Great to see the 'Two Johnnies' in a collaboration, and I cant get enough of practical MicroCap use (I bet I'm not alone) very good presentation It's up there in your top tutorials...cheers!
Hi Eddie, Thanks for doing this. Nice description of the Miller effect. The JAT501 uses a 100pf cap for miller compensation. Looks like your sim is rolling it off early as compared to the real circuit.
#handry raya Thanks John! I forgot to do the simulation with the Boucherot cell removed and the input high pass filter removed. Is that what you did? Also, are you running a Bode sweep, or running a transient analysis? On the bench, do you get it to oscillate with a specific input frequency and wave shape? For instance, are you running a square wave?
@@KissAnalog No input high pass on bench test. I'm running a sweep. I ran step response tests with various capacitive loading and no continuous oscillation (only damped ring).
Hi Eddie, you know of course that the frequency plot considers the small signal characteristics only. The internal Re of the transistor changes with current flow in Ie, this could cause HF oscillations that will not be spotted in small signal analysis only. Now one must also consider that the dynamic gain of the transistor reduces at higher current and temperature. You can step temperature in an analysis as well to see what effect this will have. I have been following all of John's JAT amplifier design, and it is pretty well-designed, one could say a blameless design. However, some pitfalls to watch out for is that the transistors assumes a perfect curve match in MicroCap and this is not always true under dynamic conditions when all is not totally stable and fixed. The final nail in the coffin is of course PCB layout, which may introduce a bunch of parasitic elements into the design which can be tweaked after the layout is practically tested. John has however done this, and it proved to be good. The DC values will almost always correlate with MicroCap, but dynamic conditions may change somewhat. Earth planes and star connections cannot be simulated at all, and one has to assume best practice by the designer.
Great to hear from you Nico! You are so right - there's a big difference from small signal and power. Temperature and sensitivity analysis can help - and trying to place parasitics in the circuit. I've done some of this to show how different grounding systems can control current flow. Thanks for your feedback - I appreciate you!
Thanks - great question! Are you referring to the capacitor to counteract the miller effect? If so, it depends on the bandwidth, but I think a ceramic might be a great option - but then it also depends on the type of ceramic. I'd think that an NPO/COG type would be stable across temperature. The voltage rating is also important.
@@KissAnalog what about silver Film mica? , sorry for too much questions but I want to know. I'm very happy to ask and seek knowledge and apply that into the future for the good. Thanks for your time and effort 🤝 By the way I'm Aneesh Murali Nariyampully 😊 glad to meet you.
Nice to meet you! Yes, Mica caps would be good too - but they are physically larger and cost more. Sometimes a capacitor that fits better is better than one with long leads.
hey sir can you please help me build a 35-0-35v 10A smps power supply with feedback system and have constant voltage even on higher load. It should be a cheap and quite easy to make at home. I need this for my custom class-D amplifier of 250watts.
Thanks for the great tutorials, I'm interested in how to do FFT, THD, output impedance in micro-cap. I believe that others are also interested in how to make measurements correctly. Vinko
@@KissAnalog excellent. Thank you. I have a cheaper scope that seems rather noisy. Now i just need to figure if that's all just noise or if i indeed do have 350khz riding on the ripple.
Great video! Silly question because I'm easily distracted... How do you keep all the DMM's behind you in the ON state? I figured they would keep trying to go to sleep to save batteries.
LOL Great question! I once was advised to give people a background to look at, so I turn on the meters;) They also do remind me if I go to long blabbing;)
Ah - I get it, the Miller capacitor across VAS section and sometimes across drivers transistors used for compensation. BTW, how things would change if we used Darlington output instead of Sziklai outputs. Any cons/pros of one against the other.
Thanks for the great question! The Darrington would show the MIller effect even to a higher degree (gain) so it would have a lower bandwidth. The Szikai is configured differently so the Miller effect is not amplified. This might make a nice video?
Hey Eddie! I was thinking about how some amplifiers are 'clean, sometimes harsh, at worst". And some are "warm" sometimes muddy. I was wondering: is that measurable on some instrumentation? Cheers,Bill
Thanks Bill, this is a great question! I do think this is something that we can characterize with tests. This is what I think we should be able to describe so that people know what they want to buy.
What is your opinion on multisim? We learned multisim in college but when my key ran out I had to find something affordable(free) . I am trying to learn microcap and still struggling in it 😭
Thanks for the great question! I liked multisim when I was doing ac input circuits with rectifiers. One thing I had trouble with is I'd see unreal current levels and voltage levels if I did certain things - so I lost confidence. Micro-Cap is hard to beat as a free - or even pain for SPICE simulation tool. Let me know if there are circuits or specific things I can put in my tutorials to help;)
@@KissAnalog thanks very much!! We mostly used it in digital logic type circuits. By the time I got to advanced semi conductors my student key expired. I am getting into ac analysis for rf and audio as well as pulse width modulation (power supplies)so I will definately go back and look at all of your passed MicroCap videos for tips and tricks. I apreciate your kindness and great videos. Thanks very much.
ltspice is extremely similar to the spice 3f5 derivatives in many ways. The core matrix solver is very similar and almost all of the shared component models are meant to be identical or nearly so. The only real difference by default is the default integrator. Ltspice wasn't created for the sake of hiding models. Pretty much the only models in that category are some of tthe smps chips which aren't super important. I mostly use ngspice nowadays.
Thanks for the great feedback Daniel! LTSpice is a useful tool, but it does have its quirks. But I do feel that it was offered as a brilliant marketing tool. It is similar to SPICE but it is also not the same - or they would have simply used the SPICE as the core, but the idea was that they wanted something that would simulate fast - as that was impressive to the customers of their chips. I have had many a conversation with LT about sharing the guts of a particular chip so that I could better simulate it in SPICE, but they will not share. Working in the defense industry we often have to prove that the input or output impedance isn’t an issue, so I have to get this information to model. They are the only vendor that has turned me down. They just say use their models - but they are behavior based and do not show faults accurately. I’ve also been to their seminars taught be the creator of the software. In one class an attendee kept asking questions regarding stability and how to insure that the feedback elements were correct so the switching regulator just wont go unstable? The creator of the software finally admitted - these models just wont go unstable, so don’t worry. So, how do you like ngspice? I have used many tools, but that is one that I haven’t used. I think I was going to try it once but for some reason that I can’t remember, I didn’t use it. TI uses Tina spice, and Analog devices (the company that bought LT) uses SIMPLIS.
@@KissAnalog There are two distinct issues with regard to performance. First is the implementation architecture. LTSpice _should_ theoretically win in analog simulation by virtue of its JIT implementation. Since the JIT backend is custom, proprietary, and X86-only, we can't say much about how much effort has gone into the optimizing code emitter without profiling or reverse engineering. Assuming all else is equal and no implementation blunders it should generally have better runtime performance vs a traditional SPICE interpreter at the cost of up-front compiler overhead (like any other JIT.) IMO this isn't very important. This would only matter in simulations much larger than the vast majority of users are interested in. I don't know that anybody is doing serious VLSI work in LTspice. But it sounds like you're more concerned with mixed-mode simulation. There the problem is simply lack of an interoperable standard for mixed-mode simulations. Even if you had access to the model, porting the LTspice digital model to run as a pure SPICE behavioral model would likely be extremely complicated and slow. Much of that depends on the sophistication of the B-source script evaluator. There are ways they could be made faster in digital simulations by incorporating a type system and being smart about deciding which B-sources should be solved by their derivatives through the matrix method. I don't think that either ngspice or LTspice have bothered to do that yet. Ngspice has xspice, which is very powerful especially for library developers due to its plugin architecture. For mixed mode simulations it can be a bit clunky but certainly it can work well. User friendly it is not.
If stepping a value and plotting, when plotted double left click a plot, third tab you will find a box marked “Rainbow”. This will colour each sweep which makes selecting which plot was which step value, but you can also tie which separate phase plot matches which voltage plot. When modifying a treansistor or such, be careful if you don't want the change to be global, model types can update values across the same model type.
Eddie.... pretty please. Sorry to be a bearer of a negative subject but here goes.... You are so often so far off-mic on this vid it's making it hard to hear you & compromising your great work! I have you cranked up but the hiss noise floor of my computer isn't helping... May i suggest you hunt down a wireless mic transmitter and lavaliere style miniature mic with a clip for the shirt collar etc.... cheers :-)
Thanks so much for your Feedback! I'll see what I can do - to get a good one it is just another expense. I just upgraded the camera and added some lighting. Along with the test equipment and projects - I spend quite a bit on this. But, good sound is very important - and I did buy an expensive mic - but it doesn't move with me. I broke another mic which was also expensive. So, I do appreciate you and I'll try to improve this. Let me know what you think;)
@@KissAnalog Yes have seen many RUclipsrs say that the audio takes far more work than the video which can be very forgiving. I have worked for broadcast TV as a studio maint engineer most my career & look after the News crews kits ENG camera systems. They chew through a lot of gear & wireless mics always cop a hammering especially the lavaliere style mics. We go through soooo many Sennheiser Gold mke2 which are very robust but still end up wrecked... One the best lavaliere mics though is the TRAM omni TR50 for its sheer performance... These are expensive pro stuff though & there are plenty other options that are 100% great for YT production...
Great to see the 'Two Johnnies' in a collaboration, and I cant get enough of practical MicroCap use (I bet I'm not alone) very good presentation It's up there in your top tutorials...cheers!
Thanks so much! I think it will be fun;) I appreciate you!
👋🏻 just subscribed
Hi Eddie, Thanks for doing this. Nice description of the Miller effect. The JAT501 uses a 100pf cap for miller compensation. Looks like your sim is rolling it off early as compared to the real circuit.
Yes i know that your original sch use 100p.
His circuit with only 10p the amp will oscillate
Sorry i try again and it still not ossilate with 10p,
but it is with 1p.
With 10p it has low gain margin (bellow 10)
#handry raya Thanks John! I forgot to do the simulation with the Boucherot cell removed and the input high pass filter removed. Is that what you did? Also, are you running a Bode sweep, or running a transient analysis? On the bench, do you get it to oscillate with a specific input frequency and wave shape? For instance, are you running a square wave?
@@KissAnalog No input high pass on bench test. I'm running a sweep. I ran step response tests with various capacitive loading and no continuous oscillation (only damped ring).
Your Channel and your explanations are slowly becoming my favourites in electronics RUclips channels.
Wow, thanks! I appreciate you!
Thanks Eddy
Thanks for the great feedback! I appreciate you Bert!
Thanks for the tutorial.
Thanks Johnson - I appreciate you!
Really great video, you were killing it this time:-) It was one well-prepared video and very much easy to understand. :-)
Thanks so much Fried Mule;) I appreciate you!
Hi Eddie, you know of course that the frequency plot considers the small signal characteristics only. The internal Re of the transistor changes with current flow in Ie, this could cause HF oscillations that will not be spotted in small signal analysis only. Now one must also consider that the dynamic gain of the transistor reduces at higher current and temperature. You can step temperature in an analysis as well to see what effect this will have. I have been following all of John's JAT amplifier design, and it is pretty well-designed, one could say a blameless design. However, some pitfalls to watch out for is that the transistors assumes a perfect curve match in MicroCap and this is not always true under dynamic conditions when all is not totally stable and fixed. The final nail in the coffin is of course PCB layout, which may introduce a bunch of parasitic elements into the design which can be tweaked after the layout is practically tested. John has however done this, and it proved to be good. The DC values will almost always correlate with MicroCap, but dynamic conditions may change somewhat. Earth planes and star connections cannot be simulated at all, and one has to assume best practice by the designer.
Great to hear from you Nico! You are so right - there's a big difference from small signal and power. Temperature and sensitivity analysis can help - and trying to place parasitics in the circuit. I've done some of this to show how different grounding systems can control current flow. Thanks for your feedback - I appreciate you!
Love the Micro Cap tutorials. More!🙂
Thanks Paul! I appreciate you!
Which type of capacitor is better for this job, ceramic or film?
Thanks - great question! Are you referring to the capacitor to counteract the miller effect? If so, it depends on the bandwidth, but I think a ceramic might be a great option - but then it also depends on the type of ceramic. I'd think that an NPO/COG type would be stable across temperature. The voltage rating is also important.
@KissAnalog thank you, so go for npo/cog ceramic.
@@KissAnalog what about silver Film mica? , sorry for too much questions but I want to know. I'm very happy to ask and seek knowledge and apply that into the future for the good.
Thanks for your time and effort 🤝
By the way I'm Aneesh Murali Nariyampully 😊 glad to meet you.
Nice to meet you! Yes, Mica caps would be good too - but they are physically larger and cost more. Sometimes a capacitor that fits better is better than one with long leads.
hey sir can you please help me build a 35-0-35v 10A smps power supply with feedback system and have constant voltage even on higher load. It should be a cheap and quite easy to make at home. I need this for my custom class-D amplifier of 250watts.
Thanks for the great question! I am actually working on a video for this. I hope to post soon;)
Hi Eddie, thank you for the great video and amazing tutoring.
I appreciate your support!
Wow great tutorial video 🤩
Thank you very much for new episode
Thanks so much! I appreciate you!
@@KissAnalog welcome dear
really interesting vid,cheers
Thanks Stephen - I appreciate you!
Thanks for the great tutorials,
I'm interested in how to do FFT, THD, output impedance in micro-cap.
I believe that others are also interested in how to make measurements correctly.
Vinko
Thanks for the great feedback! I will cover those in future videos;)
👏👏👏
Thank you! I appreciate you!
Thanks for all the great info. When you scope a power supply for ripple do you turn on averaging?
Thanks for the great question! I have done that and it does work. I usually start by using the 20MHz bandwidth limit, and am good with that;)
@@KissAnalog excellent. Thank you. I have a cheaper scope that seems rather noisy. Now i just need to figure if that's all just noise or if i indeed do have 350khz riding on the ripple.
Great video! Silly question because I'm easily distracted... How do you keep all the DMM's behind you in the ON state? I figured they would keep trying to go to sleep to save batteries.
LOL Great question! I once was advised to give people a background to look at, so I turn on the meters;) They also do remind me if I go to long blabbing;)
What's up with your micsig? New GUI or some feature I'm not aware of?
Thanks for the question Jonas, but what do you mean? Is there something that you see that you didn’t expect?
@@KissAnalog the GUI on my micsig is black, so I was just wondering if you maybe have some updated software or so :)
Great content!
Thanks Mike - I appreciate you!
Ah - I get it, the Miller capacitor across VAS section and sometimes across drivers transistors used for compensation. BTW, how things would change if we used Darlington output instead of Sziklai outputs. Any cons/pros of one against the other.
Thanks for the great question! The Darrington would show the MIller effect even to a higher degree (gain) so it would have a lower bandwidth. The Szikai is configured differently so the Miller effect is not amplified. This might make a nice video?
sir can you test the frequency operation limit of a specific transistor? the iconic C5200 power transistor? Thanks.
Thanks for the great question! I'll have to do a video on this;)
Can you such lessons for miller in Tube amp PP
Thanks for the great idea Bert!
Hey Eddie! I was thinking about how some amplifiers are 'clean, sometimes harsh, at worst". And some are "warm" sometimes muddy. I was wondering: is that measurable on some instrumentation? Cheers,Bill
Thanks Bill, this is a great question! I do think this is something that we can characterize with tests. This is what I think we should be able to describe so that people know what they want to buy.
Awesome video!
Thanks!
What is your opinion on multisim? We learned multisim in college but when my key ran out I had to find something affordable(free) . I am trying to learn microcap and still struggling in it 😭
Thanks for the great question! I liked multisim when I was doing ac input circuits with rectifiers. One thing I had trouble with is I'd see unreal current levels and voltage levels if I did certain things - so I lost confidence. Micro-Cap is hard to beat as a free - or even pain for SPICE simulation tool. Let me know if there are circuits or specific things I can put in my tutorials to help;)
@@KissAnalog thanks very much!! We mostly used it in digital logic type circuits. By the time I got to advanced semi conductors my student key expired. I am getting into ac analysis for rf and audio as well as pulse width modulation (power supplies)so I will definately go back and look at all of your passed MicroCap videos for tips and tricks. I apreciate your kindness and great videos. Thanks very much.
Thank you! Let me know if you have questions;) Here’s the MicroCap playlist ruclips.net/p/PLZ0_iMoMBSslK1NFAXIEBsTnFzP8u9BGH
Where is the link to join Patreon???
Thanks for asking! I appreciate you! www.patreon.com/KissAnalog
ltspice is extremely similar to the spice 3f5 derivatives in many ways. The core matrix solver is very similar and almost all of the shared component models are meant to be identical or nearly so. The only real difference by default is the default integrator. Ltspice wasn't created for the sake of hiding models. Pretty much the only models in that category are some of tthe smps chips which aren't super important. I mostly use ngspice nowadays.
Thanks for the great feedback Daniel! LTSpice is a useful tool, but it does have its quirks. But I do feel that it was offered as a brilliant marketing tool. It is similar to SPICE but it is also not the same - or they would have simply used the SPICE as the core, but the idea was that they wanted something that would simulate fast - as that was impressive to the customers of their chips. I have had many a conversation with LT about sharing the guts of a particular chip so that I could better simulate it in SPICE, but they will not share. Working in the defense industry we often have to prove that the input or output impedance isn’t an issue, so I have to get this information to model. They are the only vendor that has turned me down. They just say use their models - but they are behavior based and do not show faults accurately. I’ve also been to their seminars taught be the creator of the software. In one class an attendee kept asking questions regarding stability and how to insure that the feedback elements were correct so the switching regulator just wont go unstable? The creator of the software finally admitted - these models just wont go unstable, so don’t worry. So, how do you like ngspice? I have used many tools, but that is one that I haven’t used. I think I was going to try it once but for some reason that I can’t remember, I didn’t use it. TI uses Tina spice, and Analog devices (the company that bought LT) uses SIMPLIS.
@@KissAnalog There are two distinct issues with regard to performance. First is the implementation architecture. LTSpice _should_ theoretically win in analog simulation by virtue of its JIT implementation. Since the JIT backend is custom, proprietary, and X86-only, we can't say much about how much effort has gone into the optimizing code emitter without profiling or reverse engineering. Assuming all else is equal and no implementation blunders it should generally have better runtime performance vs a traditional SPICE interpreter at the cost of up-front compiler overhead (like any other JIT.)
IMO this isn't very important. This would only matter in simulations much larger than the vast majority of users are interested in. I don't know that anybody is doing serious VLSI work in LTspice.
But it sounds like you're more concerned with mixed-mode simulation. There the problem is simply lack of an interoperable standard for mixed-mode simulations. Even if you had access to the model, porting the LTspice digital model to run as a pure SPICE behavioral model would likely be extremely complicated and slow. Much of that depends on the sophistication of the B-source script evaluator. There are ways they could be made faster in digital simulations by incorporating a type system and being smart about deciding which B-sources should be solved by their derivatives through the matrix method. I don't think that either ngspice or LTspice have bothered to do that yet.
Ngspice has xspice, which is very powerful especially for library developers due to its plugin architecture. For mixed mode simulations it can be a bit clunky but certainly it can work well. User friendly it is not.
If stepping a value and plotting, when plotted double left click a plot, third tab you will find a box marked “Rainbow”. This will colour each sweep which makes selecting which plot was which step value, but you can also tie which separate phase plot matches which voltage plot. When modifying a treansistor or such, be careful if you don't want the change to be global, model types can update values across the same model type.
Thanks for the great feedback! I appreciate you!
Meh, that was OK. You've had better videos.......
Haha jk. This video was awesome. I've never seen this circuit explained so well. Good stuff Eduardo!
LOL Thanks Joey! I appreciate you!
@@KissAnalog no thank you for abiding a crazy person and making him feel like he belongs lol
Your gear skyscraper has grown a bit. Don't get trapped under it one day :-D
LOL You are right! I gotta rebuild my bench - and sort in equipment that I use more;)
Who else at 12:13 shouted "Step it, Step it!" at the screen.
LOL Thanks Ian! ;)
Eddie.... pretty please. Sorry to be a bearer of a negative subject but here goes.... You are so often so far off-mic on this vid it's making it hard to hear you & compromising your great work! I have you cranked up but the hiss noise floor of my computer isn't helping... May i suggest you hunt down a wireless mic transmitter and lavaliere style miniature mic with a clip for the shirt collar etc.... cheers :-)
Thanks so much for your Feedback! I'll see what I can do - to get a good one it is just another expense. I just upgraded the camera and added some lighting. Along with the test equipment and projects - I spend quite a bit on this. But, good sound is very important - and I did buy an expensive mic - but it doesn't move with me. I broke another mic which was also expensive. So, I do appreciate you and I'll try to improve this. Let me know what you think;)
@@KissAnalog Yes have seen many RUclipsrs say that the audio takes far more work than the video which can be very forgiving. I have worked for broadcast TV as a studio maint engineer most my career & look after the News crews kits ENG camera systems. They chew through a lot of gear & wireless mics always cop a hammering especially the lavaliere style mics. We go through soooo many Sennheiser Gold mke2 which are very robust but still end up wrecked... One the best lavaliere mics though is the TRAM omni TR50 for its sheer performance... These are expensive pro stuff though & there are plenty other options that are 100% great for YT production...
Thank you for this info!! I have a juicy box for amplification so I’ll check these lavalieres out.