Add a non-ohmic "real bulb" for DC sim

[kathy-phet]

As we consider redesign, we should consider the possibility of modeling a real bulb as an option, where the resistance changes with the current. As an advanced feature.

[ariel-phet]

@oliver-phet forwarded me a request for this particular enhancement today:

Dear folks

I am impressed with your simulations - ease of use and accuracy but was disappointed with one aspect of the dc lab - the resistance of the filament lamp stayed perfectly constant! Is there anyway you can change this to have the resistance increasing with current??

Many thanks

I believe we have received this request more than once, and it would be a nice addition. The real bulb should definitely be an "option" and not the default, but it seems worth having this as canonical example of a somewhat non-ohmic component.

[samreid]

I'm unfamiliar with the physics of non-ohmic light bulbs, can you point me to a model for this? Also a note for when I get to implementing this: adding equations to the matrix solve step seems a much more robust solution than iterating in search of a stable point.

[ariel-phet]

@samreid well...the resistance basically increases with temperature, I will see if I can find a decent model/reference, but it might be something we "hollywood"

This seems to have some good information:

ohmic-lightbulb(2).pdf

One of the the odd things to keep in mind with a "real" lightbulb will be that as it gets hot, the resistance goes up, so for a fixed voltage, the power dissipated will go down, yet it will still appear "brighter" since the brightness we see is related to the blackbody spectrum output.

I am guessing @phetsims/dubson will have some good insight as how to hollywood this well.

[ariel-phet]

@samreid I am also going to leave this unassigned for the moment, it is low priority

[samreid]

Another request for non-ohmic:

[samreid]

Based on discussion Thursday, I've placed this issue on hold.

[samreid]

Kathy says we may want to model current as a function of voltage (like a sigmoid), at steady state.

[samreid]

From design meeting discussion:

Non-Ohmic bulb would only appear on advanced screen, if implemented.

[samreid]

From July 20156 meeting:

Big feature request from teachers that would be on lab screen, if included.

As temperature increases, resistance increases

Open questions: does it need to have some memory?

The bulb needs to know how long current has been going through it. Not currently in the model, but SR thinks it can be added. Note: this will not be necessary if we are modeling the R=f(V) as a steady state.

What are the typical learning goals? Looks to be steady state. Teacher request: Is there anyway you can change this to have the resistance increasing with current??

Non-Ohmic Bulb would not have tunable resistance. The resistance would be derived from V and I.

Add a LED, modeled as a diode. Goals for LED: doesn’t turn on until certain voltage threshold (2V?) and won’t turn on if leads are connected in reverse.

LEDs are generally in current-limited circuits. Will be very easy to burn it out without resistance.

[samreid]

Normally, the current resistance curves are not a function of time.

[arouinfar]

I found another request for a non-Ohmic bulb on Unfuddle.

Unfuddle Ticket #3212

"When the battery voltage is increasing , the resistor is following the Ohm's Law (good thing), but the light to.....(bad thing)
May be is it to difficult to make the resistor value from the light bulb increase. I don't know much about informatic programs.
If you can change that , it would be very good.

[samreid]

@arouinfar is this for 1.0 or 1.1?

[arouinfar]

I would say this is a 1.1 feature @samreid.

[samreid]

I committed a proof-of-concept "rough draft" working prototype for non-ohmic light bulbs. Notes on how it was implemented:

The desired curve looks roughly logarithmic, so I modeled I = k*ln(V). In testing, I determined that base=2 with a coefficient k=3 worked well. Combining I=ln(V) with V=IR and solving for R gives R = V/ln(V). I also shift the logarithm so that at V=0 the log is 1.

        const base = 2;

        // I = ln(V)
        // V=IR
        // V=ln(V)R
        // R = V/ln(V)

        // Adjust so it looks good in comparison to a standard bulb
        const coefficient = 3;

        // shift by base so at V=0 the log is 1
        resistorAdapter.value = coefficient * V / logWithBase( V + base, base );

I used the same facility we developed for batteries exceeding their current threshold in #245 (running the computation engine twice). The first run puts all non-ohmic bulbs at 1 ohm to compute their voltage and therefore resistance. Then the second run uses the computed resistance.

I added the non-ohmic bulb to the last page of the carousel in the "Lab" screen, only in the DC simulation. The non-ohmic bulb does not have a control panel.

Some questions for the design team:

• Please test and comment on the basic functionality.
• Where should it appear in the carousel?
• Which screens should it appear on?
• How should it be labeled?
• How should it look different from a standard bulb?
• I picked a coefficient that shows a reasonable dynamic range compared to a standard bulb in parallel, but a standard bulb and non-ohmic bulb have a slightly different brightness at the default 9V. Should we tune it so they have the same brightness with a 9V battery?
• Any other adjustments to the log base, coefficient or equation?

[Nibbos]

Hi, Sam. This is looking really good. Many thanks. I'm just wondering if it's possible to tweak the function you use to give a less vertical plot around the V=0 region? Are you able to display the V and I values so that students can calculate R as the experiment is carried out? If so, what value of R is produced at small values of V? And are you able to produce the same graph for negative values of V and I? Really appreciate the work going into this. Regards Conor

…

On Tue, 20 Oct 2020, 04:51 Sam Reid, ***@***.***> wrote: I committed a proof-of-concept "rough draft" working prototype for non-ohmic light bulbs. Notes on how it was implemented: The desired curve looks roughly logarithmic, so I modeled I = k*ln(V). In testing, I determined that base=2 with a coefficient k=3 worked well. Combining I=ln(V) with V=IR and solving for R gives R = V/ln(V). I also shift the logarithm so that at V=0 the log is 1. const base = 2; // I = ln(V) // V=IR // V=ln(V)R // R = V/ln(V) // Adjust so it looks good in comparison to a standard bulb const coefficient = 3; // shift by base so at V=0 the log is 1 resistorAdapter.value = coefficient * V / logWithBase( V + base, base ); I used the same facility we developed for batteries exceeding their current threshold in #245 <#245> (running the computation engine twice). The first run puts all non-ohmic bulbs at 1 ohm to compute their voltage and therefore resistance. Then the second run uses the computed resistance. I added the non-ohmic bulb to the last page of the carousel in the "Lab" screen, only in the DC simulation. The non-ohmic bulb does not have a control panel. Some questions for the design team: - Please test and comment on the basic functionality. - Where should it appear in the carousel? - Which screens should it appear on? - How should it be labeled? - How should it look different from a standard bulb? - I picked a coefficient that shows a reasonable dynamic range compared to a standard bulb in parallel, but a standard bulb and non-ohmic bulb have a slightly different brightness at the default 9V. Should we tune it so they have the same brightness with a 9V battery? - Any other adjustments to the log base, coefficient or equation? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEEEH3A2N7NX66WFLF5TQ33SLUCLXANCNFSM4B3Y5VPA> .

[ChemersSci]

This would be amazing if implemented and would make the simulation compatible with the GCSE (high school) core practicals in the UK as already mentioned. Myself and the students have found the sims incredibly useful during COVID 19 where practical work isn't as accessible and some students are self isolating. The qualifications regulator has said that we can use simluations instead of completing practical work this year, which has made PhET and invaluable resource.

The circuits and graphs that students need to be able to produce can be found here: https://www.bbc.co.uk/bitesize/guides/zgbwpbk/revision/3 if that helps at all. The inclusion of a diode and the ability to record negative values for V and I would mean students in the UK could fully complete the practical from the simulation, which would be amazing!

Thanks for all the excellent work so far.

[Nibbos]

Negative current won't be implemented, alas. This had been requested but won't be happening. I told students to just make the current values negative when they reversed the polarity of the battery, since they could see that the current flowing was in the opposite direction. Filament bulb is looking good, and I've suggested a few tweaks to make it a less vertical graph in and around the origin. Diode, thermistor and LDR would be awesome! Conor

…

On Tue, 27 Oct 2020, 12:12 ChemersSci, ***@***.***> wrote: This would be amazing if implemented and would make the simulation compatible with the GCSE (high school) core practicals in the UK as already mentioned. Myself and the students have found the sims incredibly useful during COVID 19 where practical work isn't as accessible and some students are self isolating. The qualifications regulator has said that we can use simluations instead of completing practical work this year, which has made PhET and invaluable resource. The circuits and graphs that students need to be able to produce can be found here: https://www.bbc.co.uk/bitesize/guides/zgbwpbk/revision/3 if that helps at all. The inclusion of a diode and the ability to record negative values for V and I would mean students in the UK could fully complete the practical from the simulation, which would be amazing! Thanks for all the excellent work so far. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEEEH3CORCYRQWFBADK22STSM22LJANCNFSM4B3Y5VPA> .

[Geoffysicist]

Thanks Sam for finally picking this up. looks good so far. Is there some way we can play with the prototype?

[samreid]

We have not published a prototype version with these changes, but it should be possible to check out the sources, as described in https://github.com/phetsims/circuit-construction-kit-dc/blob/master/README.md

[samreid]

I committed the name change and i18n and opened side issues for the remaining work. @arouinfar can this issue be closed?

UPDATE: Or leave it open and on hold as a reminder to reply to all of the users that have requested this feature?

[samreid]

After some revision, the new behavior looks like this:

Here is I vs V:

Here is R vs I

@ProfStick @Fedsie @Nibbos @ChemersSci @Geoffysicist does this seem suitable for your use cases?

[Nibbos]

I-V curve looks fantastic. Thanks, Sam. R-I curve interesting but not required for GCSE teaching purposes, but increasing R with I is part of the qualitative answer required by our exam specifications. Many thanks thus far. Regards Conor PS Will negative voltage give same values for current when polarity of battery is reversed?

…

On Fri, 30 Oct 2020, 17:04 Sam Reid, ***@***.***> wrote: After some revision, the new behavior looks like this: Here is I vs V: [image: image] <https://user-images.githubusercontent.com/679486/97729986-e1c3f880-1a98-11eb-8aed-e1626eeb0bcb.png> Here is R vs I [image: image] <https://user-images.githubusercontent.com/679486/97730029-f0121480-1a98-11eb-8a35-3c520aad5400.png> @ProfStick <https://github.com/ProfStick> @Fedsie <https://github.com/Fedsie> @Nibbos <https://github.com/Nibbos> @ChemersSci <https://github.com/ChemersSci> @Geoffysicist <https://github.com/Geoffysicist> does this seem suitable for your use cases? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEEEH3CHHYLM4SQ3AWDS27LSNLW3DANCNFSM4B3Y5VPA> .

[samreid]

PS Will negative voltage give same values for current when polarity of
battery is reversed?

Yes, the computation uses the absolute value of the voltage difference:

circuit-construction-kit-common/js/model/ModifiedNodalAnalysisAdapter.js

Line 210 in aa6fe2c

const V = Math.abs( v1 - v0 );

[Nibbos]

Perfect. Thanks.

…

On Fri, 30 Oct 2020, 17:23 Sam Reid, ***@***.***> wrote: PS Will negative voltage give same values for current when polarity of battery is reversed? Yes, the computation uses the absolute value of the voltage difference: https://github.com/phetsims/circuit-construction-kit-common/blob/aa6fe2c38306c6e16840df59a859f67def7b55e7/js/model/ModifiedNodalAnalysisAdapter.js#L210 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEEEH3AVWE6WM4E7IL3PD5TSNLZCPANCNFSM4B3Y5VPA> .

[samreid]

As far as I can tell, all of the work for this issue is completed (aside from publishing, which will be done in #616 ), closing.