CircuitLab simulation

[Ioannis]

Have not tried the KiCAD yet (as well) but LTSpice, despite its primitive environment, it is easy use, at least for the basics.

Worth to give it a try.

Ioannis

[Demon]

Just a clarification, KiCad uses ngspice.

Turns out KiCad has a few "features" that you quickly pick up on when you read about spice:

- as mentionned previously, you're expected to import models.
- pin assignments in KiCad are not necessarily the same as the pins in the ngspice engine. You're expected to double-check, and use alternate pin assignments when necessary.

"KiCad uses different pin assignments for semiconductor components than is expected by the ngspice engine. For example, the standard SPICE pin assignments for a BJT transistor are pin 1 = collector, pin 2 = base, and pin 3 = emitter, whereas KiCad uses 1 = E, 2 = B, and 3 = C. This can cause great confusion for users wondering what the heck is going on. To fix this issue, click on the Pin Assignments tab and swap the collector and emitter pin assignments"

https://www.woolseyworkshop.com/2019...tion-in-kicad/


I read somewhere that this might just be leftover "holes" from early versions of KiCad. It's developed by the community, so it's bound to have some gray areas where people weren't sure how things went. So over the years they kept the ability to override stuff. It's not the end of the world.

Shoot, just look at our community and imagine if we were responsible for the development of PBP.

[Demon]

Have not tried the KiCAD yet (as well) but LTSpice, despite its primitive environment, it is easy use, at least for the basics.

Worth to give it a try.

Ioannis

I gotta say, I was able to get a simulation running with just a few clicks.

[Ioannis]

Isn't it amazing?

Ioannis

[Demon]

Yeah ... too good to be true.


Problems quickly arise when you want to test using parts off amazon or aliexpress. They rarely have manufacturer names, so getting matching spice models is a challenge. These guys build their own models.

https://www.electro-tech-online.com/...ob-led.157114/


I wanted to test this COB LED brightness control, but substituting the SC8050 for a 2N2222 (or possibly a logic-level MOSFET if the transistor still heated a lot at the 3/4 mark):




To control the brightness of a 15 inch section of this COB red LED strip:

https://www.aliexpress.com/item/1005003307581709.html


I guess I'm back to breadboards.


EDIT: This guy has a similar design using a 2N3904:

[Ioannis]

Simulating any part is impossible.

Either you should choose a similar one (search the included lib or other people's lib) or make your own library. There are a few ways but I never tried really do my own. Though there are directions on many sites and forums.

Ioannis

[Jerson]

If you can accomodate some simulation inaccuracies, I suggest you try this circuit simulator. https://falstad.com/circuit/ Extremely frugal on resources.

This is also downloadable as an offline circuit simulator.exe on the same site. To my knowledge, this is not accurate like LTspice or others, but gives a nice visual representation of the working circuit.

I suggest this as an option with a very low learning curve.

[Demon]

If you can accomodate some simulation inaccuracies, I suggest you try this circuit simulator. ...

Thanks, I'll keep that one in mind if I ever try to simulate a circuit again.


I had tried using that PWM board (FR120N, LR7843, D4184), but it the COB LED wouldn't completely turn off and I was stuck with flicker. I had tried ILD2 and LTV-825 opto-couplers, but I was serious trouble balancing the circuit (I can't remember the exact problems, but I couldn't get it to work as desired - full OFF/ON, 300mA range for COB LED, etc).

I ended up with a IRFZ44N MOSFET driven by PWM from a PIC, with 1K on the gate; looks bullet-proof, low cost, minimal parts and super simple. Best part, it is driven from a PIC, so it's a snap reproducing the same lighting on the 2nd enclosure. These FETs can take insane voltages/amps (I'm only working with logic circuits).

Small victory.