DC Voltage conversion for miniature bulbs?

[David New Jersey]

Hey folks, wondering if anyone can point me either to the theory or a solution for an issue that comes up more and more for me with my older electronics - I have a DC component that goes bad in a circuit, lets take today's issue, a miniature DC filament bulb used for illuminating things from which the reflected light gets measured. I need to replace the bulb, but it's specs are no longer available in the same voltage, physical dimensions and current draw. It needs to be a filament bulb for spectra considerations so the vast range of drop in LED's aren't an option given their spectral incompleteness. The only other filament (more specifically incandescent) bulbs that will physically fit in place with the same wattage output are of different voltage and current ratings.

In my current case, according to my meter, my device is driving the bulb at 3.0VDC and it's drawing 1.0A. But modern replacement miniature bulbs only have that kind of wattage in higher voltage bulbs, like 6VDC or 12VDC.

If the bulb were always on or off, I'd simply power it with a supply independent of the circuit, but the circuit turns it on and off for testing and calibration purposes, checking it with a light sensor during these on off intervals, so it would be nice to have the replacement bulb be powered by the circuit so the light output is in sync with the photo diode.

What comes to mind for a tinkerer like myself are (say in the case of a 12VDC replacement bulb);
1) Use a sub 1A, 3V relay and power the other bulb with a separate 12V supply,
2) Somehow use a voltage regulator (suggestions?)
3) Go crazy and program an arduino, which seems like overkill.

Any help, in advance, is greatly appreciated - thanks! David

 

[duke37]

I would experiment with a boost convertor to change the 3V DC to 6 or 12V.DC.
Where does the 3V DC come from?
Do you need a thick filament for optical reasons?
How critical is the voltage?

 

[Alec_t]

How important is colour temperature? I'm guessing (could be wrong) that using a 6/12V bulb instead of a 3V one, even if of the same wattage, may not result in the same colour temperature, because of the difference in the thermal mass of the filaments.

 

[David New Jersey]

Great questions, thanks for getting back. To start, the device is a 90's era color densitometer, measuring reflectance percentage of light off of flat objects (photographs). There are 12 sensors that are filtered to different regions of the visible spectrum. The unit self calibrates to the bias of the light source - as long as the light source is spectrally 'complete' (isn't spikey like LED's or other some other emitters), it will self adjust for color temperature shifts and biases in the bulb, which can occur from changes in voltage as you probably know already. Also, incandescent bulb's have highly asymmetric spectral output to begin with but it's correctable in calibration.
The 3V comes at least off of a TTL logic board connected to the read head where the photo diodes sensors are and light bulb I'm working to replace. The densitometer has it's own internal 12v power supply I believe and I'd guess a regulator near there (there is no regulator I can see on the little head control board which sends the 3V to the bulb).
The actual bulb is very unique in being a full 3 watts for such a tiny package size (about 2mm diameter x 5mm long), and it has a lens in it to focus light. I will work on reverse engineering that and installing a permanent lens in the optical path, but for now, I'm stuck with how to deal with this 3V power issue and replacement bulbs.

As for how critical the voltage is, well, that's the thing, the only other bulbs I can find which will give me 3 watts of power are higher voltage, like 6 or 12, so that's my question I'm putting out there, I'm not even sure what question exactly to ask, so I guess I'd say, what would you folks try?

I'm looking into boost converters right now, thanks so far for that suggestion, I've got 6 machines in parts so I'm not terribly concerned about experimenting.
David

 

[Alec_t]

The densitometer has it's own internal 12v power supply I believe

If that has enough oomph to power a 3W 12V bulb in place of the 3V one you should be home and dry. A light pipe might be handy if you need to position the (larger) bulb other than in the original bulb's position.

 

[David New Jersey]

Well, now I'm working off of the boost converter idea, but I should have noted before (here we go...) that the bulb has a standby state where it's dimmed at 2.5V, but that increases to 3V when the unit is called to make a light reading. So far however, the list of boost converters I'm finding on Digikey look like they start at 3V for Vmin input specs.

Alec_t I'm guessing by oomph you're meaning after a boost converter draws it's required current in addition to a higher voltage bulb's current requirements?

The bulb needs to fit in a hole at the top of the little head which illuminates the photograh, and the higher voltage bulbs average 12mm vs the 2mm diameter of the existing bulb, so I thought about a prism or mirror to make a 90 deg. turn into the hole which these larger bulbs can't fit into, I think I'll see if Thorlabs has anything that's small enough to work "in-head". Thanks for the light pipe suggestion - David

 

[David New Jersey]

Here's a pic for curiosities sake of what I'm playing with. The bulb on there right now is another mini bulb I tried but it's too low wattage. Got it down the street from the radio shack that's closing down. I know they've been closing down for over a decade now, but it feels like a change in an era to me.

 

[Alec_t]

I suppose one of these (600mW) isn't powerful enough?
If the densitometer already has a 12V supply, why do you need to step up from 3V?

 

[David New Jersey]

The bulb supply voltage is modulated by the control board both during readings and automatically during startup for lamp testing and calibration, with a shift in bulb voltage supplied changing from 2.5-3VDC between 2 states, standby and "in use". If the bulb doesn't change states, the startup and calibration fails, and the machine won't work properly because it didn't offset the op amps in the signal amplifier circuit, pre-ADC during calibration...so I'm kind of tethered to that 3V line I think...

 

[BobK]

Hmm. It seems like you might need a DC amplifier that can handle 3W of power. An opamp with gain of 4 plus a P-channel MOSFET output transistor might do it. The feedback for the amp would be from the junction of the transistor's drain and the bulb.

Bob

 

[David New Jersey]

Hello Bob, thanks for that suggestion, I'll look into that, at this point all I've been able to find through mouser/digikey/allied that will handle ~2.7 vmin input and output 12V is a part that won't be available until October
http://www.mouser.com/ProductDetail/Monolithic-Power-Systems-MPS/mEZD41502A-B/?qs=/ha2pyFaduggnV8zhjNStyhG7IFRk02MP8MM%2bYotI1J5kRdPnpUNRw==

So, recommendation well appreciated! David

 

[BobK]

I gave you the complete solution. You do not need the part you are referring to.

Bob

 

[David New Jersey]

Ok, thanks Bob, that's great news - oh, it just occurred to me, in the photo I attached above, one can see that the case is a hefty aluminum to stabilize temperature. Relatively small drifts (I don't know the spec unfortunately) in DC supply to the bulb can cause errors in readings (this device I'm working with is effectively a precision light meter). Given this, would you suggest researching and using additional measures to stabilize the op amp with respect to temp?
Oh, I noticed you're from Worcester, I was there not too long ago, if you haven't already and like Vietnamese, try Dalat Restaurant on Park, amazing family cooking with rural spicing, better than Boston I'd say!

 

[BobK]

Dalat Restaurant

Been there many times!

What brought you to Worcester?

Oh, and the op amp does not need temperature compensation, it works on negative feedback and is inherently stable. But, on the other hand, If you think temperature drift would be a problem, then replacing the bulb with something it was not designed for would swamp any contribution from minor temperature differences causing a different current in the bulb. The power vs color temperature characteristics of your bulb will surely not match the original.

Bob

 

[David New Jersey]

Was in Worcester for a client meeting. Color temp shifts by the bulb or new bulb, are compensated for by the 12, dye filtered sensors spanning the VIS spectrum during calibration, prior to performing light readings. If the op amp on negative feedback is stable, than great, looks like I'm about ready to go, thanks for all the advice again you may have saved a lot of headache trying to solve this issue with the gaps in my knowledge of electronics. Best - David

 

[BobK]

Here is a circuit that implements my suggestion.

V2 represents the voltage your instrument is sending to the old bulb. In my simulation, it varies from 2V to 3V. R3 represents your new 12V bulb. The chart shows how the voltage to that bulb varies from 8V to 12V when the input is 2V to 3V. Pretty much any P-channel MOSFET will work. Choose one with a reasonably low on resistance, (<100mΩ) and a fairly high threshold voltage (actually low since it is negative), about -3V or more because the op amp cannot output all the way up to positive rail and it must be able get up to the the theshold voltage to turn the MOSFET off.

If you are familiar with opamp circuits, you will see that this is a non-inverting amplifier, gain of 4, but the + and - inputs are reversed! This is not a mistake, it is because the output stage is an inverter, so the + input become the - input and the - becomes the + of the full amplifier.

 

[David New Jersey]

You're right, I would have been confused! I think I have an LM324 around but will order the mosfet. Bob, thanks for putting this in spice for me to visualize, that made things a good bit easier to understand, especially the reversal on the input. Say, I wonder if you knew a Ralph Jackson who used to teach electronics in the Worcester area by any chance, he's retired now, but I knew him some time ago, I think he did satellite comms, but now teaches and rebuilds antique radios?

 

[BobK]

Nope. I am not in touch with anyone else doing electronics in Worcester. I do know a lot of ballroom dancers though

Bob