Stack Exchange
log in

Peter Smith

 Discussion on answer by Lundin: Is au…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Mar 25, 2021 15:37
I have used Simplicity studio a lot, and many of the 'auto generated' features did not work as advertised. I also detest the libraries they provide as they are so opaque. I had to dig through all that madness when I needed to implement DMA (which is somewhat unusual in the earlier 32 bit parts). I ended up simply using the CMSIS headers and doing direct register access. Didn't really take long to get up to speed that way and the code was definitely cleaner. On a different note, I remember the auto generated Windows C code (from the Borland tools) which gave me the switch statement from hell.
 

 Discussion on answer by Raffzahn: Doe…

Imported from a comment discussion on retrocomputing.stackexch...
Peter Smith
Mar 12, 2021 11:54
When feature sizes shrunk below about 120nm (IIRC - it has been a while) it turned out that quantum tunnelling was becoming a major source of leakage and the designers had to play all manner of tricks to deal with it.
 

 Discussion on question by Essam: unde…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jan 18, 2021 19:25
Modelling this with a perfect diode is not the way I would go; the whole point of the circuit is to make the diode appear to be perfect.
 

 Discussion on answer by Raffzahn: Pro…

Imported from a comment discussion on retrocomputing.stackexch...
Peter Smith
Nov 6, 2020 18:47
The 'rant' is amusing, Where I currently work, one of the junior engineers has been tasked with making a moisture sensor that can record the moisture levels at hourly intervals and operate for up to a year or so. The proposal includes a 32 bit microcontroller running Linux. I (and many others!) could probably do that with a 8 bit device on bare metal.
 

 Discussion on question by P_M: In thi…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Oct 26, 2020 03:16
Vout will be high under two specific conditions; those conditions are the ones that make the outputs of the comparators go high - high in this sense is where either diode is on.
Peter Smith
Oct 26, 2020 03:16
You might find the solution easier if you assume the diodes are ideal. It might also help you to put values on Vu, VL and then vary Vin.
 

 Discussion on question by equave: Why…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jul 21, 2020 14:47
Have you pulled the RESET input low with a relatively low resistance? This pin is adjacent to the clock pin and might occasionally be triggered by the clock input by crosstalk (it has happened to me in the past).
 

 Discussion on answer by Andy aka: LM2…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jun 9, 2020 18:45
I have also noted that the volt microsecond rating is not stated, but it is a critical part of the design procedure.
 

 Discussion on question by Sqoshu: Tes…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jun 3, 2020 15:04
You can find the LM111 spice model at ti.com/product/LM111/toolssoftware which will give a more accurate representation.
 

 Discussion on question by Trevor Mers…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
May 28, 2020 07:28
Do you know that there one instruction set computers? Certainly makes the test cases easier :) en.wikipedia.org/wiki/One_instruction_set_computer
Peter Smith
May 28, 2020 07:28
Thousands? As I recall, the 6502 had a total of 51 base instructions and a total of around 150 or so (as noted by DKNguyen instruction decoding requires hardware to do it and there really wasn't that much space for more instructions). It was still a very capable processor for it's time.
 

 Discussion on answer by Andrew Steane…

Imported from a comment discussion on physics.stackexchange.co...
Peter Smith
May 26, 2020 08:10
only do in a restricted sense or in some limiting case - To the best of my knowledge, every law of physics is strictly bounded for that very reason :) Famously, it took 19 years for Ohm's law to become a law.
 

 Discussion on question by p1ns: Ident…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Feb 9, 2020 13:43
Do you have access to a logic analyser? It would probably not need to be particularly fast (perhaps 10MHz). Using a 16 pin dip header and even 4 channels could yield significant information about what is going on.
Peter Smith
Feb 9, 2020 13:43
The CCA is very likely Channel Controller Assembly; General Instrument (from which Microchip came) were pioneers in electrically alterable ROM, which supports your position that the device contains mileage information. I no longer have the databooks from that time after many moves. I would also suggest that you ask a related question on the retrocomputing site (asking about microchip channel controllers in particular).
Peter Smith
Feb 9, 2020 13:43
From the CCA item, I believe it is a very early channel controller based on the 8 bit PIC (and one time programmable); this part became available at around the time you mention (the original PIC one of the few surviving processing devices after Microchip was spun out of General Instruments and it was updated to be an I/O controller). I will note that the original definition of PIC (from GI) was Programmable Interface Controller. There are no 16 pin controllers from Microchip now, but there certainly were in the early 90s.
 

 Discussion on question by Bloodmoon: …

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Feb 3, 2020 13:42
The Nyquist sampling theorem relates to the maximum frequency of the signal being measured, which in your case is a constant. The rate at which it is being measured must be > \$2f_{max}\$ (not equal to 2f).
 

 Discussion on question by Hugo Sereno…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Dec 31, 2019 12:52
Something that may be contributing; Volp (not in the datasheet but a known phenomenon) - when some outputs switch and the other inputs are held stable, those other outputs will be affected and the effect is via the shared internal ground. See ti.com/lit/an/szza038b/szza038b.pdf
 

 Discussion on question by K.Mulier: D…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Oct 30, 2019 21:01
@K.Mulier - you can get plates pre-drilled with multiple drill holes at regular intervals so all you need to do is simply space the PCB mounting holes accordingly; here is an example: amazon.co.uk/Heavy-Straight-Fixing-Plates-Screws/dp/B07MVJ7X‌​H1
Peter Smith
Oct 30, 2019 21:01
I usually leave holes for stand offs and screw them into the board with the other end screwed into a metal plate.
 

 Discussion on question by vianna77: D…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Sep 8, 2019 11:51
@NickAlexeev - the comment is on page 3 of the first chapter (foundations) in the third edition. My first edition has long been consigned elsewhere due to the many moves I have done since getting it in 1982.
Peter Smith
Sep 8, 2019 11:51
@NickAlexeev In "The Art of Electronics" they refer to amperage (and wattage in the first edition) as "...that's strictly bush league" (which I agree with - we have the perfectly usable terms current and power or current rating and power rataing when talking about such things).
 

 Discussion on question by Max: RC Ana…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jun 6, 2019 15:52
I suspect signal loading due to the cascading of filters; it will become clear once you post a schematic.
 

 Discussion on question by MFlop: How …

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Apr 6, 2019 22:01
The collector voltage at 7260 and at the top of resistor 3280 is the same. My calculation shows 400 microamps through 3280 and that leads me to believe that diode 6262 is short circuit.
Peter Smith
Apr 6, 2019 22:01
@Ale..chenski - this may be trivial to those who have dealt with such amplifiers, but it is hardly trivial to someone who has not. I would agree that in all likelihood one of the driver transistors has failed. Upvoted as an excellent question.
 

 Discussion on question by Michel Keij…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Jul 16, 2018 13:06
It has been a while... With both CANH and CANL the same at about 2V, they are recessive (not driving). Most logic analysers allow you to change the trigger levels and most default to TTL (about 1.4V). What you need to see is CANL go low although not usually to ground. CANH will always show high even when recessive to a TTL trigger level. This might help: e2e.ti.com/blogs_/b/industrial_strength/archive/2015/06/04/…
Peter Smith
Jul 16, 2018 13:06
What are the trigger levels for CANH and CANL? They are recessive at 1.25V (both of them). Do you have a termination resistor on the CAN interface?
 

 Discussion between Peter Smith and Pa…

Imported from a comment discussion on electronics.stackexchang...
Peter Smith
Apr 16, 2016 15:55
ti.com/lit/an/slyt356/slyt356.pdf
Peter Smith
Apr 16, 2016 15:44
There are some really low Iq regualtors around; this one ti.com/product/… has an astounding 500nA Iq. There is a paper that showed regulating ultimately yielded a longer battery life in one question. I will dig it out.
Peter Smith
Apr 16, 2016 15:35
No: the adjustable version (and many low Iq LDO devices) need an external resistor programming string of two resistors; this device does not require them - and yes, even two reistors is something a consumer manufacturer would not want to add.
Peter Smith
Apr 16, 2016 15:30
Note also that the lack of external components (except for decoupling capacitors) also keeps the cost way down.
Peter Smith
Apr 16, 2016 15:30
I would not be surprised if the manufacturer happened to have several thousand of these very cheap (and abundantly available) devices on hand. Cost is a major driver for something like this. If I were doing a battery powered unit for extended operational life, I would not personally use this regulator.
Peter Smith
Apr 16, 2016 15:30
Thanks - You are correct, and I meant to say fixed, and I have corrected that in the answer.
Peter Smith
Apr 16, 2016 15:30
The implication is that there is a load provided internally to the IC package.