response if applied properly.

Best,

This then begins to defeat the design criteria of cheap and simple by adding cost and complexity.

[mawyatt]

The 2N3904 would be too slow for this app, a better solution would be a high Ft SiGe HBT like the BFR840, BFP620, BFU725, and others. These inexpensive SiGe devices aren't designed nor specified for pulse use (they are RF devices), however should yield very fast pulse response if applied properly.

Best,

This then begins to defeat the design criteria of cheap and simple by adding cost and complexity.

These are about $0.50 HBTs 

Best,

[mawyatt]

These inexpensive SiGe HBTs brings up a question IF they would perform well in an avalanche mode for fast pulse edge production?

Anyone given this a try?

Best,

[David Hess]

Do you expect that ordinary 3904 could be good enough to compete with "bare" LVC logic in terms of rise/fall time?

No, a 2N3904 is too slow.

Or are we talking some special RF transistors?

Nothing exotic is required and RF transistors can be inexpensive.