EEVblog #306 - Jim Williams Pulse Generator

[Nermash]

Here is something for all those who hate waiting for pcbs and just don't have that required 2N2369 in the junk box

Old faithfull BC547B in FSMT (Flying Surface Mount Technology - as originated by Robert Lacoste in his book "The darker side" ) deployment, connected to +170 V DC Nixie supply that I am tinkering with at the moment, and with 470k current limiter.

Captured using Owon SDS7102 100 MHz DSO. I am guessing that excessive amount of ringing on the trailing edge is due to using the standard ground lead. I will probably try connecting it with 1 m of coax...

[jdesbonnet]

Just got a a small selection of low cost, low voltage RF BJT transistors to see what might avalanche. So far just got around to testing two individual devices:  1 x 2DD2661 with a nominal V(br)ceo = 12V. Tested up to Vce of 80V, no joy.

1 x BFG424F with nominal V(br)ceo=4.5V: got avalanching at about Vce = 30V.

[Rufus]

I had a quick play around, not that my 200MHz x3000 is fast enough to know what is going on.

A transistor socket soldered between two SMA sockets, 15k from base to ground. On the collector about 3 feet of rigid coax fed with 250vdc from an insulation tester via 1M. On the emitter a 10db attenuator and coax to the 50 ohm input on the scope. If I got the probe settings right the traces show what is on the emitter.

I only tried about 10 transistors so far and about 2/3rds avalanched. The traces are from a BC184B. The pulse amplitude depends on the transistor and jitters a lot. The pulse width was very stable. The rise before the pulse is I guess leakage before avalanche and varied with transistor. Not sure what is going on with the fall after the pulse.

[free_electron]

Boards came in yesterday from itead. Parts from digikey should arrive today.


i will build one somewhere this week and assemble the kits.
i have 8 spare boards. The kits will appear on ebay and you can order there using paypal. That solves all payment / currecny exchange problems.

[carpin]

The board looks great, the track length from the transistor to the bnc is short, exactly what Williams did.

 

[max-bit]

FREE_ELECTRON Only one, the question ?
 Did you test pulse generator at a voltage of 3 V power supply? When power is above 1.6 - 1.7 V, the inverter is not stable and at a voltage of 1.8 V the current reaches 0.8 A!!!
 Here is a video: (test)


 I do not know, maybe the fault of the elements used (coil? DIODs), but they are consistent with the specification.
 And so can not supply the "Pulse Generator", directly from the lithium battery?
 Please test whether other models of the LT1073 (DC-DC converter) is also a problem occurs.

And ...
The transistors have been tested properly work only orginal Motorola transistors T
hey did not want to operate transistors: PHILIPS, National

[tnt]

@max-bit: This is really strange. I can feed 5V to my pulse gen without any problem, current is less than 20 mA.

[max-bit]

It's interesting what's the problem ... with the coil?
Used this model:
http://uk.farnell.com/wurth-elektronik/7447779215/choke-smd-150uh/dp/1635996?Ntt=1635996

[DRT]

I built one of these pulse generators using a BSX20 which seems to work very nicely - I checked it out on a Rohde & Schwarz RTO (2GHz BW 10GS/s) at work which measured a 270ps rise time.

[max-bit]

Today,  solder the second PCB ....
 We'll see if the problem persists ?
 Can someone this problem occurred?
And this BCX20 what manufacturer?

[DRT]

And this BCX20 what manufacturer?

The BSX20 I used was made by ST.

[max-bit]

Case explained. Stupid mistake, capacitor C1, the reverse (polarization!) Soldered (tantalum). ..... always have to double check.

[max-bit]

Here are my measurements (unfortunately, not so fast oscilloscopes)
 Oscilloscope OWON SDS8202 - 200MHz 2GS/s
 BW Measurement of rise time of ~ 1.3 nS That is: BW = 0.35 / tr BW ~ = 270 MHz
 And another interesting measurement - TDR (Time Domain Reflectometry)
 Coaxial line to the antenna.
The first reflection after approximately 18 ns / 2 (divide by two because it's time to reach the signal and return) - on the connector cable - Transition from BNC to N
 Second, reflections on antenna surge protection at the time 40 ns / 2
 Final reflections from the antenna 86 nS / 2
 Lc (cable length) x-factor reduction (H-1000 cable x = 0.82)
 Lc = tr * C * x
 LC1 = 2.2 m
 LC2 = 4.92 m
 LC3 = 10.5 m

[KTP]

The Owon SDS8202 looks very nice!

I think you need a flat top for a few nanoseconds however.   If you know the exact amplitude the pulse is supposed to be into whatever load you have, then I guess such a short pulse could be used for BW measurements.

For instance, say on a 1Ghz scope this pulse gets to 10V but on a 200mhz scope it only gets to 7V because the scope rise time is longer than 1/2 of the pulse length.  Wouldn't this make the risetime calculation on the 200mhz scope in error?  It seems you kind of need to know where the signal is going to "level out" in order to get the true 10% and 90% points...

[KTP]

I just did an example to try to prove to myself that it does make some difference if you have a flat top vs a short pulse for rise time measurements.

In the first pic I have the best rise time I have been able to get on my MSOX3024A with a 600ps rise time, flat top (20ns long) pulse.  This was measured at 1.59ns meaning a BW of 220mhz.

In the second pic, I took the same tinylogic circuit, but this time just brushed the input across the table, generating a bunch of very fast short pulses with no flat top.  Now my scope measures the rise time at 1.47ns meaning a BW of 238mhz.

Ok, so they aren't *that* far apart...but it is almost 10% off and makes me wonder even how much shorter I could measure if I had a circuit capable of 200ps rise time.  I think you need the flat top.

[max-bit]

OWON is not bad, especially as the price taken into account (approximately $ 800)
2GS (for the record 10k) 1GS (over 10k), but here to 10M!
 A little clunky, but something like Agilent, Tektronix is a price above $ 5000
 So do not complain about If the Chinese improved by housing (quality) components, and software.
 It's even a 1300 - 1400 USD was by this revelation.

[T4P]

Unfortunately 1GS onwards only has 10k memory length
That said if i want a 2GS scope i would go for a Rigol 4000 for 4ch/s or the 2000
But it looks like rigolna is having a problem with their site now
But it's actually onpar with DS1202CA above 1GS (and cheaper)
Can't say the same for the software .. yeah the scope is good and i love the SDS7102 but i'm not sure if the 8202 would be a good buy ... just for the sake of the software if i want a 2GS scope i better be getting me money's worth

[max-bit]

Find cheaper oscilloscope from SDS8202? With these parameters! Rigol is not bad, but the price of x2

[T4P]

Find cheaper oscilloscope from SDS8202? With these parameters! Rigol is not bad, but the price of x2

Sure it is, but 14MB of memory. Why not?

[max-bit]

Model and Price

[T4P]

Model and Price

DS2202  And plus it's upgradeable to 56Mpts ...

[max-bit]

Compare: Rigol DS1302 Memory 300MHz BW 2 GS 10k only - price 1100 Euro (1375 $)
 Rigol DS2202 14M Memory 200MHz BW 2 GS price 1300 Euro (1645 $) + 300 Euros ($ 360) Total 1600 Euro (2000 $)
and
DS8202 OWON price 650 Euros ($ 825)

 It is a price higher than 2 x ....

[max-bit]

Although RIGOL, replanted .... These 54 MB built-in oscilloscope, it is only activated? They were able to sell immediately with 54 MB, and not even take it for a $ 350

[T4P]

Hmm ... where's your 300 euros coming from?
I do not mind paying extra for good software, the owon SDS7102 is a good scope the software is just adequate (i like it ... but i face it, it's not gold) but i won't say the same for the SDS8102
Plus the DS2202 has quite high a acquisition rate (refresh rate or acquisition?)
and 14M ... fullspeed i think
Can somebody confirm?

[IvoS]

Can someone tell me where to get 2N2369 transistor that avalanche? I got some 2N2369A from Mouser and they don't work.