HTML-код
- Опубликовано: 1 окт 2024
- In this video we explore the good, bad and ugly aspects of switching time measurements on bipolar transistors (BJTs). We find that the switching time measurements (rise time, fall time, delay time and storage time) are all highly dependent upon the test circuit and test conditions - and that these conditions are not standardized by any particular device type of manufacturer. Notes for this video are here:
www.qsl.net/w2a...
Hi Alan,
Very informative video with excellent visuals. Deserves more than just a thumbs up. 73 WB3BJU
Another great tutorial Alan. Thank you very much for taking the time to put these types of videos together to help folks that really want to learn the art of electronics. Your presentation is always clear, professional, and easy to follow. As a content creator myself I know this takes a lot of your time to put these together. You really have to do it for the love of the hobby because the pay from RUclips is very little. What really gets me is someone comes by and gives a thumb down. This always makes me wonder why? Not like folks are paying for the free information. Alan I appreciate what you share.
How about FETs, can you please make a video about their switching parameters and to compare VS BJTs ?
Me too
Hi w2aeq. Great video. Do a follow up video where you add a speed up cap across the base resistor and show how it actually speeds up the switching speed. Then do a bakers clamp circuit which keep the transistor from going into deep saturation and thus mitigate storage delays. Great video. I will probably do a similar video on this subject on my channel. Robert
Agreed! I'd love to see a video clearly explaining how to get much faster switching with a BJT
This devil in the details kind of video really gives you valuable insight into what's going on. Thanks.
Sir, I am an electrical engineering undergrad and i give you standing ovation on your every video. You've helped me understand many concepts and i consider you one of my lecturers. Keep up the good work!!!
You would be hard pressed to find a college lecturer that is able to give the hands on experience coupled with the highly technical theory side and the maths as well as Alan can. I can totally see your parallel as him as a pseudo lecturer.
I agree. You can learn a lot from Paul at Mr. Carlson's Lab as well.
no offense , but why are you telling everybody who knows something a Sir? That corporate culture long dead by now in everywhere in the world. Every exists in military and polices. Just address him with his name, he got a name. Just enough respect to the person who learn from and his time and courage is enough. If he have a Dr rate then address him with Dr. but not Sir, Sir is too formal.
@@dhammikaperera4059 Hi Dhammika. Not everyone (or every culture) is comfortable addressing someone by their first name, especially if they haven't met them in person. I agree with you that it seems too formal in some cases, but I consider it a title of respect.
A very nice video as usual, thank you!
P.S. I really love the 'good, bad and ugly' thing.
Fantastic video! Both enjoyable and informative.
Thanks for the excellent video as usual! I'm going to have to put together one of these test jigs - especially using your method of connecting probe leads to the pcb. Nice trick!
Wonderful video, Alan; I learned a lot. Many thanks. It would be great to see a follow-up showing a similar discussion and testing with MOSFETs. The varying gate charge characteristics -- especially in large power FETs -- would make for some interesting measurements. That, in turn, could lead to a video about how to properly drive a FET.
Again, thanks. Your videos are the best of their kind!
Hi, Alan!
I have finally received my first ever HAM license on Sunday and wanted to give you a big thanks, since I consider you to be one of the people who inspired me to take up the hobby!
Going down memory lane a little, I had first found your channel because I wanted to find a video on scope triggering just before I bought my first scope. A Tektronix 2235, by the way(Yep, I'm one of the family, too, lol). That was in November last year. And then I watched those seminars on scope applications with antique radios that you had hosted and it all went down from there haha!
Anyways, I wish you many healthy, productive and fun years ahead!
73s!
Congratulations on earning your ticket - that's great! Very nice to hear that I've helped you along the way.
Hi Alan,
thanks for such informative video.
I have a question, when you were testing the RF diode and measuring the rise and fall time, I think you have to zoom in more horizontally as the scope relies on the horizontal resolution to give you an accurate rise and fall time. it would interesting to see what they would be on 400ns or 200ns per box scale if you still got the setup up and running.
That is a very good observation, and worthy of a little bit of discussion. Many (most?) digital scopes make their automated measurements on the waveform points shown in the display. Since most scopes have 1000 points (or less) resolution across the display, measurements like rise and fall time on very fast edges can be inaccurate if the horizontal scale is too slow and the edges are nearly vertical. It turns out that Tektronix scopes operate a little differently - they make their measurements on the full waveform record - not just the displayed points. You'll noticed from the annotations in the bottom of the scope screen that I have the scope set to use a 10,000 point waveform record, and the scope's sample rate is running at 1.25GS/s, based on this combination of record length and horizontal setting. That means that there is a 800ps interval between each time point. Therefore, there are probably 18 or 19 sample points on the 15ns edge, so the accuracy is still good. In fact, the scope will give me a "low resolution" warning if the sample rate isn't sufficient for a given measurement. As I said, a very good point that you raise, and something to really watch out for on many other digital scopes.
Thanks for the clip. Very interesting subject. BJTs are not really fast switches, FETs and even faster super conductor devices are better at it.
However, there are circuits to improve the switching speed of BJTs, like totem pole for MOSFETs.
In my opinion, it would have been good it you monitored emitter current to demonstrate that during transition net current entering BJT is not zero! (That is Ie != Ib + Ic during the transition)
Very good, I would expect to get this education at a higher level of college. My thoughts on this are that many transistor circuits used as a switch to take the load off of a microcontroller, or in some other similar circuit, won't matter much. But when the circuit involves much more complicated switching times that correlate with switching times elsewhere in the same circuit, that is where the selection, and tweaking of control parameters of each stage are important. The testing you are doing can be turned into a book of practical transistor data. You can simplify what each testing parameter means, and list the transistors in or of best to worst for that particular test. I would buy a book like that. In most cases, simple transistor circuits can work around any common transistor, but even those have their limits. Usually those limits are in power handling, and how fast or slow they are isn't of much concern, as long as they go off, and on with no noticeable delay. Other circuits involving audio, have requirements about low noise, and that is another area that can be tested. Thanks for the excellent way you explain things in great detail.
Really good explanation!
Thanks for this great video!
Hi Alan, Excellent explanation. How does the hFE of a transistor affect its rise and fall times? Thanks a lot.
I didn't measure that, but I suspect that the affect is minor.
The hfe of the transistor change the storage time and a littlebit the switch off time. The storage time gets longer for high hfe transistors due to effect that the electrons in the charged base have to flow primary back to the base drive circuit. The discharging path from the base to the emitter is therefore less effective for a high hfe transistor. Cheers from Berlin
@@christophschuermann7920 On the other hand, a high β transistor may be able to be driven into saturation with a smaller base current. Sometimes the manufacturers specify Vce(sat) for high-gain transistors with Ib = Ic/20 instead of Ic/10. In that case, there will be less stored charge to remove, so it's somewhat "swings-and-roundabouts".
I fully agree with all of the positive comments. You are certainly a 5 star and/or Gold content provider on YT. Thanks for all your time and dedication!!!!!!!!!!!!!
Great video.
I've been wanting to do the same comparison between genuine 2N2222A transistors I've had for years in my parts bin, and the 100 I bought from AliExpress for real cheap (< $2), to see if there are any rise and fall time differences between them.
Unfortunately, I haven't had the time to do that yet, and I was wondering if you've conducted such a test?
Do you plan on doing any such tests in the future? Or a way to tell the difference between genuine parts and the Chinese knock-offs. Maybe a test fixture like what you have?
A homemade fixture like mine could certainly be used to understand the switching behavior differences, if any. But, other tests might be necessary depending on the application (amplifier, etc.). I haven't purchased any cheap knock-off parts, so have never done any comparisons.
Thanks. I had to ask before doing it myself. I will give it a go since I need to know if the Chinese stuff is worth buying or not.
Well done!
Thanks Alan! Another area of misty understanding gets blown away by the clear breeze of reason :-) I'll definitely be building one of these. Are the switching times pretty consistent among batches of transistors of the same type?
Sorry, one more question; is the switching time, particularly thinking of the on time, what gives rise to the propagation delay in TTL ICs?
Switching times should be consistent from the same manufacturer for the same device type - there may be more variation when looking at other manufacturers of the same device. I would imagine that the TTL propagation delay is dominated by the storage time of saturated transistors, but I haven' studied this to state it definitively.
Best preparation, best explanation, best voice and best demonstation avalable on the internet. 👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍
👏👏👏👏👏👏👏👏👏👏👏👏👏👏
great video. I wish you would also show, what does antiparallel diode + resistor to gate resistor do to waveforms
I just love your videos! Not sure if you have a video on the subject but a video on the power dissipation during switching (MOSFETs, BJT's) would be fantastic. Thanks again for your videos!
I did a video on that, but it is on the Tektronix youtube channel: ruclips.net/video/e3GjBcjP6e0/видео.html
With all the trolling on YT and FB it's a great comfort to know that guys like you exist and provide straight up great information with no BS. The internet is often a sewer, but there are oasis of intelligence here and there that make it worth it. Thank you!
Wow! Thank you very much for this tutorial. Please keep them coming.
Excellent
Great video as always! Thank you very much.
Also it would be interesting to see Ts & Tf times without negative bias, just 0V at the base. I think Ts will incease by ≈50% or so. Still I'd think it's kinda missing in the video.
But again, it's great anyway.!
Yeah, I should've done that.
i tried measuring the rise time of a mps2222a recently and now i know one way of how they do it. thanks, this helps a ton.
Awesome video Alan, can you do this for FET's?
Great Video, you really helped me alot
This also means that with sawtoothish drive waveforms you can shorten the storage time while still having a square wave at the collector. This may come in handy at some times where you want to minimize pulses to spread out
I really like your videos !
What a nice surprise, I love this style of videos.
Thank you!
:o)
Excellent but... it is a bit annoying when we are looking at your paper text and diagrams and it keeps moving around. I would suggest putting the paper on the desk and not touching it while talking. We all shake or move papers while doing this but on video (Hfe of our screens to the real paper size!) it can be a bit distracting. *That being said, as always I really enjoyed this video. Thumbs up.*
Thanks for the constructive feedback, I will try to be more aware of this in future videos.
Good to hear. I am a firm believer in providing constructive feedback. It's all about getting the wording right. Keep up the good work. And remember hFE is everywhere! ;-)
Thank you very much for this visual. You did this BJT set up with a basic common Emitter circuit, I would really be interested in seeing it also done in a Common Base, and also a true Common Collector (reverse mode) set up if you can have the chance. Thanks again, and keep up the excellent work!
Man, I just realized how much I missed your back to basics videos! Hope all is well, Alan
This material would take someone many days to understand reading from a textbook.
Yep. That's why I like to go to the datasheets for transistors and other components and build a test circuit on my breadboard and play around with them. Building a test jig like Alan's would be much better.
Great video.
You are very good at this. Thank you for wanting to do this for us all
Well the BJT performance in switching is generally very depressing, you almost have to be desperate to use them (see IGBTs) :-P
Thank you very much!!!!!
I come here and watch your videos periodically and am really amazed at your vast knowledge and really superb delivery. But mostly I want to say thank you for taking the time to do this at all. You really help a lot of people all over the world understand things much better. Thank you. Mike K8MB.
First, I am VERY glad that you got the rise and fall times correct; I am disappointed by how many engineers get those backwards. Second, I love your probe test points and I wonder where you get those coil springs? Third, you can reduce the "storage" time by using the 74LSxx trick of placing Schottkey diodes between the base and collector. Fourth, the fall time MEASUREMENT depends substantially upon the collector resistance/collector capacitance RC time constant -- 220 x 10p = 2.2 ns -- the minimum fall time you can measure might be 2.2 ns if the Coss is 10 pF or larger. Whatever the Coss, the intrinsic fall time will be RC shorter than the measurement.
We work these same dynamics with the Klystron on Radar Systems...amazing devices.
You explain things so easy to understand! I wish I had these videos when I was in college.. (25 year ago LOL)
Extremely very good video! Thank you very much and it give deep help! Thx!
Yes digital circuits went from TTL logic families, PMOS, NMOS, CMOS. Digital circuits had switching loss issues that is why they settled with CMOS for faster switching times and lower power consumption.
Very interesting and informative video. Thank you very much and keep up the excellent work ✌🏼
you are the best teacher ever! 73 PD1FR
Good video, Thanks... more videos like That, please...
The differences in the generic house types is quite revealing. Some makers probably don't do much extensive testing if at all. Guess you get what you pay for especially when it comes to speed and quality.
I stumbled upon the phenomenon of storage time last night while testing a mosfet driver circuit. Interesting stuff!
Missing examples with the other values of base resistor and collector, when it doesn't get a deep saturation. Anyway great vid.
It would have been interesting to see the switching speed vs temperature, like get a hot air gun or better yet an adjustable soldering iron so you can look at the relationship between temps and operating speeds.
this was very insightful. thank you very much!
It's Awesome Alan! I am refreshing all my undergraduate concepts from your video before going to my master's :P
All of your videos are very interesting. Thank you.
Like 1K5 Dislike 4 that set a new RUclips record almost 400:1.
The dislike perhaps was happen accidentally.
I've got several videos that have better than 99.5% likes vs dislikes. But, there's always a few.... The video on the basics of FETs has over 3400 likes with just 12 dislikes...
I thought youtube have to ban such a porn!!!! Thumb up!
Oky have learned the concept. Liked...
If I'm not mistaken, at 05:28 you said that the delay time is measured from 90% of the input waveform but you pointed at 10%.
Good catch - a little verbal mistake. (every one of my videos has a few, unfortunately).
I really appreciated this video. Thanks!
Hi Alan, appears that the link to PDF is down. Getting error "Destination
Unknown
Everything's working on our side, so the short link you clicked is either wrong or has been retired."
I've updated the link. Here is the link copied here as well:
www.qsl.net/w2aew/youtube/bipolar_transistor_switching_time.pdf
Question: When the drive voltage went from negative to zero volts, right at the end of the sequence, you can see a llittle gulp of current on the base. My best informed guess is that that is the capacitance of the base being charged up from -ve to zero. Is that correct?
Yes, I believe that you are correct.
Thanks for sharing. Very interesting topic.
The Switching Loss means lower output voltage and lower amplitude? I'm confused on what the loss the loss of what?
The loss is a loss of efficiency. Switching regulators are used when high efficiency is needed. Efficiency is the measure of power output divided by the power input. The desire is to have this as high as possible. Any power (heat) wasted in the power conversion circuitry is lost and not delivered to the load. Thus, the power dissipated by the device when it is switching on or off is wasted power, and is referred to as switching loss.
Thanks yes i get it. But most circuits specifications don't list in the spec the switching loss value they only tell you the power efficiency. So the switching loss value is included in the power efficiency specification it seems its "not separated" in the specifications? What I'm confused about is that all transistors and FETS will have a resistances between the switching transition which is what the switching loss it from the heat dissipated but how would the transistors and FET components have very low switching loss when you will always have that transition resistance and heat dissipation while the transistor and FETS are switching on and off. Yes lower voltage and faster switching frequency helps with the switching loss and heat dissipation but still the transistors and FETS transistor resistance will always do a problem. I'm not sure how the EE designers will get away from this because it will always be there so I have no idea how to lower switching loss. What else can lower switching loss?
There are different sources of losses in switching regulator circuits - the manufacturers typically list only the total efficiency. It is the *designers* that will measure individual loss sources such as switching loss. Switching loss is worse with faster switching frequencies because you switch more often. The key to minimizing switching loss is to switch quickly, so that the transistor doesn't spend much time between the "fully on" and "fully off" states.
I was hoping you would show without negative base voltage, just going back to 0v.
Storage time and fall time will both increase. I observed a 50-100% increase compared to -1V drive.
The higher the switching time , the higher the switching loss , the smoother the signal? The Switching loss means that the transistors or FET's output amplitude voltage and current is lower when the switching time is higher in frequency?
A switching transistor dissipates the most power during the switching time. So, faster switching times are desired to minimize loss and device heating due to power dissipation.
Thank you for all the detail!!!
Very useful. Thanks !
Hi Alan, sorry to be a pain, but I wondered if you (or your subscribers) know of, or have a way of finding out, what a recommended replacement would be for RT1010 (NTC thermistor, 7.5 Ohm, 10% - I think 7A) on the A2 regulator board of the 24XXB 'scopes? There are two, the smaller, inboard one is not a problem, as the SG200 (5 Ohm, 5A) is still produced.
There is one p/n for it in the 1989 Service Manual for the 2465B/2467B (75DJ7RK5-RO-220), and a Tek p/n in the 1993 edition (307-0350-00). The closest I've found so far is a 7 Ohm (within 10%) 10A part, but it is quite a bit larger and radial rather than axial.
There is a gap at the junction of the leads and encapsulation of the old one, and the cold resistance has shifted upwards (I'm guessing moisture ingress thru the gap has caused this) to just under 10 Ohms. I could dry it out and re-seal the leads with very low viscosity epoxy, or use the over-sized part I found, but I'd really like to get a new part, closer to spec as it's in there for current inrush limiting on start-up.
I totally understand if you are too busy, but maybe one of your viewers has the answer.
I think I may have an addiction problem developing, though. I've got a 2465B n its way, an Ex-Sony Gmbh. owned one with a Tek cal sticker on it from '09. Happy days!
It couldn't hurt to send a note over to the nice guys at VintageTEK.org. They may know of a replacement, or even a source of an original.
Great idea, i shall give it a whirl! Thank you.
It would be interesting to use these measurements to derive the SPICE models parameters for the part. If done correctly, the simulation should match the measurements over a wide range of conditions. Too bad the manufacturer's don't do this (correctly) for us. The procedure is in "Modeling the Bipolar Transistor" by Ian Getreu, and in the ICCAP software from Keysight EEsof.
You've giving me a better understanding what's really involved with "switchers". They were somewhat mysterious to me with my "analog" mind. Thanks for taking the time to share!
Great video! That is useful in my work. I’m curious if this would vary over temperature. My intuition isn’t very good, but I would guess it would, mainly since almost everything a transistor does varies with temperature. You’ve given me some good ideas for some experiments. (I’ve only got a 2021, so I’m suffering from AFG envy though.) Thanks!
long live The Frisca
How fast would the fall off be if you didn't drive the switch off negative?
Both the storage time and the IC fall time will increase. For the couple of transistors shown, these parameters increase 50-100% without negative pulse of -1V.
Hi Alan, someone asked me what exactly does "PAD" stand for? Passive Attenuation Device? Or something else? I looked in the ARRL handbook and interestingly there isn't a definition in there. Thanks as alway, Vince.
I always thought it was because they "pad down" a signal, but I really don't know the true answer.
No worries, Thank you.
Great video, would have liked to see some big TO3 power transistors in the mix.
Let's see SCR's and TRIACs
Fantastic. Thank you. If you change the "rest time" between AWG pulses, does anything change within the "operating frequency" of the DUT? I haven't seen much information about switching performance near the speed limit. How do BPJs fail as they approach their speed limit?
Thanks for another good lesson.
That was very interesting results!!!
I have a ---- IRF 250 (200Volt @30amp Power Mosfet) that I have been playing with, and was wondering if you could run the same tests on it, as the video I just watched. I don't have near the type of test equipment that you have. I've been working on running a Kind of 3Phase motor I wound, And can run it up very fast with different input pulse rates, but won't run on any kind of load. Maybe they are not used for that kind of use.
Nice video Alan. Have you considered a follow-up video showing how to speed up the switching by using various techniques such as Baker clamps, parallel cap across base resistor and proportional base drive? RBSOA explanation would also be nice to see.
Its almost a lost art nowadays with high speed MOSFETS.
Very interesting thanks
If you are looking for ideas to discuss, perhaps how to use function modules from china and how they vary from real test equipment might be an idea. How good are they and where to use them. I just ordered a AD307 module and wonder just how good are these and others like broad band amplifiers, also AD active mixers.
These functions are very cheap and seem, maybe, very high performance for the price. Can we replace these functions in older radios with these modern equivalents that the early engineers never had?
Ken
K7KLW
I found the answer in the #111: How to make a high performance oscilloscope probe socket.
Thanks...
you should show how long the storage time is with 0V turn-off drive
It will increase - typically by 50-100% compared to -1V drive.
Very cool thing to experiment with and again a really nice, well explained video. One question though - do the parameters vary much between transistors of the same type? So will two different 2N2222 transistor exhibit about the same times?
They *shouldn't* vary much among the same device type from the same manufacturers - but beware of different manufacturers' devices - there may be some additional variations.
Thanks!
EXCELLENT. NUFF said.
Hello Alan,
thanks for these precise explanations abour switching time. Great video. What kind of wire do you use for your homemade probe connector?
I'm curious how the architecture is different in the transistor to change the parameters so drastically. Excellent stuff!
Awesome video, dude.
Excellent presentation Alan :) !!! Showing the influence of both, the input signal and the operating conditions, which each manufacturer uses to their advantage, to improve the results in the tests of their devices.
Nice little fixture.
Very interesting. While not related to bipolar transistors I would have found it interesting to see how a small fet would compare just to give a visual comparison of the switching times. Datasheets tell me the fet would be maybe 5x to 10x faster but I guess it really depends on the fet.
Awesome !...cheers.
Awesome. Could you make video on how to use the function generator in conjunction with the oscilloscope. Not the 50 ohm story, but actual usage as you (too) briefly mention in this video.
I have the scope and function generator (sadly Rigol, please don't be offended), but I'm not feeling safe about how my fg and scope can be connected in applications.
Great video as always 😊 thanks.
Also please make some videos of how vintage radio controlled cars were built. I think my equipment only have 30 mhz
Generally, the only way you typically can get in trouble when using both your function generator and oscilloscope is to be careless about ground connections. As long as the scope, generator and power supply grounds are common, you're unlikely to damage anything.
@@w2aew Thanks very much,
I have a Rigol (Scope DS1054z and Function generator DG1032z)
On the FG there is a input counter on the front. Can I connect a FG to that one to count the frequency? What waveform does this kind of input require?
Meaning what is the real life application?
On the back of my FG there are [CH1&2/Sync/Ext Mod/Trig/FSK] What is the difference between sync and trigger?
What inputs and output is applicable and what are real life usage?
What is real life usage of burst and sweep?
In this video they discuss the setting of 50Ohm vs. high Z.
ruclips.net/video/3PYAez9CmcQ/видео.html&ab_channel=Rose-HulmanOnline
How do I determine the impedance of a circuit in a way that I can adjust the setting correctly (and what are the boundries herefore)?
Is electronics always high impedence as a general rule?
Would you consider to make a video on usage of the function generator in the same way you demonstrate the usage of the scope in your videos?
To you this might be silly questions, but I struggle and I also sense that others do and it would help us get going. Thanks in advance.
Also a 101 video series get started with building circuits would be fantastic. E.g. a radio controlled device (below 30mhz) and all sorts of 70-80's electronics.
Again thanks for the best videos on the web :-) The world needs more of this :-)
@@Infinitesap That's a lot of questions. Most of the answers are in the user manual for the function generator. The frequency counter is typically used to measure the frequency of a signal being produced by a circuit you're working on, maybe an oscillator or clock generator. There's no specific signal type is it is expecting, just that it is repetitive and periodic. The SYNC output is a pulse that coincides with each cycle of the waveform being generated, most often used by a scope to trigger on. Ext Mod is an input for when you're generating a waveform that is modulated (AM, FM, etc.) from an external source - such as using an audio signal to modulate an AM carrier. The TRIG can be an input or an output, depending on how it is setup. Trigger is usually used in Burst or Sweep mode. As an input, it accepts a signal that initiates the burst generation or the start of a sweep. As an output, it is a pulse that coincides with the start of a burst or a sweep. Burst is often used when only a specific number of cycles of a waveform are needed to test a circuit or device. A sweep is used often to measure the frequency response of a circuit.
Good and informative video. Thanks. I've been aware of MOSFET's gate capacitive issues, but your video has taught me bipolars aren't perfect either.
The test was interesting, but would most applications use negative voltages to pull a transistor low to deplete it in switching application?
When you mentioned that there was little of a test setup, I immediately thought of level shifting from 3.4v to 5v, or vice versa.
Driving an 30ma LED from either a Raspberry PI's 3.3 paltry "general purpose I/O" pins or driving one or more LED's from an Arduino 5v pin would be an interesting response time test. I am not even sure if a PI can even power most opto-isolator/couplers, safely.
An Arduino can handle an LED from a pin, but not more than one. For current draw LED's 20-30ma seems like a nice current draw for a test, though biasing aside most Arduino people just use a 1K resistor for current limiting because they close enough, most people have them, and when you are holding a hammer, everything looks like a nail. : )
Thanks again!
True - most applications wouldn't pull the base negative to turn off (unless it's specifically designed to minimize storage time), but that's the way that the test conditions are defined by the manufacturers.