Temperature circuit Idea

Well, we all know that the laser diodes (especially the expensives 445nm ones) are suscetible to heat. The intense heat can damage their junction, rendering they unusable, only maybe as a LED.
Well, I found a VERY old article about using a common diode (as 1N4001) as a temperature sensor.
How to a Diode attach to a LD:



obviously the diode would need some thermal compound/grease.

There is the circuit that I have:


translated parts:
"Temperature Controller from 30~60 ºC"
"This circuit (figure 1) release a relay when the temperature passes the value determined by P1. The sensor used is a general use diode..."
ignore the part that insert it on a screw, please
"... The trigger point of the circuit is fixed on P1, based on a reference that may be a common termometer..."
"... The powering of the circuit should be made with stabilized tension of 12V."
"C.I." means Integrated Circuit
"LED" could be the place to insert the wires that disconnect the driver from the batteries (I think).
And I wouldn't be here unless I had some question to the electronics specialists:

- The actual proposal is using the common diode to sense the LD temperature (My drawing) and when the temperature exceeds 60º (the diode safe area is until 70º) it shutoff the LD to prevent damage. Maybe the driver would need modification (TTL? don't really know).
- The circuit works at 12Vcc, want it to run at 3v or maybe 4v, dunno.
- It ranges from 30º celsius for 60º celsius, a bigger range would be needed.
- It uses a relay: pretty old and big devices: would need a conversion to use a SCR (maybe a SCR.. don't know)?
- 741??? big IC detected
- All that maybe in a somehow bigger host (3cm long?) for diode protection?

hope this turn out into a great idea! need help here!

Using common silicon diodes as basic temperature sensors is a very old and proven concept. Their forward voltage drops roughly 2 mV for every degree of temperature, and this property can be used for numerous applications.

Nowadays people may prefer digital temperature sensors or sensor diodes with a bigger mV/K curve, but the basic concept has not changed much over the years.

So yes, you could construct a TEC cooler with a simple silicon diode as its sensor - or even use silicon diodes as the sensor in thermal laser power meters as i've described elsewhere on this forum.

Thats some good DIY stuff there!
I learn something new every day here:beer:

Benm said:

Using common silicon diodes as basic temperature sensors is a very old and proven concept. Their forward voltage drops roughly 2 mV for every degree of temperature, and this property can be used for numerous applications.

Click to expand...

Interesting, I didn't know this. Vf drops 2mV for every degree up? Or down?

Meatball said:

Interesting, I didn't know this. Vf drops 2mV for every degree up? Or down?

Click to expand...

IIRC, that number is only a rough estimate - one can get higher levels of precision if one does the appropriate analysis while propagating errors, and the relationship is not linear, thats why there is only a small range of temperatures it would work for.

Common signal diodes like 1N4148 are decently precise and "repetitive" in this, having usually a -2mV/C derive, and if needed, 2 or 3 of them can be placed in serie, for increase the variation too ..... usually, is used an amplification stage (op-amp) for increase this value .....

Benm: the problem using another sensors is their cost (here in Brazil a NTC or a PTC costs $10). This is a idea to make it "DIY" with homemade items (except the op amp and some more )

People, someone could try to explain my questions, and if it would work? I'm planning to try this. Thanks.
(I have tried it with a burn out LOC (the diode sensing). It worked pretty good for sensing the temperature, and the aixiz module went with 2 wires, case GND, red diode + and yellow 'sensor' +, with a multimeter it works )

edit: Thanks, 532 With Envy. Keep reading here and some people may take interest on this experiment :beer:

but the real question is: is it possible to use at 3v and the current will be low enough to use on a handheld circuit?
if not, we can use it on a Labby style circuit, with 12v. > It could be of use < !

It depend entirely from the driver type that you use (as example, i place a NTC on some of my handheld units that uses linear sink drivers, and i think none have already done this ) ..... that circuit (or any other equivalent), can be used as "safety turn off" for some type of drivers, and as "power limiter" for some others, but not knowing what's the circuit that you want to use as driver, is a bit difficult to give you working suggestions .....

Ofcourse, you must use different values and a different type of op-amp. for work with 3V, cause the 741 serie works stable only from a minimum of 4V (or better said + and - 2V), and also the reference voltage must be changed ..... you can redesign the circuit around a 5230, that work from 2V ..... or also use a 358, that work from 3V, and being a dual op-amp, you can also use the second one in place of the input transistors group, for amplificate the diode signal ..... and as reference voltage, you can use a red LED, that act decently as a 1.5V zener, at low currents ..... also, no need the relay or SCR, for drive the LD driver circuit .....

Lots of possibilities, but still need to know what LD driver you are using .....

I think it would cut the battery contacts to the driver.

LD ---- Driver --Temperature SCR-- Battery

Uhm, no, it's an inefficent way for do this ..... better to blank directly the driver, or limit the current from it .....

As example, if the driver is an LM317 or similar, a transistor (also the output transistor from the circuit) between the ADJ pin and GND, turn it almost off (or, better said, turn it to 1,25V output, that for the most part of the LDs is equivalent to be unpowered) ..... where, in a sink current driver, placing the transistor on the VREF circuit blank it off when it's closed to GND (sort of TTL modulation)

I thinked about TTL
well, if there is no difference than turning the 317 to 1.25V we could put it on 3V output, so the LD get very low powered (assuming it is not a A140) and the user see it and turn it off.

maybe as Benm said, we could use it to control a TEC for a labby unit.

Yes, but remember that with an LM317, you need at least the FV of the LD, plus the total dropout of the driver, as power supply .....

Or do you mean 3V as FV of the LD, and the battery is more than 3V ?

If not, you probably need a sink driver, for use a Li-Ion battery ..... except if you are speaking about a 3V green module (they already use a sink driver) ..... in this case, you can use a mini NTC from ebay (suppose cheaper than from a store, there), for insert a self-regulation loop in the already existing driver .....

BTW, you can use that circuit for control a TEC, if you want, but don't forgot to place a power resistor in parallel to the relais contact ..... this cause, when it's not powered, the TEC re-route the heat from hot to cold side ..... so if you're not using a linear driver for power it, is always better to keep a minimum current through it, for prevent this .....

Limecat said:

IIRC, that number is only a rough estimate - one can get higher levels of precision if one does the appropriate analysis while propagating errors, and the relationship is not linear, thats why there is only a small range of temperatures it would work for.

Click to expand...

The exact mV/K value varies with diodes and depends a bit on the sensing current too. But in practise, it is easily usable. The only downside is that you have to calibrate the circuit, its not like an I2C temperature sensor that gives you a straight digital answer.

Once calibrated it is very reliable though. If you measure a 4148 at 1 mA, you could get values like 700 mV at 20 degrees and 680 mV at 30 degrees. Once you have established these values for a diode, they will not change over time. People build thermometers that are accurate within one degree over a reasonable temperature range with this approach, even someting as wide as 0 to 100 celcius would be possible.

HIMNL9: I don't want that circuit, I want to get this working and post the schematics so everyone can build it or buy it from someone, and prevent their diodes getting destroyed by excessive heating.

Benm: Good, I was looking to read this!

Now we need someone that is available to try to help me with this little circuit..

Feel free to post any circuit ideas you have, i'd be happy to take a look at them!

Benm said:

The exact mV/K value varies with diodes and depends a bit on the sensing current too. But in practise, it is easily usable. The only downside is that you have to calibrate the circuit, its not like an I2C temperature sensor that gives you a straight digital answer.

Once calibrated it is very reliable though. If you measure a 4148 at 1 mA, you could get values like 700 mV at 20 degrees and 680 mV at 30 degrees. Once you have established these values for a diode, they will not change over time. People build thermometers that are accurate within one degree over a reasonable temperature range with this approach, even someting as wide as 0 to 100 celcius would be possible.

Click to expand...

I'm scientific minded and OCD about doing things PERFECT even if its just a one time thing - if it can be calibrated, I will want it calibrated lol
I've 'built' a few PHR lasers and a 6x laser and they all died at some point. Didn't consider them builds since they never went into a housing. I think I have photos of one from about a year or so ago.
I was giong to make a 445 ~1W with an ADC to show current on it. Already have the circuit board sort of laid out with the power supply and current source, but it doesn't look like I'm going to have the space on the PCB to put an ADC on it lol

Anyway, I remember determining Boltzmann's constant by using a transistor, in a physics lab course, but I can't find those lab books. I hope I didn't just leave em. I like to hoard my school work, even if its crap.

PS sorry for tangential discussion.