User KyranF | chat.stackexchange.com

KyranF

Oct 14, 2024 05:28

@rdtsc each capacitor is expected to leak about 85uA according to the datasheet so is definitely not a negligible amount being wasted. I'm going to change up the caps too to reduce overall capacitance and i'll look for better leakage specs if I can

KyranF

Oct 13, 2024 23:31

ah, so I think the small inductor value + 500 Khz + the fact it's a dual MOSFET synchronous converter is what is screwing me. I am going to change it to a non-sync diode-based boost converter, then it doesn't have to pretend to be a diode and only switch when it needs to, for lower current consumption when it's idle 99% of the time.

KyranF

Oct 13, 2024 22:58

Anyone have hints for how to improve the no-load current draw of a Boost converter going from 6v up to 26v? I'm using a TI LM51231-Q1 switching at 500Khz with a 4.7uH inductor, two Mosfets, and fairly high output bulk capacitance ( 4 x 330uf aluminium caps)

I am getting a crazy 60-80mA drain on my board with no output load. The caps charge to 26V and it seems fairly stable, yet there's huge wastage on the input- which is a battery in this case and no good for battery life.

KyranF

May 15, 2024 06:25

@Lundin Thanks for the tips. Yes in this case the product is quite small, 30mm circle by 15mm deep, most of which is taken up by coin cell. Not enough room for a panel mount button - I do agree those are generally better designed for exterior button mounting with weatherproofing

KyranF

May 7, 2024 15:52

Does anyone have any online materials for reading or examples of good ways to integrate buttons and switches in small portable electronic products? I have a small sensor i'm working on, which has a single illuminated tactile button as the only way users interact with it, but I have some uncertainty in how to properly enclose/seal/integrate these switches into the outer plastic enclosure (3D printed for now, likely injection moulded later). I'm wondering about things like gaskets, sealants, epoxy or caulking etc more as an assembly/manufacturing question and switch enclosure integration th…

KyranF

Oct 22, 2022 14:50

@NickAlexeev Thanks for the tips Nick!

KyranF

Oct 15, 2022 23:38

How precise in voltage/drive current must you be? any optical or performance specs you are meeting? That will dictate how you control them and how precise the circuits need to be. If it's not really an important factor for their output precision, you can look at doing a low-side N-channel MOSFE load switch which is driven by your PWM signal

KyranF

Oct 15, 2022 23:36

one thing to note is they will consume more than 5mW each, that's the optical power. usually the actual power driven through them is like 2-5x higher depending how efficient they are.

KyranF

Oct 14, 2022 03:54

I would love to throw money at a professional, industrial design guide book for how to achieve some of these common and 'expected' protections.

Making a robust industrial product that my users can't constantly blow up is really important to me, I've had so many idiot end users it's insane

KyranF

Oct 14, 2022 03:53

anyone have any go-to reference design handbook type things for digital and analog circuits with practical features like over/under voltage protection, ESD protection, current limiting resistors, and other industrial ruggedization methods for common/simple circuits?

For example, i'm doing a high side P-channel MOSFET load switch for 12V to a motor driver input, from a 3.3V microcontroller, I have an NPN BJT to control the 12V MOSFET and I have a small value series resistor for the output load, a reverse blocking diode, a pull-down resistor to ensure it can't float... next best thing would …

KyranF

Apr 25, 2022 04:13

mine was powered from a 24V battery bank from the electric vehicle. Pulse current was 120 amps for 5milliseconds

KyranF

Apr 25, 2022 04:12

Simpler pulse circuits could be used if you were okay with a rapid decay of optical output, with a very high current surge and as capacitors discharge the brightness dims to the end of the pulse

KyranF

Apr 25, 2022 04:11

@Shalvenay my circuit was a constant-current LED driver circuit comprised of 1 main power mosfet, a high power sense resistor, and feedback/limiter BJT (which could be a MOSFET, or an op-amp with gain so you can use a smaller resistance current sense resistor). As the current through the LED going through the low-side FET reaches the setpoint, the limiting circuit kicks in and holds the main MOSFET in a partial on/off linear state, for the duration of the pulse.

KyranF

Apr 25, 2022 03:57

@NickAlexeev haha, that's hilariously the same application and similar power output, probably driven by the same/similar requirements. Funny. I enjoyed making my LED strobe light, I hope you enjoyed yours. My professor was astonished at the optical power output, saying that someone dangled a black USB cable in front of it and after only a short time, the cable become so hot the jacket was melting off

KyranF

Apr 21, 2022 05:19

@NickAlexeev The lights were a smart-sensor, using ethernet interface with a ROS (robot operating system) driver layer to configure the pulse durations, pulse start delay (from camera trigger, PointGrey Grasshopper cameras for stereo vision was used for vision guided navigation), and general brightness settings etc.

It was 24V, with 120 Amps peak pulse power. Duty cycle was only around 5% due to camera frame rate and shutter speeds.

I suppose the answer to the question is it was used for night time navigation by camera

KyranF

Apr 21, 2022 05:17

KyranF

Apr 21, 2022 05:17

@NickAlexeev It was for my Honours thesis, at Queensland University of Technology in Brisbane, Australia, for the Robotics Lab. They were working on a prototype of the AgBot which is an autonomous farming robot project for lighter weight electric farming robots for weed control and other tasks. You can see the strobe light mounted on the front under the canopy

KyranF

Apr 19, 2022 05:42

@rdtsc no, I use octopart to desperately find any stock of any of the chips I use, and it's a desolate landscape across the board (pun intended)

KyranF

Apr 15, 2022 17:25

@rdtsc I tried to get parts from TI directly, a power chip and an audio chip, I emailed their various contact points saying I needed even just 1 chip of each for design verification of a product i'm working on... no good. "sorry sir, out of stock". I thought for sure they'd have a secret reserve held somewhere they could hand out on a case-by-case basis.. but no!

KyranF

Apr 13, 2022 04:58

For a recent project, I had to buy $80 USD Dev boards and de-solder a $10 chip, just to get the prototype out the door for the customer. Crazy!

KyranF

Apr 13, 2022 04:51

audio chips too, and most capacitors and diodes. Are people just shelving their electronics projects/businesses for a year or two while this blows over?

KyranF

Apr 13, 2022 04:47

I am sooo frustrated at this chip shortage... It's just beyond ridiculous at this point. Any IO expander, PMIC, DAC/AIO, MCU I try to spec or continue with from a previous project is all 0 stock and like 10-12 month lead times. /rant

KyranF

Apr 11, 2021 20:25

but what about techniques like sizing a branch off the 5V supply where you know the operating current of an IC is maybe only 20mA DC, having a resistor (like 22 or 33 ohms) and then the local supply capacitors for each IC? normally I just have a flat 5V logic supply to everything, but branching it like that might be useful for diagnosing issues?

KyranF

Apr 11, 2021 20:10

@W5VO I know of the thermal camera option - and as to why the part died, I know the general failure was the user (not me) being very poor with wire conductor hygiene and all sorts of sparks and ESD and silly nonsense was happening to the board. I'm surprised it even lasted as long as it did. This guy has it dangling by wires outside of a control box on the machine (it's an all-electric mobile farm utility tractor thing)

KyranF

Apr 11, 2021 06:56

I have encountered an issue where a logic chip on my board has died, and it dies in a way which shorts the logic supply of that chip to ground, but it's a weak short (or the regulator handles it well) and only pulls it down to 2.5 or 3v (from a 5V logical supply).

I had to go around my board and de-solder every IC on the 5V supply one at a time and power up and test to see which one was actually causing the short. after the 10 chips were each soldered and tested, the final one being the microcontroller in a super annoying 28-DIP package, ended up being the culprit. It wasn't obvious enough…

KyranF

Apr 1, 2021 18:43

@rdtsc wow that's super tiny and high current for its size, and 5 microvolts RMS noise is pretty neat too. it's the size of a fly shit but does better than any LDO i've used yet.

KyranF

Apr 1, 2021 16:30

they make them all in their own "Digi-Key Red" solder mask colour.. haha

KyranF

Apr 1, 2021 16:29

wow cool, DKred

KyranF

Apr 1, 2021 05:49

A bunch of capacitors and ferrite beads and other random stuff I need also went out of stock in the last two weeks... I always make an effort to select stuff which is in stock, but coming to order my BOM 10% of the parts are now out of stock! Crazy. Spent a bunch of time finding alternatives :(

KyranF

Apr 1, 2021 04:22

@Shalvenay strangely most of the other distributors don't have this particular chip, maybe it's so new or mouser/digikey have some exclusivity deal with TI? I don't know, but it's frustrating.

KyranF

Apr 1, 2021 04:15

@W5VO I am in the process of requesting samples, not sure on the result yet. Hopefully they have some kept in reserve for samples/eval boards.

KyranF

Apr 1, 2021 03:27

anyone happen to know where I can get some (5), or have some I can buy?

KyranF

Apr 1, 2021 03:25

so, has anyone run into chip shortage issues lately? I'm trying to source the texas instruments LM76005 and Digikey and Mouser are all out of stock until september, with many thousands on backorder..

KyranF

Feb 15, 2018 00:54

i love component shopping, it's a nice aspect of design

KyranF

Feb 15, 2018 00:54

The Infineon BTS5200 is a similar chip i just found, looks nice, quad output and great specs

KyranF

Feb 15, 2018 00:49

that's perfect

KyranF

Feb 15, 2018 00:32

thanks!

KyranF

Feb 15, 2018 00:24

simple high side driver x 3 in two ICs. I did an 8-channel circuit like this before with discrete (although, they were at least dual PMOS in one SOIC-8 package) FETs and resistors everywhere. it was ugly. could have been way more integrated

KyranF

Feb 15, 2018 00:24

I only need 3 outputs on this, I might just get a quad PNP IC and whack that after the 7407, where the open collector pulls down the PNP base to turn it on

KyranF

Feb 15, 2018 00:22

are there PNP versions of those darlington arrays, typically?

KyranF

Feb 14, 2018 23:57

I should have just used a PNP darlington array IC instead of this open collector one

KyranF

Feb 14, 2018 23:56

thanks guys! I know my current circuit is working, but i don't like being uncertain about it

KyranF

Feb 14, 2018 23:55

it's more like 0.8V with the current through it at the moment

KyranF

Feb 14, 2018 23:55

yeah. i get just over 0.7V on it when its low

KyranF

Feb 14, 2018 23:40

so many unanswered questions from a TI datasheet! This is unusual in my (small) experience

KyranF

Feb 14, 2018 23:39

the IC can handle 40mA per output it says, but is that constant? Datasheet says typical/recommended operating conditions is 40ma per output channel. Says typical power dissipation is 140mW, is that with all 6 channels sinking 40ma?

KyranF

Feb 14, 2018 23:37

the same circuit should work on the PLC's input optocouplers but i haven't tested that yet. i'm sure it's similar, both expect 5-10ma to operate

KyranF

Feb 14, 2018 23:36

it's just enough to make it work, who knows if it's reliable or efficient (i don't think it is :( )

KyranF

Feb 14, 2018 23:35

the opto coupler has a series 2.2K resistor, which makes the 'logic high' actually only source around 5ma

KyranF

Feb 14, 2018 23:35

I ended up with a 2k pull up resistor to source the current from 24V rail, and i have a schottky diode on the output (instead of series resistor), and the open collector pulls low to re-direct that ~12mA current to ground through the IC

KyranF

Electrical Engineering