#286

episode 286
fast rise times

[@listerdave1240]

I don't think an oscilloscope with 70MHz bandwidth can reliably display a rise time less than a few nS. I think if you were to measure a pulse with 3nS rise time and one with a 100pS rise time they would appear almost the same on the oscilloscope, the faster one probably inducing some ringing (in a digital scope). With regards to the rise and fall times in the datasheet keep in mind that these are assuming a relatively high capacitive load and are dictated by the output impedance and current drive capability of the output stage rather than the actual switching time. For an open output or even feeding directly into a 220 ohm resistor it is likely to be considerably faster. My oscilloscope is only a little faster at 100MHz so I am unable to confirm it that way but if instead I just connect a short piece of wire to the output I can see strong harmonics way up into the GHz on a spectrum analyser which is sort of indicative of rise times much faster than 3nS. To see them clearly I increased the frequency of the oscillator inverter to about 20Mhz as with the original 100Hz the power transmitted from the 'antenna' was too low for the spectrum analyser to capture. Of course if you are using that same 70MHz oscilloscope to do the TDR measurements all the above doesn't matter. It would be interesting though if someone here with access to a faster oscilloscope could try out this same test. My guess is that both will give a considerably faster rise time probably around 2nS for the HC and less than 1 for the AC.

[@dodinho4909]

Yes, you are right, I measured today with a 4GHz scope (almost same circuit, 74AC14). Rise/fall time of

[@jacks5kids]

ListerDave1240 is absolutely right, but to observe sub-nanosecond rise times correctly, you need ALL of the following. 1) An oscilloscope that is fast enough. 1GHz is the bare minimum, but at least 2GHz should be used 2) Either a really fast scope probe, or a circuit that has an output impedance of 50 ohms and the use of 50-ohm impedance cables throughout into the 50ohm scope input. 3) Proper decoupling capacitors connected just millimeters from the silicon of the IC. Matrix boards, through-hole resistors are not suitable, and a DIP package for the IC will degrade the measurement. In this case, use 1nF, 10nF AND 100nF all in parallel with the smallest capacitor closest to the chip 4) A proper ground plane to connect input, power and output grounds. This demo had none of these four, and just having a fast scope is not enough. The ringing that was observed is generally not a function of the properties of the IC, but of the test circuit

[@PebblesChan]

I just discovered your channel. Excellent stuff! You should upgrade your Z80 and 8080 stuff to play videos games as I did back in the early 1980’s. As for this video, it appears you hit the head of the nail - the 74ACxx part has a different impedance and you’re seeing the consequences of it (although a relatively minor case). This illustrates that you just can’t necessarily insert a new and faster logic family in place of an older and slower one. These days we even get problems using the modern versions of supposedly old logic families because of the smaller and faster semiconductor geometries to same cost by increasing yields. If you get too fast, the good old DIP packaging also contributes towards impedance challenges. Many DRAM designs of the 1970’s and 1980’s used series terminating resistors on their address lines to quell the potential unwanted under/overshooting and signal ringing.

[@colinprewett4480]

as i was screaming at the tv its 99hz, thankfully, you corrected yourself!

[@flemming2jacobsen911]

Nice circuit. I spent 1/1 playing with it :-)

[@ohboy4784]

Hello dear sir, What you are measuring is actually the scope's input analog bandwidth, not the DUT !!! The results you get are in the ball park of 4.5nsec for a 70MHz scope like the Rigol you are using. ( BW~= 0.35/Rtime) Therefore ,You should use a vastly faster scope( 500Mhz or higher),could be an analog scope as well. Furthermore , the measurement of the 74AC part is flowed ,since you shouldn't (or the auto mode of the scope shouldn't )measure rise time of the ringing amplitude but of the logic levels "settled voltages". Lastly you should terminate the output to avoid the ringing. I usually enjoy your videos, this one ...not so much :)

[@alejandrosantiago6080]

Very interesting video, you hould make the part to count the inductance, comparisson. -Alex.

[@MIKROWAVE1]

This is a fine video. This circuit can be adapted to making a fast square wave driver for a level 7 DBM.

[@MIKROWAVE1]

Wow all kinds of drive into 50 Ohms; In fact I had to add attenuation.

[@rtybn2012]

Using 220 ohm equates to 44 ohms. That is assuming that the output is zero ohms of the IC's. You need to use 250 ohms and use a 50 termination on the scope.

[@MsFireboy2]

Thanks for posting. Question Back when these came out who was the first manufacturer?

[@IMSAIGuy]

TI

[@MsFireboy2]

@@IMSAIGuy Thank you. In you opinion will this produce a fairly noise free output signal?

[@IMSAIGuy]

I don't understand the question. this is a digital system zero or one@@MsFireboy2

[@MsFireboy2]

@@IMSAIGuy Square wave RC oscillator circuit.

[@IMSAIGuy]

kzbin.info/www/bejne/pWSUkJmkidaIhdksi=dLArShyD0NBJaXlg

[@norbert.kiszka]

3.3 nS of rise time equals 106 MHz of this oscilloscope bandwidth.

[@bretfuzz925]

I can’t help but point out that your power supply decoupling capacitor is too far from the device to work properly for the AC family part. The socket, leads, and lead frame inductance of the part are causing the ringing. There is no need for the 1uF tant if decoupled properly.

[@ivolol]

LVC parts should be able to go a bit faster again, at 3.6V tho

[@user-zc8sd8jx8s]

LVC (at least the -04 and -14 etc.) work fine with 5V as well.

[@andrewbell185]

Could the rise time measurement be affected by the bandwidth of the scope?

[@IMSAIGuy]

maybe

[@MahmoodElnasser]

@@IMSAIGuy That's why you can't measure the speed difference. Your scope bandwidth is 70MHz which is useless for measuring sub 4nsec rise / fall times. Use faster scope and you will see the difference. Also you need a better probe higher bandwidth and short gnd clip. your long gnd wire gives measurement errors. Use something like gnd spring or even better SMA cable to connect to your scope.

[@IMSAIGuy]

@@MahmoodElnasser Yes you are right. However since that is the only tools I have then I don't need a faster edge for the tools I have. If I ever get a faster scope then I will revisit.

[@tonyfremont]

@@IMSAIGuy it's time to revisit now. ;)

[@IMSAIGuy]

@@tonyfremont you are right. I just did. 2.7nS rise time for HC part 2.2nS for AC part. HC part nice and damped, AC lots of ring.

[@nreitcheck]

I looked up the datasheet that you were looking at and the specification was not the rise or fall time of the output but the propagation from the input to the output. Here is possibly the original source of the circuit. www.electronicdesign.com/technologies/displays/article/21762705/build-your-own-cable-radar A good build up of this design is. www.epanorama.net/circuits/tdr.html

[@nreitcheck]

I found it, the AC series is faster it is approximately 1.7 ns vs HC at 3.6. Here is a link to the document check page 5. www.mouser.com/catalog/supplier/library/pdf/FairchildLogic.pdf

[@aldob5681]

raise time is not the delay time