10 Tips for Analog & Mixed & OP Amp Designs
HTML-код
- Опубликовано: 26 май 2024
- What to consider when designing boards with analog, digital and op amps. Thank you very much Arthur Kay.
Other Links:
- Arthur's LinkedIn: / arthur-kay-a5201929
- Texas Instruments: www.ti.com/
- FEDEVEL courses: fedevel.com/courses
Chapters:
00:00 What is this video about
00:28 Floor plan - component placement
04:55 Return current
06:24 Crosstalk vs. height
07:18 Crosstalk vs length, spacing and thickness
10:41 Split planes, analog and digital grounds
14:29 Slot / split in reference plane
18:22 OP amp layout example
20:25 Decoupling
26:48 Electrical overstress
30:10 TVS diode protection
37:25 Component specification
42:06 Common mode noise rejection
46:34 Power supply noise rejection
49:30 Simulations
1:02:36 Measurements - don't rely upon them
1:05:51 Measure with oscilloscope
1:08:09 Clean your boards
1:09:21 If it works, maybe fix it
1:12:48 Use evaluation modules
1:17:45 Real example: Common mode noise rejection
1:20:58 Real example: Power supply noise rejection
------------------------------------------------------
Would you like to support me? It's simple:
- Sign up for online courses hosted on our platform: fedevel.com/
- You can also support me through Patreon: / robertferanec
- FEDEVEL Discord: / discord
It is much appreciated. Thank you,
- Robert 
Наука
Arthur wrote many good articles, glad to have him here!
A video on star ground vs. a continuous ground would be awesome. There is some very respectable equipment with star ground, but if a continuous rectangular ground is better, that might help a lot of new designs out.
My 2 cents here is that a star ground has the possibility to mess up your frequency response or stability if you have a design around a few hundreds of MHz. It's all about parasitic inductance.
I really enjoyed the last "if it ain't broke" part. I have had many clients over the years that strongly believed in the old meaning of the saying.
As soon someone tells me this these days I translate this to "we have actually no freaking clue what we are doing".
You either want to know all the ins and outs if a circuit, or prevent bad things from happening if you don't know all these things.
But just kinda believing and hoping the entire circuit just works is a REALLY bad idea for so many reasons.
The biggest one, which was really obvious last couple of years, if what happens when you're forced to change components (in this case I am referring to all the stock shortages).
Happy to see Arthur in this channel. His lectures are super insteresting.
very nice video this time. without fancy tools no-one can afford other than large companies(with the great respect for the legend Bogatin). This time great presentation with very good clear and concise examples. for almost every main concern od issues you can have on PCB design.
Have not seen more than 10 seconds of this video, but a like is already given. Looking forward to watching this video!
I'm from Russia. Robert you my you are my best friend, no matter what! We love you very much, all the smart people of America.
The "star ground" was used in analog power-amplifiers, where the resistance of the ground track was non-negligible, and the return (ground) current of single-ended loads (speakers) would create mV-level ripple on the ground-reference, which would play havoc on the analog front-end. It is a total non-issue with good differential driving and signaling.
I agree! Good point. There is always an exception to the rule. Best regards, Arthur
Thank you wey mutch for this video, I always watch your videos with great interest, because they are very informative and there are a lot of good, smart people. Thank you
This tutorial is great, Could you share the PPT with us?
Robert please do some video on magnetics design for smps
may be coming soon ...
@@RobertFeranec and also layout best practices for power supply design
Hello Robert, I watched more your videos it was very intereting and usefuly. I heve a question depend on DRC High-speed rusles.
As you know Altium allow define min/max peopagation delay, but only absolute valuse. is it posible use queries (or oyher trick) for check more advanced ? instaed of min/max compare with other value, such as propagation of delay other nets with defined tolerance.
Thanks.
From about 30:47 to 41:10, in a couple of different slides, an ADS8660 ADC is shown with a somewhat surprising input protection network. The "clip positive excursions" diode is shown clipping to the REF input rather than AVDD. I actually went and looked this up on the datasheet, and that's indeed what it shows! I am 50% inclined to think this is an error, as the datasheet section describing the required REF driver doesn't mention absorbing over-voltages.
I'm now very curious what the actual truth is on this detail! Arthur hedges his bets by putting TVS diodes on both REF and AVDD (where the latter would be unnecessary if clipping to REF is really how this chip is designed.).
Anyhow, thanks once again Robert and Arthur for a hugely informative video! Even when you cover points that we know already, these videos make fantastic references that we can direct colleagues or junior team members to for a particular topic.
Thanks for next great video!
Thanks for great video - I try always take notes in my notebook... As for Crosstalk - Arthur show microstrip line... what's about microstrip? It seems distriburion of return current will be other - it's interesting for me watch such pictures like microstrip in this video...
Thank you!
Hello Robert, thank you for the very nice and informatice video, this video raised a question for me (regarding operational amplifer layout), do we need to remove the GND and PWR plane just under the inverting pin and outout pin? or not only we need to reomve GND and PWR plane under the inverting pin and outout pin, but also we need to remove the GND and PWR plane under the feedback track?! if yes, can this make the feedback track noisy? Considering there is no return path under this track? thank you for you help inadvance.
Nice!
спасибо, хорошая работа.
Correct me if I am wrong, but at 4:30 wouldn't you want the amplifier as close to the DAC as possible, and the DAC filter as close to the amplifier as possible. That way you minimize the length where lower voltage level signals live and thus can pick up noise, and by placing the filter as close to the amp as possible, maximize the length between DAC and filter (which might pick up noise) which will be filtered?
You are correct. I should update the slides. I would only comment that the DAC output is already a high level signal, so somewhat less susceptible to noise than the low level amplifier input. But ultimately I think you are correct. Best regards, Arthur
Its antiailiasing/reconstruction filters, so in both cases it's no. You are avoiding crosstalk of unwanted spectral content from/to the data converter. "Lower voltage signals" makes no sense here, what you should be thinking about is signal source impedance in first place.
@@Konecny_M It might be an antialiasing filter, but it still restricts the bandwidth over which noise is captured. It might not be its primary purpose, but it still does limit the noise bandwidth. And 'lower voltage signals' is relative to the voltage levels after the amplifier. If you capture any noise before the amplifier it gets amplified by amp resulting in noise levels higher by a factor of the gain of the amplifier on the output signal.
Video on PE earth connection on PCB noise
Nice video, thank you a lot!! What simulation software is used in 50:00
Tina TI