Hyperion LPM + Vritonuzz Data Logger Platform

[Pman]

Watching what happens with this. Decent price point although I already have a couple.

 

[Astralist]

That's great as my Scientech has an analog output as well. It doesn't change, however, when you change the scale you are using. It is locked into a 3 volt full scale output for every scale.

I've been looking for information about your scientech, at first thought the BNC out analog port was for something other than that because it simply strange for the meter with analog sensor input, to have an output as an analog signal :thinking:
Now i know for sure that the analog output from this meter is already compensated/calibrated/amplified as it outputs 3V per full scale.

Ran out of space on the MCU? Guess there usually is a model with more memory etc with the same pinout, albeit more expensive.

In any case i'm mostly curious about the sensor assembly. As long as that's solid there always is the option to combine it with better adcs, more powerful processors and what not in the future.

At first, when i just create this as a mere LPM (that is just displaying the mW after being converted with sensor coefficient/calibration) it leaves a huge room for creativity. As my hand feels itchy to coded more feature, then i added more one by one until i realized that it run out of space. :crackup::crackup::crackup:

With better ADCs you mean either 24-bit or 32-bit ADC?

Watching what happens with this. Decent price point although I already have a couple.

Me too, is watching for what happens with all of my suppliers, altough i already secured the raw material for 50 units. :crackup::undecided:

 

[Benm]

The required adc depends on how non-linear the relation between laser light and sensor output really is. If it is linear and you can match things well you could measure up to 65 watts in 1 mW steps using only 16 bits.

I guess 18 bit adc's are fairly common though, giving you something like 26 watts in 0.1 mW steps if linear.

I doubt this would be a limiting factor in any case though: you only need a single shot adc for this purpose, not a realtime one as you would need for audio applications and such, although those are available too in 24 bit for audio applications.

As far as code goes i reckon much of it went into trend prediction instead of raw measurement to get reasonably fast measurements instead of waiting for temperatures to stabilize. There are some tricks around this that can be done in hardware as well, although those require 2 sensors to work best. If you're interested i'd be happy to have a go at that once the sensors/tec's are in with coating on them and all that.

 

[Trevor]

Been meaning to leave a comment... kept forgetting! Great to see someone else working on LPM stuff.

Speaking to datalogging protocol, a long time ago I was working on putting together the OpenLPM protocol. Only two data packet types were ever used.

d  [READING_IN_MILLIWATTS]\n
sp [THRESHOLD_IN_MILLIWATTS] [TIME_IN_MS]\n

Note that both packet types are followed by a newline (\n) but no carriage return (\r).

The "d" packet is just data - so a "d", a space, then the reading in milliwatts, followed by a newline.

The "sp" packet is kind of fun. What it does is reset your graphing software when you press a reset button on the LPM. The LPM theoretically has a testing time and threshold power to begin, so those are both transmitted to your PC-side logging software so that it's ready to start receiving data when you point a laser at your sensor.

I can give you some code snippets to use, if you want. I believe my currently released version of Peregrine takes advantage of the "sp" packet, if not I'll upload the new source code.

You mentioned your LaserBee... if you want to use that graphing software, you need to transmit data in this format:

\r[VALUE],[VALUE]\n

So, for instance, a value of 10mW would be \r10,10\n.

Will be exciting to watch as you develop this!

Trevor

 

[Trendkilla254]

I am very interested in this LPM. I need to be able to check low power builds and this seems perfect! I can't wait until you get these done. My birthday is in July...

 

[Benm]

@Trevor: what would be the point of sending the same value twice?

I'm especially interested in the sensors used here... I have some ideas on a somewhat different approach but that still requires a TEC with a good coating that is very black and can handle a decent amount of power. Due to lack of that i've put it on hold, but if such sensors become available i'd love to do some work on it.

 

[Trevor]

@Trevor: what would be the point of sending the same value twice?

Dunno. If you read the protocol sent by most LaserBees on a serial monitor, that's what it is.

I think the original plan was for one to be power and the other current, but that hasn't happened... even some eight years after the protocol was imagined.

It is what it is, I guess!

Trevor

 

[Astralist]

Been meaning to leave a comment... kept forgetting! Great to see someone else working on LPM stuff.

The "sp" packet is kind of fun. What it does is reset your graphing software when you press a reset button on the LPM. The LPM theoretically has a testing time and threshold power to begin, so those are both transmitted to your PC-side logging software so that it's ready to start receiving data when you point a laser at your sensor.

Will be exciting to watch as you develop this!

Thanks! :beer:

The protocol is quite easy i think, but i still don't quite understand the sp packet. I can think of the threshold as a mW threshold for auto-logging on PC software just like mine. But how about the testing time? i've no clue at all or does it translates to a delta-time for each reading??

As i'm using USB protocol, i need to add some RS232 COM protocol to my datalogger to make it work with all RS232 based LPM (or USB CDC emulated of RS232). I just hope that it can become an alternative datalogger for windows OS without the need for separate software for USB protocol and COM protocol. :beer:

I am very interested in this LPM. I need to be able to check low power builds and this seems perfect! I can't wait until you get these done. My birthday is in July...

Actually it's done, but i just need to recheck and double check again, just in case i can add more improvement to this meter. Especially the sensor and coating, right now i just have the 8x8mm and 10x10mm sensor that has nearly same performance. And maybe i can give you a discount as for your birthday and newborn baby :beer:

I'm especially interested in the sensors used here... I have some ideas on a somewhat different approach but that still requires a TEC with a good coating that is very black and can handle a decent amount of power. Due to lack of that i've put it on hold, but if such sensors become available i'd love to do some work on it.

I've been testing many combination with coating, unfortunately not a single of them is focus/pinpoint resistant even with carbon deposition, the surface will degrade if using laser that pinpointed (focused for the size about 1mm circle) to it. Then this just like the other DIY LPM, i need to inform the user to avoid using focused/slightly focused laser to measure the power.

The alternative for this it to use silicone carbide deposition, but that is way too expensive. There is much more consideration actually, like the coating type vs flatness of absoptivity. Some black coating just act as a mirror for some spesific wavelength, for example black Al anodize vs IR, the IR will reflected just like when you point it on a mirror! :undecided:

BTW you can see the attached pic as a blackness comparison of my LaserBee II OEM sensor VS matte black paint coated Rp500 Coin VS HPCVD micro carbon coated Rp200 coin.

 

[Benm]

That looks pretty black to me - and i guess those coins are ideal test surfaces given their low value and aluminium composition.

I would not worry too much about being resistant to laser focussed to a tiny dot. As long as it can handle a couple of watts in a beam a few square mm in surface it will be perfectly usable. Afaik even those ophir heads have limitations when it comes to power density and can be damaged by lasers under their rated power focussed down.

The problem is the coating always has limited thermal conducitivity but also some thickness - i'll bet you can burn it off that rp 200 coin without getting the coin even warm to the touch.

Is the coating mechanically durable in any way - say strong enough not to be damaged when you rub your fingers over it?

 

[Astralist]

Actually the blackest coin cannot be burnt at all, because it's carbon coated, not like the ones with paint coating.

But the problem is that since the carbon coating has a microscopic suede surface, it doesn't mechanically durable at all. The sensor will be damaged with just a slight rub. On the other hand, using only matte/flat black paint produce a coating that is reflective and doesn't looks too black, but has a relatively durable surface.
Another consideration is that while you may want to use high temperature resistant paint, you might get a coating that really "heat resistant". I can think that heat resistant paint will have lower conductivity than normal paint, but of course it also depends on its thickness.

To keep the balance, i consider making two layers of coating, so with the base of high temperature reistant followed with micro carbon suede surface, it will have a good balance.
I have a reference (on google) that there was exist a suede black paint, but too bad that AFAIK it was stopped in production

BTW my 6x6mm sensor will arrive on 27 Apr.
For your algorithm testing, i guess i can offer you the 8x8mm sensor for free, provided that you cover shipping cost. Since it has transient and slow response, it will be a good choice for testing the extrapolate algorithm.
But IMO i'm still in doubt about using curve extrapolate, it's only good if the power measured was relatively constant (and without negative slope). If you measure a DPSS system, then i guess there will be a problem.

 

[Benm]

My idea would actually need two identical sensors, though only one really needs to have the coating on it.

What i basically plan on doing is mounting them on the same heatsink, one heated by the laser, the other heated by a SMD resistor. Using circuitry to keep their temperatures identical it's possible to determine how much electrical heating is needed for the reference sensor to match whatever amount the laser deposits on the sensing one.

If you can get them in indonesia i'm sure we can work shipping costs out: either send them directly or if it's somewhere during this summer i can have someone pick them up (maybe from jakarta, otherwise from a friend on bali so you only have to ship domestically).

Also i don't mind paying manufacture cost for them obviously - and i can probably find a matching heatsink around here so that doesnt have to be shipped either if it's a problem.

 

[paul1598419]

Are you looking for a heat sink like the ones used on the Laserbee, radiant X4? A TO-220 heat sink? These are quite cheap, I have a half dozen myself. I have some black matte paint I was going to use again that is thermally protective to 2000 degrees F. I'm not sure about the range of it as being completely linear through all wavelengths and power settings.

 

[Astralist]

@Benm

I guess i can see through the logic of your design, but as per your sensor configuration i'm not sure if i can do that, does it mounted side by side or stacked?
Also i'm quite confused about which setting of sensor you need/want to test. If you only need the sensor that has been coated and not mounted to heatsink yet, i can provide it that way, but please note that when you mount it on heatsink you might accidentally deteriorate the coating, and if i just mount it the way i mounted it on the heatsink, it may not as you expected it to be.

It will be good if i can just ship it domestically, and btw my 6mm sensor has already arrived but it has to be tested yet. I'm guessing that the response time is below 3s for 0-95%.

FYI my thermopile stock:

• Assorted 8x8x3mm and 10x10x3mm 60+pcs (supplier missing)
• 8x8x2.55mm 1pcs
• 6x6x2.2mm 100pcs

Like i said before, if you want to test your idea, the best practice is to use the assorted ones because it has a quite long transient time. And that ones, i can give to you freely.

Feel free to PM me should you want to not to spill your design here.

 

[Rivem]

Definitely interested in this astralist. Let us know when you sell. If you want to send out beta units, I can tweak electronics and software.

 

[Astralist]

Of course i will post a sale thread later. Right now i'm still testing a buch of coating, and trying to reduce the response time mechanically.

 

[Benm]

My idea for setting up a measurement using two sensors works like this:

You take 2 sensors mounted on the same heatsink (which is no problem to obtain). One of the sensors is heated by the laser (and has to be coated), while the other is heated by a SMD-package resistor that is heated electrically. This sensor does not need to be coated to absorb light though it would be best if it were because the principle relies on identical sensors.

Having this all set up, you use an opamp based circuit that keeps the sensors at equal voltage by heating the resistor of the unlit one. Basically this results in an amount of power dissipated in the resistor mounted on the unlit sensor that is exactly equal to the amount of power the laser puts on the lit sensor.

What you are doing is keeping two surfaces equally hot, one by the laser you want to measure, and the other by electrical heating from a resistor you know the value off, voltage across and hence power dissipation.

This may seem complicated at first, but it has a big advantage: it does not require a reference laser to calibrate: you know the electrical power to the resistor is equal to the laser power if both sensors are identical, so you just measure the voltage required to keep both at the same temperature and calculate from that.

To see how well this approach works with these sensors i'd need to have a couple, preferably of something that has a large supply and will be available in the future too.

If you could provide a few i'd be happy to cover the cost of the units and postage. I suppose they will fit a standard letter envelope so you could send them directly to me in Belanda, or otherwise i can have someone pick them up from a friend living on Bali in 2 months or so.

I realize this approach is somewhat unusual, but i think it is important to have something that can be built and used without calibration equipment. How well it works mostly depends on the speed of the sensor elements, which is why it requires hands-on testing.