Black Body Thermal References and Fraser's new Black Body project :)

[Fraser]

Another little project that I have started.

I already own some Black Body thermal reference sources for my thermal camera work but they are relatively rare so I buy any 'bargains' I find. This lead me to become the owner of a very nice quality unit that I will detail in this Post. Before I do that though, a little about Black Body thermal reference sources and why they are needed.

The Black Body Thermal Reference source

Thermal cameras and IR thermometers are factory calibrated to provide a measurement accuracy of around +-2C or 2% when correctly used. This calibration holds for many years with decent quality cameras and thermometers but it is a good idea to check the equipment for accuracy at least every year. In some cases, every time it is used ! (Lab work) Calibration can be effected by component aging and contamination of optical blocks. It should not be blindly assumed that if a thermal camera is functioning that it is still making accurate measurements within its original specifications. There are various simple ways to check a thermal camera or IR thermometer for reasonable accuracy. I will not cover that topic fully here but two that come to mind are the 'melting ice bath' for 0C and the 'Boiling Water bath' for 100C. Thermocouples attached to a suitable ambient temperature metal plate can also be used but errors are possible.

The 'industry approved' method for checking the performance and accuracy of a thermal camera or IR thermometer is to use a purpose designed Black Body thermal reference unit. Such units contain a Black Body thermal emission plate or chamber that is driven by a suitable heating or cooling system. The heating and cooling system is designed to provide very accurate temperatures at the emission plate and this is why these units are classed as an accurate temperature reference. The Black Body emission plate of chamber is often coated in a special high emissivity paint that appears matt black in the optical domain. This surface is delicate and should not be touched with hard objects or fingers ! The emissivity value of the coating on the emission plate or chamber interior is stated by the Manufacturer in the specifications and is often 0.98. It is important to know the emissivity of the emission plate for accurate measurements but it need not be 1.0, 0.98 or even 0.96, so long as it is a decent value, known value, and a relatively flat response across the operating temperature range of the Black Body reference source. Some reference sources use plain flat emission plates, whilst others use the superior concentric ring format. The raised concentric rings help to avoid thermal reflections causing measurement errors. Any surface with an emissivity value of less than 1.0 can produce thermal reflections.

So what is the common construction of a Black Body thermal reference source ?
There are several designs for thermal camera Black Body reference sources in use. Some are cheaper to produce than others and some provide greater accuracy of temperature. You tend to get what you pay for with these units as they are specialist and so all are relatively expensive. Expect to pay around US$800 for a cheap one and many thousands of US$ for the better quality units. I did say they were expensive !
The form that the heater or cooler takes in a Black Body thermal reference takes dictates the complexity of the design. A few common types are detailed below:

1. Heater element

In this design, the Black Body plate or chamber is directly heated by a suitably constructed heating element and cooling is by natural radiation, convection and conduction paths. It cannot produce temperatures below ambient. The heater is controlled by a closed loop analogue or digital temperature controller that employs Fixed DC, Variable DC or PWM as the drive for the heater. The temperature sensor is normally a Thermocouple, Thermistor or RTD.
 
2. Heater element and fan cooler

Similar to the 'heater only' design, this Black Body uses a suitable constructed heating element that directly heats the plate or chamber. The design also incorporates a fan that can pass cooling ambient air over the plate or chamber in order to reduce their temperature more quickly when required. The fan is not normally part of the temperature stabilisation process. The heater is controlled by a closed loop analogue or digital temperature controller that employs Fixed DC, Variable DC or PWM as the drive for the heater. The temperature sensor is normally a Thermocouple, Thermistor or RTD.

3. Hot air blower utilising heater and blower fan unit

In this design the emission plate of chamber is heated to the required temperature by the flow of air over its rear/outer surface. Some designs employ air paths within the plate of chamber walls. The air is heated above ambient temperature by a motorised turbine / heater combination. Such is commonly found in hair dryers. The temperature of the air exiting the blower assembly and the emission plate/chamber is monitored by the heater controller. The drive to the heater is adjusted to provide the required reference temperature. Depending upon the design, the heater is driven by DC, PWM or Triac chopped AC. The fan speed may be constant or variable to suit the air path design and Delta T response time required.

4. Water jacket with temperature controlled reservoir utilising a heater or refrigerator

A 'Wet' Black Body reference source is normally intended to be static and not portable. It comprises an emission plate or chamber that is directly coupled to a relatively small water chamber that either sits behind it (plate) or around it (chamber).
The water chamber has inlet and outlet ports that permit water to be pumped through it at the desired rate. The water is part of a closed loop system that incorporates a main (large) water reservoir, water heaters and/or coolers, a water pump and a small reservoir located at the emission plate or chamber, as already described.

The water in the main reservoir is temperature controlled using heaters and/or coolers to maintain the required temperature. The heated or cooled water is then pumped around a closed loop water circuit that includes the small reservoir and so heats or cools the Black Body emitter. The system has a high thermal capacity and slow Delta T rate. This makes it an accurate, but slow to respond system that is often used at a set temperature without the need to change that temperature quickly. The water in the main reservoir is normally heated by a conventional heating element and drive circuitry from the temperature controller. The temperature controller often monitors the temperature of the emission surface, main reservoir and even in the path to the small reservoir. It can control the temperature of the main reservoir and the flow rate of water around the closed circuit. Where water cooling is required, this may be provided by a Peltier cooler or conventional mechanical refrigerator. The Peltier can act directly upon the main water reservoir metal casing, whereas the refrigerator tends to operate using a cooling circuit coil within the reservoir. A motorised stirrer may also be included in the main reservoir to equalise the temperature throughout its volume of water. With cooled main reservoirs it is sometimes necessary to use a mixture of glycol and water to prevent freezing at sub Zero C temperatures.

There are versions of the 'wet' Black Body reference that have the emission plate or chamber integrated into the side of a single large reservoir. This removes the need for the small reservoir, interconnecting hoses and pump. The single reservoir system can suffer from a vertical temperature gradient across the emission plate. For this reason a water stirrer is desirable to ensure the thermal layering within the reservoir remains disturbed and the temperature throughout the water relatively well distributed. 

It will come as no surprise that 'wet' Black Body thermal references are intended for Lab use and are very expensive.

5. Peltier combined heater and cooler controlled by accurate temperature controller

The Peltier heating and cooling element is a very useful tool when considering how to create heating or cooling at low loads. Their power consumption can be a significant drawback however. The Peltier element has a lot to offer the designer of  Black Body thermal reference however.

1. It can heat or cool a surface depending upon the polarity of its supply
2. It is solid state with no moving parts
3. It can change its surface temperature quickly, only limited by the thermal load placed upon it.
4. It is relatively easy to drive with either variable DC or PWM drive types.
5. It is relatively compact but it does require a heat-sink and sometimes a fan.
6. It is relatively long lived and robust if treated properly
7. Relatively inexpensive these days.

A Peltier element is capable of producing a Delta T across its two plates in excess of 70C. Operating at very high Delta T does reduce its life however. Special 'high temperature' versions are available however. With an Ambient temperature of 20C, a Peltier element system can easily produce a surface temperature of -20C to +80C and even -45C to +100C if auxiliary cooling and heating is employed.

The Peltier element is usually driven with either a varying DC voltage, DC Current or a PWM signal. They should not be driven by simple low frequency switched DC if long life is desired. Slow DC switching causes thermal shock within the device. They require a decent controller in order to be used as a thermal generator for a Black Body thermal reference as their temperature needs to be closely controlled to prevent unacceptable levels of overshoot in the system. A PWM controller is the favoured choice in modern microprocessor controlled Black Body Thermal References. Analogue closed loop control systems are manufactured for controlling Peltier elements and these can be very accurate if designed correctly. The controller for a Peltier element differs to that used for a simple resistive heater in that it is often capable of reversing the polarity of its output drive to permit the Peltier to both heat and cool a surface. The peltier element based Black Body thermal reference is therefore able to produce temperatures above and below ambient. A very useful feature. When working at very low temperatures, consideration must be given to condensation production and icing of the Black Body emitter surfaces.

When considering the design of a Block Body reference source, the Peltier element certainly ticks a lot of boxes ad is a well understood technology. Its one major drawback is efficiency at less than 6%. The elements have limited load capacity and high current draw normally measured in Amperes ! It is not uncommon to drive a common Peltier element with 6A at 12V. Power hungry beasts ! Sadly this hunger for power also effects how well the elements can cool. At some point in the cooling process the self generated heat within the Peltier element starts to degrade the cooling performance. It is then time to consider using multiple Peltier elements in a 'Stack' to reduce the drive needed on each element. Such is not normally needed for a Black Body Reference source however so I will not expand on the issues related with Peltier stacking.

The temperature sensor used by the Peltier controller may be any type that offers decent accuracy in this application. It is common to find Thermistors and RTD's such as the PT100 and PT1000. A combination of well designed Black Body emitter, Peltier elements and Peltier controller can produce an excellent Black Body reference source accurate to within 0.1C of setting. At these levels of accuracy, air currents across the emitters surface even have an effect ! This may be controlled using appropriate housings and tubes to limit such air movement contamination. Operating the emitter plate at sub Zero C temperatures can necessitate the use of Aluminium coupling tubes to the camera that are flushed with Nitrogen gas to prevent condensation and frost issues. That is a whole other topic however.

The Peltier element based Black Body Thermal Reference can be a very nice item to have in your thermal camera test equipment arsenal

6. Ambient temperature reference with accurate thermometer

An Ambient temperature reference source is the simplest of references for use with thermal cameras and IR thermometers. It's thermal emission plate of chamber is totally passive. The only active components used in such a reference are within the thermometers that provides the user with an accurate temperature reading from the emission plates surface. The emission plate sits at whatever the ambient temperature of its location may be. These are best used in an environment that has reasonably stable ambient temperature. Operation is simple. The thermometer is switched on and allowed to stabilise. The sensor on the rear of the emission plate of chamber provides a good level of accuracy that the user notes. The emission plate of chamber has a specified emissivity that is entered into the camera or IR thermometer. The camera or IR thermometer is then aimed at the emission plate of chamber and the reading noted.
If the reading is not within the expected tolerance then it s likely that the camera or IR thermometer requires attention or calibration. It does not get much simpler than this type of reference source. Point of failure are limited to the thermometer module and degradation of the emission surface. For this reason, such easily portable check sources are quite popular in some industries. Some also contain ports in the emission plate/chamber for insertion of conventional fluid based thermometers to provide further confidence in their ambient temperature reading.

Such a simple check source is eminently suitable for DIY construction. The required parts are available cheaply and construction is simple. I detailed such a project in another Post some time ago.

The disadvantage of the Ambient temperature check source is that it can only provide a single point of temperature reference as opposed to the wide range of temperatures available on other thermal reference sources. It cannot truly be used as a calibration check source as a result.

Fraser

[Fraser]

The Black Body Thermal Reference Project

Well to be honest, this is more of a restoration to service than a true building or repair project. I  am detailing it just in case it is of interest.

I have been watching an auction on ebay for some time and negotiations with the seller had been 'challenging'. The auction was for a Galai BB-50 IR Source that was Military surplus. Those in the thermal camera Industry will recognise the name 'Galai' as an Israeli company that manufactures Military hardware. The Galai name is synonymous with quality

The Galai BB-50 is a Peltier element based Black Body Thermal Reference that offers excellent accuracy and an extendable temperature range. It is also interesting in that it comprises the BB-50 head that is connected to the Galai 204E controller unit via an umbilical cable. This makes the unit very versatile with respect to positioning.

The stand alone temperature range of the BB-50 head is -20C to + 100C. With the use of a temperature controlled +5C water feed, the lower temperature limit is extended to -45C. Emission plate Temperature accuracy is excellent at +-0.1C . The emission plate Emissivity is 0.98 and it incorporates the concentric raised ring system to reduce issues with thermal reflections.

As stated, the BB-50 uses a Peltier element to produce the required temperature on the emission plate. The Peltier can both heat and cool the emission plate as required, hence its ability to produce temperatures below ambient. The head contains only the Peltier element, a PT100 RTD sensor and the metal housing that incorporates the water jacket and large finned heat-sink. The heat-sink is what the Peltier element drives against and it may be supplemented by the water jacket to extend the temperature range. The use of a water jacket is not that common on these Peltier Black Body References. It is not required for general use however.

A special Peltier temperature controller is required to drive the BB-50 head. This was not included in the auction and was my basis of negotiation with the seller. Sadly he was not great on the communications front The controller used with the BB-50 head looks like a common PID at first appearance. It is somewhat different however. The controller provides a very fast response time, very narrow dead band and a bi-polarity high frequency, high current, PWM output drive to the Peltier Element. A standard PID is generally slower to respond, has a wider dead band and the output is often just on/off relays or SSR drive. Simple ON/OFF switching is not recommended for driving Peltier elements. For these reasons, the missing controller for the BB-50 was a significant problem that could prove very costly to rectify. The seller was asking £100 for the BB-50 head but after protracted negotiations we settled on £80. For me, that price was just about acceptable for such a high quality Galai thermal reference head.

Upon purchasing the BB-50, my search for a suitable temperature controller unit began. I quickly realised that new Peltier temperature controllers that used an  appropriate drive type were very expensive. There are cheap on/off/reverse polarity output types available from China but they would be less than ideal due to potential damage to the Peltier Element over time and their very slow response combined with wide dead band. They also required a 10K thermistor temperature sensor and I preferred to use the PT100 RTD built into the head. That was not a showstopper but the ON/OFF type of control was.

Consideration was given to building a Peltier controller from scratch using an Arduino as the 'brains' and a high current H bridge to drive the Peltier element. Whilst such a PID is definitely achievable, its development would likely take time to perfect for the role required. I found some simple Arduino designs that are intended to drive Peltier elements for PC chipset cooling. They are far from optimal as a Temperature Reference controller. Without the opportunity to buy a complete dedicated Peltier Controller at a reasonable price, the Arduino route did appear my only option however.

I got lucky on eBay however......

I was carrying out searches for all manner of PID's and temperature controllers. I spotted a unit described only as a Wells-CTI temperature controller and noted it had no further details about its functionality. I marked it for follow up but continued to search for more obvious Peltier temperature controllers. After much searching in vain I eventually returned to the Wells-CTI EM series controller. Some Googling provide all the information I needed. It WAS a Peltier Element controller, and a very capable one at that ! It had great accuracy, it could use a Thermistor, Thermocouple or RTD sensor, it could provide up to 24V PWM at up to 19A ! It is purpose designed for the precision temperature control of a Peltier based test rig for IC testing. The cost of the unit new is $5000. It also provides the ability to remote control it from a PC and to set temperature profiles for ramping or stepped temperature programs. All in all an amazing bit of control kit perfect for my BB-50 head

I struggled to find a manual for the Wells-CTI temperature controller as the company was taken over and no manuals or the PC control software are available for download. I then searched on the PID that was used in the unit and still carried its original manufacturers name and model number....... Accuthermo FTC100D..... jackpot ! It soon became apparent that the PID manufacturer provides everything to control a Peltier element except the high current power supply. They had the user manual for the Specialist Peltier PID, the 700 Watt H Bridge and the software for PC control The Peltier control system is created 'building block' style as the manual showed in pictures. The PID connects to the H-Bridge and the high current SMPSU of appropriate voltage is connected to the H-Bridge. The SMPSU voltage sets the PWM signal amplitude.
The clever stuff is all done inside the Peltier PID Controller unit. It can accept temperature measurement inputs from Thermistors, RTD's and many types of Thermocouples. It creates the correct PWM signal output and a polarity control line to set the H-Bridge to the correct polarity output to the Peltier Element depending upon whether heating or cooling is required. It goes far deeper than this though. The Peltier PID is a dedicated controller for precision temperature control. It has a user definable  settings to fine tune the controller to best suit the system that it is incorporated within.
The Wells-CTI EM series controller cost £60 so was not exactly 'super cheap' compared to an Arduino or Chinese Peltier PID based solution, but this unit is the gold Standard in controlling a Peltier element and it is a complete solution that may be pressed into service quickly. I could not have built such a unit of similar performance for £60. My time is precious too.

I may have to modify the Wells-CTI controller depending upon which version I receive. They come with different PWM output voltage levels to suit the needs of the various Peltier elements. Until I receive the unit I do not know whether the integrated 'Universal' SMPSU offers voltage output setting or whether I will need to change the whole SMPSU if the provided output voltage is too high for my BB-05 Peltier element. If it is the 12V controller version, I am OK, if its is the 24V version....changes will be needed.

Both the BB-05 and Wells-CTI EM series controller are due to arrive with me this week so hopefully I will have the system up and running soon.

Fraser

[Fraser]

More pictures of the BB-50 system and components

[Fraser]

The Accuthermo FTC100D PC control application......

As can be seen , it is very configurable and is programmable so can create thermal profile runs for automated testing at different temperatures 

The unit also has a temperature logging capability to check on the Peltier thermal output temperature stability

The FTC100D does not need a PC for configuration as all can be done from the front panel if required. The PC setup is just nicer to view and quicker use than the units limited front panel controls.

Fraser

[firewalker]

Bookmarked for further study! 

Alexander.

[Fraser]

The Galai BB-50 original specification when used with the Galai 204E controller

Galai BB-50 specs when new

Temperature Range:
-20C to +100C (The lowest temperature achievable is dictated by the ambient temperature)
-45C to +100C (-45C requires fluid cooling of the Peltier element 'HOT side' down to +5C)

Accuracy: better than +-0.1C over entire -45C to +100C range
Reproducability: better than +-0.05C
Temperature Stabilization time: less than 10 minutes after a 30 minutes warm-up period
Emissivity plate diameter : 50mm with Anti Reflection concentric raised rings
Emissivity of emission plate: 0.98 over -45C to +100C range
Emissivity plate temperature Sensor: PT100 Platinum RTD


Fraser

[Gyro]

Hi Fraser, interesting project.

I'm just wondering how that emission plate works at below dew point and sub zero temperatures where it is presumably going to be coated with water or ice. I'm sure it must be accounted for, LWIR transparent clingfilm for instance?

On a slightly lower-end tack, I've been wondering about a simple reference myself, my needs are humble as will be seen from my TG130 review! I found a table that indicates that soot (coal or candle) has an emissivity of 0.95, so I was thinking a simple soot covered plate with attached resistor and thermocouple (I'm not interested in sub-ambient). Workable do you think?

Thanks,
Chris

[Fraser]

Condensation and frost/ice are an issue when operating a Black Body at very low temperatures.

In the lab test pictures in this thread, an aluminium tube was attached to the emission port and the thermal detector. The tube was flushed with Nitrogen gas to prevent dew and icing on the emission plate.

Making your own simple Black Body is relatively straight forward but the accuracy of an active (heated) type is very much dependant on the heater and temperature controller used. You want even heating of the emission plate so I would recommend a surface heater similar to those used in 3D printing platforms. The controller needs to be of reasonable quality to maintain the plate at a stable temperature. If it cannot hold the emission plate at less then +-1C drift, it is not much use really.

I will add a link to a nice surface heater in a minute. They are simple to drive and exactly what AGA/AGEMA used on their Black Body reference that I own. The Emission plate on the AGEMA Black Body is around 100mm diameter x 5mm thick aluminium.

As to coating the plate for high Emissivity, candle soot does work but is too fragile for practical use. I recommend buying some Matt Black spray paint from an Autoparts shop or, better still, Matt Black Krylon used on Military vehicle camoflauge. It has excellent emission properties. VHT Matt Black cat paint for car manifolds us also very good and can handle up to 600C without problems

Fraser

[Fraser]

The surface heaters used on the rear of some black body emission plates, like that of the AGEMA are Polymide self adhesive types as shown in the attached picture.

These are available cheaply from China. I intend to buy some different sizes for experiments.

https://www.ebay.co.uk/itm/1W-5W-7W-13W-20W-30W-PI-Polyimide-Flexible-Adhesive-Thermo-Heater-Heating-Film/352012311342?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160908105057%26meid%3D7a0d993ba34d4ce0bf0f88459db69914%26pid%3D100675%26rk%3D1%26rkt%3D15%26sd%3D352012311342&_trksid=p2481888.c100675.m4236&_trkparms=pageci%253A64278aaf-0040-11e8-9aff-74dbd180f9ee%257Cparentrq%253A232eb5eb1610a860f7f2b306ffd63ceb%257Ciid%253A1

Fraser

[Gyro]

Thanks Fraser, I think I missed the Nitrogen flush (I think they do that to Chickens too!  ).

That same table did have a reference to the Krylon paint too (as well as a 3M one I think), I found a link on Amazon, not cheap, but not extortionate either .... there again, I have loads of assorted 'no-name' cans of matt black in the garage, including BBQ high temp ones. I was thinking soot would be an easily re-formable surface and 'known' emmissivity (I've seen different figures for different paints). Thanks for the surface heater links. Yes, I think 1'C ought to be achievable with a reasonable thermal mass.

Chris.

[CatalinaWOW]

A good quick reference to the issues of plate coating is

https://wp.optics.arizona.edu/optomech/wp-content/uploads/sites/53/2016/10/Black-Coatings-to-Reduce-Stray-Light.pdf

Although is discussing how to make good baffles it is the same problem from another point of view.  Be aware that the values for the paints listed are very dependent on how they are applied and on the surface they are applied to.

A good option if building your own surface plates is to use honeycomb material.  This creates a multitude of cavities, which enhance the inherent emissivity of the surface.  Most infrared references will give simplied formulas for estimating this enhancement.  There are many trades involved in doing this.  Obviously temperature control of the honeycomb itself is an issue, with control decreasing as the thickness of the honeycomb increases, while emissivity enhancement drops as thickness gets down to the neighborhood of the cell size.  Cell size is also important, and the impact varies with the instruments used to view the surface.  In many cases cell sizes larger than the pixel size of the instrument are acceptable, and even desirable as pixels that are well contained within a cell can be reliably evaluated while those abutting or bridging a cell are suspect.  Ideal is for a cell size small with respect to the pixel, but this is often difficult to achieve and may require fabrication of your own structures.

[Fraser]

The Wells-CTI Peltier controller arrived today :

It is in great condition and looks perfect for my needs...... except.....

I took the cover off of the unit and found the expected configuration within, but the PSU fan exhaust blowing directly onto the PID's underside is 'interesting'. There is another fan pulling the air through the unit and out of the rear though. It could have benefited from some tidying of the cable loom, especially considering its $5000 price tag ! The 24V 12.5A  Mean Well SMPSU is securely mounted on the base plate and the H-Bridge is mounted on a mezzanine level created by a nicely made bracket on top of the power supply. The Mean Well power supply is nicely made and has its own integral fan for cooling. Sadly it is 'Universal' in terms of its 100-240V input but it is a fixed 24V DC output That will be too high a voltage for my needs so another SMPSU will be required. Thankfully the Mean Well power supply is a common 210x115x50mm case format, so other brands will fit

The Mean Well power supply will not be wasted as I need a decent 24V supply for another project

I started searching on the internet for a 12V 30A power supply to fit into the Wells-CTI casing. These power supplies are commonly used to drive LED strings and so are thankfully quite common at reasonable prices. I looked at many on eBay and could see that most were in fact the same unit. UK stock ranged in price from £12 to £24 depending upon who it was purchased from.         

I eventually settled on a decent looking 12V 30A SMPSU for £19, delivered, from a company that specialises in 3D printers and accessories. The company is UK based and I suspect they stand behind their products more than your average Far East eBay seller.

https://www.ebay.co.uk/itm/DC-12V-30A-360w-Universal-Regulated-Switching-Power-Supply-LED-3D-Printer-CCTV/322769962251?hash=item4b2696c50b:g:i-gAAOSwjPdZxK1N

It includes an integrated cooling fan as per the Mean Well unit so cooling is not a concern.

The terminal strip connections are helpfully similar to the Mean Well original unit:

Mean Well   +V  +V  -V  -V  PV  RC  E  N  L

S-360-12      +V  +V  +V  -V  -V  -V  E  N  L

The S-360-12 SMPSU should be a drop-in replacement for the original power supply and then I will be ready to test the controller with the BB-50 head.

Fraser

[Fraser]

The replacement SMPSU.....

[Fraser]

Oh dear, I have had a DOH moment today !

I just downloaded the manual for the Mean Well power supply module.

The 'PV' pin permits changing the output voltage to anything in the range 4V8 to 26V4 

I can change the SMPSU to 12V output by applying approx 3V to the PV pin.

A quick message to the seller of the 12V replacement power supply has hopefully cancelled the order.

As the old saying goes... read the flippin manual !

Fraser

[Fraser]

I have now had a chance to quickly reverse engineer the H-Bridge board that is used in the Wells-CTI controller. It is a modified version of the Accuthermo 700W H-Bridge design.

This gets better and better for me 

The H-Bridge is controlled by the PID controller and receives both a PWM signal and a polarity control line. Normally the H-Bridge PCB would just act on the PWM signal and output the polarity dictated by the the PID. The H-Bridge PCB in the Well-CTI is slightly different as we will see.

The main IC's used on the H-Bridge PCB are as follows:

Alegro A3941KLPT H-Bridge External MOSFET driver

National LM2574HVM-ADJ+ Step Down regulator (Adjustable)

Linear Technology 3080 Low Dropout adjustable regulator

And the four power MOSFETS of course.... type IRF2907ZS-7P

Nothing very exciting really but there are some additional features included on the PCB.

1. There is a 'Thermostat' input to the board from the Peltier drive connector. This is a safety feature that shuts down the drive from the H-Bridge if the over temperature mechanical thermostat is operated in the Peltier head.

2. There is an output from the H-Bridge PCB to the Mean Well SMPSU 'PV' input. The 'PV' input permits external control of the output voltage from the SMPSU. So why would the H-Bridge PCB be driving that control pin ? Well this Peltier controller unit is unusual in that it can apply a different voltage to the Peltier depending upon whether it is in teh Heating or Cooling mode. How does it do this ? .....

The LT 3080 Adjustable LDO drives the 'PV' control line and its adjustment pin is connected to two potentiometers that are individually selected according to the Heat or Cool polarity signal coming from the PID. this effectively means the output voltage of the Mean Well power supply may be customised simply by changing the value of a couple of resistors and the potentiometers may be used to fine tune the voltage for Heating and Cooling.

So to summarise, I have a 700W H-Bridge, to drive the Peltier element, that incorporates a safety thermal cut-out system and a means to control the SMPSU output voltage 

I will modify the resistor values on the PCB's SMPSU PV control circuit to set the required 12V for the BB-50

Fraser

[Fraser]

A quick update.....

The BB-50 Peltier head arrived today. There is some minor coating damage to the emission plate but nothing I cannot repair with the Krylon paint that I already have 

A quick test of the Peltier element on a lab power supply proved it to be in fine condition and working very well indeed.
I did a test run on the head up to 7V to see how it performed in cooling mode. I plotted a graph of voltage against current and the Peltier is tracking perfectly.

Results:

4V = 1.8A = -10.1C
5V = 2.4A = -12.1C
5V5 = 2.7A = -13.0C
6V0 = 3.0A = -15.0C
6V5 = 3.27A = -17.0C
7V0 = 3.5A = -19.6C

The lowest measured temperature was -20.6C after the head had been sat with 7V on it for a few minutes.
This is the temperature that Galai claim the head can achieve in normal Ambient temperatures of 20C without additional cooling. The heat-sink temperature rose to +33.8C during the test.

After plotting the V-I graph I extrapolated the current drawn at higher voltages than 7V in an effort to identify the Peltier element that is being used in the sealed head assembly. My graph suggested the following approximate V-I specs to be used to search for likely Peltier model candidates.:

5V 3A
8V 5A
12V 7A
15V 9A

The cable used on the BB-50 head to power the Peltier looks to be of around 5A capability but I did not exclude the possibility of it being run at 9A using PWM. I had previously suspected that the element would be a common 12V type. In fact, the common model of Peltier is rated at around 15V and you just under-run it at 12V if desired. For this reason I suspected the head contained a 15V 9A Peltier element that was either run at its full voltage or throttled back to reduce internal heating effects. The Peltier needs to be of approximately 40mm x 40mm to match the emission plates 50mm diameter. This also suggested a 15V 9A Peltier as the lower voltage elements are of smaller dimensions.

With all of this in mind and knowing that if I was careful, I would not fry the Peltier element, I decided to apply 14.6V to the head using the Meanwell SMPSU that I already have. I set it up to provide 14V6 instead of 24V. At this voltage the BB-50 head would draw around 8.5A. Interestingly, the emission plate temperature did not tumble as I had expected. in fact it seemed slower than when running on 7V ! I monitored the emission plate and heat-sink temperatures carefully. The emission plate did get down to a low temperature but I noted that the heat-sink was having to work a lot harder during this test as its temperature rose to 35C yet the emission plate stabilised at only +4C 

This was all the proof that I needed that the Peltier element was not at all happy running at 14V6. What was happening ? Well you can run a Peltier element above its stated Max V without killing it provided you have it attached to suitable heat-sinks so that it does not cook itself. What you discover though is that the energy that is being forced into the Peltier Element is not all being used  in the heat pump process. The excess energy heats up the Peltier Element, even if it is in cooling mode and the cooled surface is contaminated with heat energy. Hence why the emission plate only reached +4C with a heat-sink temperature of 35C present.
Running the same Peltier element at 7V with an initial 4A draw (reducing to 3.5A) the emission plate achieved -20C with a heatsink temperature of 33.8C.

My portable lab power supply is limited to 3.5A so I have not tested voltages higher than 7V but I suspect that I am close to the optimum voltage to use with the BB-50 head. This also makes sense when the connecting cable cross section is considered. Approx 4A current draw can be handled by it with ease, 9A less so. I will carry out more tests using the Wells-CTI controller after I have tweaked the SMPSU voltage.

Wells-CTI controller update

As readers will already know, I discovered that the Mean Well SMPSU is capable of having its output voltage reduced using the 'PV' control line. The specifications state that it is happy to provide any voltage between 4V8 and 26V4 with the 12.5A current capability unchanged. As I do not need 12.5A in my application this was not a great concern anyway.

I have now planned the required modification to the H-Bridge PCB that drives the SMPSU 'PV' line.

As supplied, the two small potentiometers, one for heat, the other for cool, have an SMPSU output setting range of 14V6 to 24V. They are used in series with a 300K resistor on the SET pin of the LT3080 adjustable LDO Regulator that generates the PV voltage. I will not bore you with detail but basically a single resistor sets the LDO output. 1M Ohm = 10V, 500K Ohm = 5V  200K = 2V etc. Simple 

For my application I need to generate a PV voltage range of 1V1 to 2V1 which will provide an SMPSU output range of 5V to 10V
The two potentiometers are 230K Ohm so it did not take a rocket scientist to realise that bypassing the 300K resistor in series with them would provide the required adjustment range. With that resistor bypassed, the two potentiometers will provide an SMPSYU output voltage range of 4V8 (minimum possible) to 10V8. Perfect 

Once the Wells-CTI controller is modified I can carry out tests using its Mean Well SMPSU to ascertain the optimum voltage for driving the BB-50 head. I will then set both the Heat and Cool potentiometers to the same 'PV' output voltage setting.

After this, I only have to fit the required AMP plug to the BB-05 head umbilical cable 

In case anyone is interested...... when the emission plate got down to -20C it was covered in frost. You could clearly see warm air swirling past its face. Quite interesting to watch. I used my E60 to monitor the emission plate and you could see when the ice crystals started to form during the cool down process... they looked like twinkling stars on the emission plate  . I will post pictures from the E60 shortly. To test below -20C I will need to use another camera that is rated for such low temperatures.

Fraser

[Fraser]

I have just completed another set of tests after modifying the H-Bridge PCB 'PV' control voltage circuit. The results are excellent.

A summary of todays activities.....

1. The H-Bridge PCB design was studied with regard to potential issues associated with reducing the SMPSU output to 7V. The SMPSU output provides the power to the H-Board controller and 'PV' control circuit. The LM2574 is a step down buck converter set to provide 12V output from a 24V input. It provides power for various LED indicators and the 'PV' control circuit. Whilst not ideal, undervolting the input with 7V would just cause it to drop out and provide 6V at its output. I considered bypassing it but thought it interesting to see if that could be avoided. Good voltage regulation is not an issue with the simple transistor and LED circuits that it drives. The 'PV' control circuit makes use of a LT3080 LDO to generate the required 'PV' control voltage. The 6V supply from the LM2574 is more than enough for its task. The H-Bridge controller was a concern, but thankfully the A3941 is designed for automotive use and is VERY tolerant of supply voltage variations. It contains its own boost converters and charge pumps so will operate on voltages as low as 5V5. 7V is specified as perfectly acceptable for normal operation  The A3941 takes its supply directly from the SMPSU 7V output. So there were no obvious issues with reducing the SMPSU output from 24V to just 7V.

2. The modification to the 'PV' control circuit on the H-Bridge PCB was simply a case of bridging the 300K resistor in the 'SET' leg of the LT3080.  This left just the 230K resistance of the Heat and Cool 'PV' setting potentiometers in circuit. They are a good fit in terms of value. The modification was completed and the output range of the SMPSU was tested. Using one of the 'PV' setting potentiometers I was able to set the SMPSU to any value in the range 4V8 to 11V5 .... excellent. As the voltage was reduced, the case mounted fan and that of the SMPSU did reduce speed but both are more than adequate for the 3A that I will be drawing from the supply. I set the SMPSU for 7.0V in readiness for further testing of the BB-50 head at voltages above 7V.

3. A test of the BB-50 head was carried out using voltages above 7V to establish the Peltier Elements most efficient operating voltage. The voltage would be used to cool the emission plate as in yesterdays test as cooling to low temperatures gives the best insight into the Peltier elements efficiency and potential thermal contamination of the cold side through self heating. The power source for the test was the newly modified Wells-CTI controller but I was just using its 'PV' control circuit and the SMPSU. I was not using the PID or H-Bridge for these 100% cooling supply tests. The emission plate was monitored with my FLIR E60.

4. Upon completion of the above test, it was obvious that the Peltier performed at its best when supplied with 7.0V. Any higher voltage resulted in increased self heating and a loss of cooling performance. The results of the test are shown below:

Voltage     Emission Plate      Heat-Sink

7.0V         -23.3C                 +32.5C
7.5V         -22.8C                 +34.2C
8.0V         -22.0C                 +35.9C
8.5V         -22.0C                 +38.0C
9.0V         -21.0C                 +39.2C
9.5V         -19.0C                 +41.5C
10.0V       -18.0C                 +43.2C

Note the decrease in cooling ability and increase in Heat-Sink temperature as the voltage is increased above 7.0V. A sure sign that the Peltier element is no longer within its 'comfort zone'.

5. Another test was set up with the SMPSU providing 7.0V to the BB-50. The emission plate and heat-sink was monitored over a period of 30 minutes for temperature change, using the E60. The results were further proof that the BB-50 head worked very well with a 7.0V input. The results of that test are shown below:

Voltage     Emission Plate      Heat-Sink

7.0V         -21.6C                 +30.9C
7.0V         -21.0C                 +34.8C
7.0V         -21.4C                 +34.8C
7.0V         -21.2C                 +35.0C
7.0V         -20.6C                 +35.5C

I was happy with these results. No control is being used on this Peltier Element. It is running 'free' with a 7.0V supply driving it in hard cooling mode. The Emission plate stays below -20C and the Heat-Sink does not display an excessive increase in temperature. The ambient temperature in the room was 20.3C. As the Heat-Sink temperature rises, so does the cold side of the Peltier element. This can be clearly seen in the test results. this is why Galai state that -20C can be achieved without additional water cooling provided the ambient temperature is not too high.


Well that is the testing completed for now. The Wells-CTI controller will be configured to produce 7.0V and I am expecting the required TE CPC 206708-1 connector for the BB-50 umbilical cable in tomorrows post.

Once all is connected up and running, I will turn my attention to resurfacing the emission plate with the appropriate paint.

Fraser

[Fraser]

No progress on this project since my last posting. Zero available time sadly. The correct Tyco-AMP connector has arrived and will be fitted onto the BB-50 umbilical soon. I considered the issue of the now unnecessary pre-regulator that has gone into drop-out due to insufficient input voltage. It drops 1V across itself in this mode. That volt is quite a lot in percentage terms so I will be lifting the input, output and feedback pins on the chip before bypassing it. That will provide the full 7V to the 'PV' switching and control circuits. I did not like the idea of leaving a switching regulator permanently in drop-out mode anyway.

I also considered fitting a 7.5V Mean Well SMPSU in place of the 24V unit that I have trimmed down to 7.0V. The 7.5V unit has output trim of around plus or minus 1V so could produce 7V. I was just a little concerned about trimming a 24V SMPSU all the way down to 7V. It did not sound a great idea as the PWM output must surely have dropped considerably to achieve a third of the nominal output voltage  ? Having read the data sheet carefully and seen that the SMPSU uses a 100kHz oscillator, I can see that it is an expensive unit that is designed to have its output lowered all the way down to around 5V without issues. For that reason, it is staying in the Controller as it is under no stress and should be reliable.

More when I get time.

Fraser

[Fraser]

Another tiny update.......

A combination of OCD and a use that I have for the Mean Well 24V 12.5A SMPSU have resulted in a nice shiny new Mean Well SP 320-7.5 SMPSU arriving today. It produces 7.5V with an output voltage trim control, so 7.0V can be set on it. It was being sold as excess stock with some minor spots of corrosion on the aluminium case, so I got it cheap at £20. Not bad for such a SMPSU from a well known brand  The power supply case and connector block format are identical to the original 24V Mean Well unit so it is a straight swap out job.

I will install the new power supply and can ignore the 'PV' control circuit on the H-Bridge as it is not needed on the SP 320-7.5. The new power supply is rated at 300W (40A) so it will not be 'breaking into a sweat' providing up to 4A for the Peltier element in the BB-50. Plenty of headroom.... just how I like it 

Fraser

[Fraser]

The project is now almost completed.

I have fitted the new 7.5V SMPSU, trimmed its output to exactly 7.0V and tidied up all the cabling inside the controller. No modifications to the H Bridge PCB were needed.

The umbilical cable from the BB-50 head was easily terminated with the Tyco/AMP CPC connector. I have the option to connect the BB-50 directly to the controller or via the extension umbilical cable that came with the controller 

Initial testing of the controller with the BB-50 has proved the combination to work perfectly together. The PID in the controller is a very clever bit of kit that is equipped with a very effective auto tune capability.
The auto tune mode carries out a thermal profile test on the Peltier load. With the gathered information, the PID can predict the required PWM signal needed to maintain a very tight temperature tolerance at the head. The resulting load profile is stored in non volatile ram in the PID. The auto tune function works really, really well  The PID now knows how best to drive the BB-50 and the temperature stability of the emission plate appears superb  I can program any temperature in the range -20C to +100C with ease. I can go beyond that temperature range but I would need to use the supplementary water circuit for that and I really do not need to go down to -45C very often. I am fortunate to already own a lab grade water cooling unit that is more than enough for the BB-50's needs if I ever need to head to the lower temperatures.

I still have to recoat the emission surface as it is slightly damaged, but I first want to test the unit and establish its current emission characteristics. I will then test and select a suitable paint to use on the emission plate.

Pictures will follow but I concentrated on getting the unit up and running today. This has been a fun little restoration project and my initial hopes for the Galai head have been fulfilled. It is a great thermal camera IR test source 

I will have to play with the PC software and control of the PID next. That enables me to set complete thermal programs to change the emission plate temperature over a period of time. That in combination with FLIR ResearchIR's thermal profiling capability would enable me to check my thermal cameras accuracy over a range of temperatures and over a period of time. Very useful 

Fraser

[Fraser]

Some quick pictures of the BB-50 head and modified Wells-CTI controller working together 

It is easy to create high temperatures with a Peltier driven emission plate, but the real challenge is getting down well below 0C.

As can be seen, the controller holds the temperature at spot on what is selected  It is rare to see the digit after the decimal point change +-0.1C  Remember, the head is using a 3 wire PT100 RTD so the controller is getting decent temperature information from the emission plate.

To reach -20C from an ambient of +20C takes around 10 minutes. Increasing temperature is a very fast affair, with the emission plate very quickly reaching +45C from +20C ambient. As I said, low temperatures test a Peltier cooler more than generating higher temperatures.

When at very low temperatures, the expected frost appears on the emission plate. I have tried to capture this on the pictures. You can also see the minor damage to the emission surface. I have yet to decide whether to just repair the localised damage or recoat the whole emission plate. I can always try a localised touch in repair and then recoat the whole plate of the result is not satisfactory.

Fraser

[Fraser]

I have been watching the PID's behaviour whilst controlling the H-Bridge and it really is interesting to see how it drives the Peltier module. The H-Bridge has 'Heat' and 'Cool' LED's on it so you can see what is happening. The PID also displays its PWM behaviour on the front panel.

So what does the PID do when you set it to a new temperature......

1. The PID sets the correct voltage polarity on the H-Bridge and ramps up the PWM signal to 'full power' in order to drive the Peltier module to the new emission plate temperature quickly. The temperature rapidly increases when heating, but slowly decreases when cooling. That is just the physics of the Peltier technology and BB-50 head.

2. As the target temperature is approached, the PID decreases the PWM drive to the Pelier module which slows the Delta T. Once very close to the target temperature, the PWM drive is further reduced and the PID starts applying the 'thermal brakes' to control 'overshoot'. This is done by by short pulses of the opposite polarity to the Peltier module. This generates the opposite temperature generation and helps to bring the Delta T under tighter control. Once the target temperature is reached, the PID alternates the polarity of the PWM drive to the Peltier module and keeps doing this until it is 'happy' thatbthe target temperature is stable on the emission plate. Once 'on target' the PID adjusts its PWM signal to maintain temperature and occasionally it will change polarity to maintain thermal balance at the Emission plate.

3. When the target temperature is higher than the current setting, this PID activity detailed above occurs very quickly and a very stable target temperature is quickly reached and maintained, thanks to the active PWM and alternating polarity drive. When the target temperature is lower than the current setting the PID activity is the same but over a slightly longer time period. If the target temperature is very low, and sub zero, the PID drives the Peltier hard for a long time and then starts reducing the PWM drive as the target temperature is approached. There is far less polarity reversal used as the target temperature is reached as the Peltier has to fight thecambient temperature hard to keep the emission plate at a low temperature.

4. As already detailed, the Peltier PID uses an 'Auto Tune' program to collect information on the Emission plate and Peltier modules capabilities. It determines how much energy is needed to be driven into the Pelier element for a given Delta T positive, or negative, at the Emmission plate. It effectively determines the emission plate thermal mass, temperature overshoot characteristics and natural (unpowered) drift. It uses this information to create its heating and cooling profile and minimise overshoot. All very smart and it does its job very well.


A little test to see how an alternative Peltier control method performs.......

The PID has many modes and options in its menu system. One option is to use on/off Peltier control rather than PWM. Drive Polarity is still used to control the Peltier but the drive level is set at a static percentage by the user and the PID just turns it on and off to control the Emission plate temperature. This is the mode of operation commonly seen on the cheap eBay Peltier PID units. They look suitable for the task but are nowhere near as sophisticated as the high quality PID fitted in the Wells-CTI controller.

I decided to set the PID to the on/off drive mode to see how stable the emission plate temperature would be. Drive was set at 100% as that is what the cheap Chinese PID's use.

The results were interesting and 'as expected'. The on/off + polarity reversal drive method is pretty brutal to the Peltier module. As an analogy, think of it like driving a car forwards at 40mph, dropping the clutch, selecting reverse, and hitting the throttle to get to 40mph in reverse, then clutch, 1st gear and 40mph forward again. And repeat ! Brutal as I said.

So what happened in my test with the PID ? Well the PID applied 100% drive in the correct polarity to reach the target temperature. No PWM to slow things down so it just used a change in polarity to apply the 'thermal brakes'. Overshoot occurred so the opposite polarity was applied to bring thevemission plate back to the correct temperature, another overshoot occurred in the opposite direction. The PID then applied constantly alternating polarity at 100% drive in order to stabilise the emission plate temperature. It was all rather hard to watch after having seen the delicacy with which the PID had operated in PWM mode. It was just a brutal full power heat, full power cool, mode of operation and the result was a wandering emission plate temperature that was plus or minus at least 1C around the target temperature and there was significant Delta T ripple present. Not a great Black Body reference IMHO. A better mode of operation may have been to not alternate between polarities and just let the ambient temperature work against the PID drive that was being pulses on and off. A bit like the thermal control used on a soldering iron, active drive that is offset by natural cooling or heating through convection, radiation and conduction. This would likely have decreased the Delta T ripple frequency significantly. The Agema Black Body operates in this drive vs ambient temperature mode of operation and manages to maintain an accurate emission plate temperature using an analogue control loop.

In the case of my PID, I could, and should have, reduced the drive level to say 20% as that would have made the temperature control of the emission plate far less frantic, but much slower to reach a new target temperature.

So in conclusion, I can recommend a Peltier element based Thermal reference design if it is controlled using a good quality Peltier PID that utilises PWM for drive control. I cannot recommend the simple 100% drive on/off type Peltier PID's that may be purchased for less then £20. Simple On/Off control of a Peltier is not recommended by manufacturers either. Such can cause thermal shock damage to the Peltier event over time. It is a brutal control method for a Peltier as I have stated.

A simpler thermal reference design would be based on just a heater attached to a thermal emission plate or cavity. The heater would be controlled by either an analogue temperature control loop or a heater PID. The analogue control loop offers the advantage of constant drive variation to maintain the correct emission plate temperature. A simple on/off digital PID would work but a far better choice would ce a PWM drive PID that could effectively replicate the fine control over the heater of an analogue control looop. The down side of a heater based thermal reference is that it's minimum emission temperature is dictated, and limited to, ambient temperature. Temperatures below ambient are not possible without an active cooling system in addition to the heater.

 What is an active cooling system based emission plate ? It is where the emission plate is connected to a cooler, be it Peltier or refrigeration plant based, and driven to a low temperature. Cooling system temperature stability is not critical in this case. The emission plate is lowered to whatever temperature the cooling system is capable of and then a heater is used to offset the emission plate to whatever temperature above that point is required. So you could have the emission plate cooled to 5C and the heater the. Heats the plate to 10C whilst working against the cooler.This system permits temperatures to be set below ambient and with some precision, using just the heater drive to set the target temperature. It is useful but very inefficient in terms of energy consumption. You basically have a heater fighting against a cooler..... a bit like a 'tug of war'  This system is sometimes used in climate chambers where the chamber is cooled by a refrigerator and the temperature is fine tuned using a PID controlled heater.

Well that is enough from me for the moment.

Fraser

[Fraser]

The Chinese WH7016K model Peltier controller.

https://www.ebay.co.uk/itm/WH7016K-10A-12V-Digital-Semiconductor-Temperature-Controller-Thermostat-Peltier/322573356278?hash=item4b1adeccf6:g:5RcAAOSwiQ9ZVkXv

It uses on/off switching plus polarity reversal to drive the Peltier. I was considering one of these before finding the Wells-CTI controller. They are certainly reasonably priced at around £15 delivered, but I do not think it uses PWM, and if that is the case, I cannot recommend it for use as part of a Peltier thermal reference system.

Fraser

[Fraser]

In case anyone is wondering about a suitable coating for a thermal emission plate, I use Krylon Ultra Flat Matte paint. It isn't very expensive at £7 a can, and works very well. It was recommended to me by a thermal camera designer so it is a good bet that he knows about decent Emissivity paints 

http://www.patrolbase.co.uk/krylon-spray-paint/KrylonBlackSprayPaint

I will be using this to touch in the scrapes on the BB-50 emission plate. It sticks to metal well but aluminium can be a challenge to many paints so use some fine 'wet and dry' abrasive paper on aluminium to create a key for the paint. Krylon can also cope with my need for a max Temp of 100C.

I attach the NASA JPL plot of the Krylon Ultra Matte paint Emissivity. It is paint 1602 in the plot. That is the brown plot and the best performer in that groups test.

https://masterweb.jpl.nasa.gov/reference/paints.htm

Fraser

[Fraser]

If only I could get my hands on some Nextel 3101c10  paint shown in this plot of high emissivity paints.

[Fraser]

NASA's plot if the four best performing paints. Krylon Ultra Matte Black (1602) is in the top 4 and thankfully easily purchased, unlike most specialist high Emissivity coatings.

Fraser

[Fraser]

The Nextel paint is a suede paint, as found in cameras. It has a soft plush finish ! This explains why the Matte finish works so well. Not quite what I want on a thermal emission plate though.

Fraser

[Fraser]

FLIR's comment on enhancing Emissivity........

http://www.flir.com/science/blog/details/?ID=71556

Who would hav thought that 'white out' correction fluid was useful in thermography eh ?  As the article says, the visible colour is not that important, more important is the finish. Matte is best.

Fraser

[SeanB]

Looking at your scratches you probably will have to remove the plate, strip it down to bare metal and repaint it Fraser, those scratches reveal corrosion under the paint, which will flake off the rest with time as it progresses. Probably came from condensed water containing salts, that went through the existing porous paint and attacked the underlying casting. You will be best off ( unless you are happy with the small difference in temperature in spots) stripping off all, applying a decent thin layer of chromate primer to the well sanded metal and then putting a few even coats of the matt over the primer, probably around 6 thin even coats.

[Fraser]

More comment on Krylon Ultra Matt Black

(2006)
A. The emissivity of Krylon Ultra Flat Black Paint is listed in the web site below as 0.97 at 5 um. I have recently used a 3-5 um InSb radiometric thermal imager to measure the emissivity. I sprayed the paint on a copper plate and mounted it on a cold finger of a variable temperature dewar with a calcium fluoride window. I rotated the dewar window so as to not back reflect. I measured an emissivity of 0.89 for the setup. If we exclude the reflection from the window (~0.04 x 2), this would give an emissivity for the black paint of ~0.97 in agreement with the web site.

http://www.infrared-thermography.com/training.htm

From this site.....

https://www.finishing.com/1600-1799/1670.shtml

Fraser

[Fraser]

SeanB,

Many thanks for the helpful comments.

Indeed, I was expecting to recoat the whole emission plate.

Note however that the damage was caused by rough, uncaring, handling during the disposal process 😔

The unit had not been well protected in transit to MoD auction, and to me. It had a badly deformed base plate and radiator fin as a result of being dropped. Something entered the emission aperture and scratched the paint off. I did complain to the seller that none of this was mentioned in the auction listing but he was an awkward person to deal with at the best of times. Nil response, as expected. He is an MoD auction buyer who cares not a jot about what he sells, its condition, or serviceability. He just covers himself with 'sold as seen, untested'. Not someone I would normally do business with but these thermal source heads are uncommon so 'beggars cannot be choosers'.

Fraser

[SeanB]

I would suggest for the stripping of the existing paint is to contact somebody local who can do CO2 bead blasting, as that will safely strip the existing coating off the unit, and not leave any residue either. I would also mask off the plate as well before taking it there, after stripping off all parts, and make sure to put sacrificial screws in all tapped holes, and mask off the entire area other than the areas to be blasted with some painters tape and card, using a firm grip tape so that the beads will not peel it, and at least 2 layers on the edge where it will be exposed to the beads, and probably not on the front face at all, but around the sides. A clean front with a coat of paint, hidden by the bezel is not going to be a concern.

I did the same for having heater plates recoated with a non stick coating by a local ship repair company years ago, helped though that the one guy was living next door to me in the flats at the time, and we got on well. Got a nice well blasted frosty plate back with a nice green non slip PTFE coat on it, emissivity definitely changes as you go from the oxidised rough matt aluminium extrusion to the PTFE, and it normally runs around 140C with 5 200W cylindrical heating elements in the body. Did have to do a lot of helicoil inserts though, but the existing holes were in any case well worn, the bead blasting with fine grit just finished off most of them rather well, but I was planning on doing the inserts anyway, had already ordered the inserts for the lot, simply as the stainless inserts survive heating cycles better.

Might take a spare one, with a PT100 pocket adaptor instead of the Fenwal thermoswitch, put in a PID controller and run it off 40VAC instead of mains ( slow down the ramp) and see about giving the flat face a coat of mystery emissivity Aerolac matt black as a poor man's black body, seeing as I do have a couple of them spare. Not going to win any beauty contests though for looks.

[Fraser]

Thanks SeanB

Beyond camera accuracy checks, a thermal reference can be a very useful tool to have around if experimenting with lenses, attenuators and windows

Fraser

[mahony]

Hi Fraser,
I am searching for high emissivity paint too. Seems like the nice Nextel 3101 c10 is available from a swiss company:
https://www.ibzagshop.ch/de_CH/e-shop/nextel-lackiersystem/lack/3101-nextel-suede-coating

Unfortunately the price tag is significant... 186 CHF per litre. 

Do you, or someone else have any experience with paint from those guys:
http://paints.labir.eu/

I think for my purpose this might be sufficient:
http://paints.labir.eu/paint/thermographic-paint-for-high-temperature-applications

Emissivity is claimed to be ~95+% except a small dip at 13µm. My current cheap rattle can flat black seems to be around 93 to 94% using a simple reflexion test with a hot soldering iron.

[Fraser]

Hi,

The 'official' high Emissivity coatings are expensive and hard to source, as you have found. The Nextel product is a surface coating rather than a paint. It is what you find in SLR bodies and is more of a flock coating. That is to say, it is furry ! This coating is VERY fragile and easily damaged if not protected from fingers or contact with other objects. Great for a lab, but not too practical elsewhere.

As detailed in this thread, I opted for the very Matt Krylon paint for my Black Bodies. Some other paints have very variable Emissivity characteristics across the LW thermal spectrum and at differing temperatures.

The paint you referenced looked interesting. It is the first I have seen that is easily available and claims to be specifically for high Emissivity thermal imaging applications. If you get that paint I would be interested to hear of your experiences with it.

Fraser

[Fraser]

I am in the process of ordering the LabIR high Emissivity paint for testing purposes.

Fraser

[mahony]

So am I.
Just finish testing initial testing with my Arduino/Peltier based cheapo-DIY-Blackbody. Temperature control and stability looks good now I need some black paint ... hopefully a bit better than some std. flat black.

[Fraser]

Nil response from the company today...... disappointing start.

Fraser

[mahony]

The same for me - maybe they are not interested in single can sales?
Do you have any reliable source in europe for some similar product? Otherwise I might just go with a can of Krylon Camouflage Ultra Flat Black - at least there is a German guy selling some of them on eBay...

[Spirit532]

It would be interesting to do a group buy of the Nextel 3101 C10 paint. The $187/l price is a bit painful, but think about it - how many liters of paint do you need to cover a dozen blackbody sources? It would probably end up costing ~$5-10 to cover a single emission plate of acceptable diameter.

However, it should be noted that, as per the datasheet, this is a 2-part coating that requires a thinner to make the right consistency, and a hardener to activate it.
The article numbers for those are "NEXTEL-Härter 6018"(hardener), mixed in 8 parts paint, 1 part hardener by weight, and the thinner is "NEXTEL-Verdünner 8061", of which an additional 10-15% are required for air spraying.

The coating itself is available here: https://www.ibzagshop.ch/de_CH/3101-nextel-suede-coating-14068-0015/7751
The hardener is also not very affordable: https://www.ibzagshop.ch/de_CH/e-shop/nextel-lackiersystem/haerter-verduenner/6018-nextel-haerter
The thinner has an acceptable price: https://www.ibzagshop.ch/de_CH/e-shop/nextel-lackiersystem/haerter-verduenner/8061-nextel-verduenner

[Fraser]

It is worth noting that I have yet to see a commercial blackbody thermal reference covered in the Nextel 3101 C10 coating. There is likely a very good reason for that as Such references are definitely not cheap so coating cost is unlikely the reason. All the reference sources I have seen or used looked to have a dark grey flat paint coating simikar in appearance to normal matte black chalkboard paint. I know it is more than likely the unobtainium 3M high emission paint coating though.

It would be possible to become the equivalent of a 'Voltnut' with high Emissivity coatings, but such is not necessary in thee real world. Even my most expensive industrial thermal cameras are still stated as +-2C accuracy under perfect conditions. The important factor in a reference plate is knowledge of its emission characteristics over a range of temperatures. The NASA testing has kindly provided that information for a number of paints. I am fortunate to own some very accurate spot temperature measurement and plotting equipment so I can check the temperature/emission characteristics of my references.

The classic accuracy check fir thermal cameras used at home is the 100C boiling water and 0C melting ice in water test. These provide some measure of confidence check without the need for expensive or exotic reference equipment. A PT100 RTD can be used as a water temperature check and the Emissivity of water is a known value.

Then there is the cheap and simple "Ambient temperature thermal reference" I previously detailed in this sub forum. With a decent PT100 RTD thermometer fitted, such would be a very useful test source. The Emissivity plate would beef to be tested though.

Fraser

[Bill W]

Some commercial sources have micro-machined surfaces, so creating an array of cavity black bodies which are then coated / painted.

The difficulty with even 2°C accuracy on a camera over temperature is how much this scales from the typical two point calibration at 20°C and 40°C black body and say in 10°C steps of camera temperature.  The accuracy at calibration has to be a lot better if you expect it to read 200°C +/-2°C with the camera at say 35°C (right between calibration points).

Cold accuracy is harder to get right, but mainly down to residual energy messing it up.

Bill

[Fraser]

Indeed Bill,
This is why I warned against the 'Voltnut" or "Timenut" mentality when it comes to thermal cameras and black body thermal sources. Thermography is a mine field of errors that could send a "Tempnut" crazy

What I had not previously realised was how variable the Emissivity of a paint can be over a pretty small temperature range. For me, it is important to have as flat a response as possible, rather than the highest possible emissivity at any particular temperature. Provided the thermal reference is of known Emissivity and temperature stability, it is good enough for most tasks.

The very best Black Body references seem to use the physics of the heated chamber/cavity to create a quality product without the need for very exotic coatings. As you say Bill, micro machined cavities seems a good way to achieve repeatable performance.

For those interested, I have several different Black Body thermal references of varying capability. I use them to carry out confidence checks on cameras when repairing or maintaining them. I have yet to have a camera that required true calibration.

I thought about posting a review of the Black Body references that I have but I realised that they are quite exotic and not easily found on the secondary market. It is therefore unlikely that fellow forum members will bump into them on eBay. A small hint to those wanting such a reference though....... think laterally.... these equipments are often not sold as Thermal Camera Black Body References... more obscure descriptions are common   Some of the units I have are actually intended for user confidence checking of equipment and are supplied by the OEM for that specific purpose. As it turns out, they are also perfect for thermal camera use The units are superbly made with precision thermal cavities and various temperature drive and stabilisation technologies. These range from integrated multiple heating elements to Peltier Modules driven by either simple temperature controlled 2 range analogue heater drive loops or complex TEC controllers. The units I purchased were worth the money paid just for the excellent thermal cavities they contained. I bought one the other day as part of a bulk purchase of medical equipment. The whole palette cost be £35 and I would have paid more than that for the Black Body thermal reference alone ! I have the bonus of some very nice medical electronics to sort through as well now

Fraser

[Fraser]

I will likely regret giving away my secrets but here are some current auctions that may interest those wanting a temperature reference......


https://www.ebay.co.uk/itm/Sherwood-Davis-Geck-First-Temp-Genius-Calibrator-3000-PC/162440799779?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160908105057%26meid%3D14f5b6dcd0a743f7ad2d9cb363000dca%26pid%3D100675%26rk%3D1%26rkt%3D15%26sd%3D162440799779%26itm%3D162440799779&_trksid=p2481888.c100675.m4236&_trkparms=pageci%3Aec7442d9-5d02-11e8-95f7-74dbd180a1a4%7Cparentrq%3A8317fc421630abc2dbd6e095fff963d3%7Ciid%3A1


https://www.ebay.co.uk/itm/ALARIS-IVAC-9000-CORE-CALIBRATOR-Thermometer-Calibrator-W-O-ADAPTER-9035/301874060006?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160908105057%26meid%3Dfcf9e1e5dc4144be900efbef522ca835%26pid%3D100675%26rk%3D3%26rkt%3D15%26sd%3D162440799779%26itm%3D301874060006&_trksid=p2481888.c100675.m4236&_trkparms=pageci%3A5323f339-5d03-11e8-a15e-74dbd180f925%7Cparentrq%3A831a9d281630aad71d55ee9cfff870a9%7Ciid%3A1

https://www.ebay.co.uk/itm/IVAC-9000-CORE-CALIBRATOR-Thermometer-Calibrator-W-O-ADAPTER-7630/291688478943?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160908105057%26meid%3Dfcf9e1e5dc4144be900efbef522ca835%26pid%3D100675%26rk%3D2%26rkt%3D15%26sd%3D162440799779%26itm%3D291688478943&_trksid=p2481888.c100675.m4236&_trkparms=pageci%3A5323f339-5d03-11e8-a15e-74dbd180f925%7Cparentrq%3A831a9d281630aad71d55ee9cfff870a9%7Ciid%3A1


See what I mean about thinking laterally .  These units do not advertise themselves as thermal imaging Black Bodies or anything like that. They are IR Thermometer calibration check and calibration units. They have to be of decent quality to be licenced for medical use. Normally very expensive but their specialist use makes them cheap on eBay. I will not detail how to modify them for a particular need, but needless to say, they have a pair of very nice thermal reference cavities in them that are easily removed and compact. Some use a heating element whilst others use the very nice TEC Peltier module method of temperature creation and stabilisation. Both have a microprocessor controlling them but this can be bypassed if needed and replaced with a custom driver. As supplied, most of these units will automatically set themselves to produce a low and high temperature for the calibration check of in ear IR thermometers. From memory, the low cavity is around 34C and the high cavity is 40C but do not quote me on that. The fact that these units auto set the cavities at the required temperature ready for a calibration check is great as no hacking is needed to get the cavities running. Not all units are this friendly though so buy at your own risk!

Fraser

[Fraser]

A couple of grotty pictures taken with my iPad. This is a Genius unit that I am in the process of modifying.

Its modification is stalled due to other projects taking priority at the moment.

The Genius unit uses a heating element to control the temperature of the cavity. The IVAC calibrator uses a pair of Peltier modules. Of the two, I prefer the IVAC 9000 calibrator as it has potential for more versatile temperature coverage due to a Peltier based design being able to create below ambient temperatures. It can also use active positive and negative thermal drive to the radiation plate for fast temperature setting and stabilisation using dual polarity PWM An arduino would be perfect for the task and only an H Bridge driver would need to be added to the BoM.

Fraser

[Fraser]

On the LabIR paint front ..... no acknowledgement of my enquiry. Nil response so this company is not interested in a review of this product on the EEVBlog forum, even though I was offering to buy it at the advertised price, I.e. not asking for a free sample.

Dead End me thinks

Fraser

[Fraser]

Inside the IVAC 9000 calibrator. Just two screws and the cover slides off

This unit uses the TEC Peltier system to set and maintain the cavity temperature

These units are often sold without their power brick. They need 5V at 4A so nothing very special or hard to source.

Rubbish photos but you can clearly see the MAX133 ADC and those eight power transistors all lined up on one side look suspiciously like a pair of H Bridge drivers. One for each cavity. A custom reference would just need an arduino based temperature controller to be grafted into the unit As supplied, the unit just boots up and configures itself for a 'dumb' calibration check using the two cavities. It reports when they are at the correct temperature. Any other functionality it offers is bespoke to the calibration of the IR Thermometer and so of no interest to us.

Fraser

[Fraser]

One of the USA based IVAC 9000 units has now been sold. $25 for a pair of quality Peltier driven thermal cavities, a decent case and a PCB containing most of the parts needed for a custom thermal reference.... cheap, darn cheap 😄

I just searched for IVAC 9000 on eBay .... there is one in the USA for $10 and another for $20

Holy cow, if you are in the USA, you can pick these up for next to nothing !

Fraser

[Fraser]

Well three IVAC units sold on eBay USA today ... I wonder what caused that ? 😄

For anyone who buys one of these units, a hint regarding powering it.

There are two input pins that require +5V to be applied to them. One is the main 5V power input and the other is of unknown identity and purpose, but it needs +5V on it to boot the Calibrator. Maybe a PSU 'OK' line like on a PC power supply ?

The power connector is a standard 5pin 180 Degree DIN type. The socket is in the 'U' configuration or some would say it is inverter to that commonly found.

From left to right looking at the power socket ......

+5V / ? PSU OK ? / 0V / Chassis / 0V

Connect the '?' Pin to +5V to boot unit. The unit will draw less than 3A at start whilst it drives the two cavities to temperature. There is a separate chassis ground pin that is not connected to 0V on the controller PCB.

Once started the unit announces its revision and flashes "not ready". After the two cavities are at the correct temperature, the display will change to indicate that it is ready for tests. The instructions are helpfully printed on the top of the unit !

Fraser

[JohnG]

Well three IVAC units sold on eBay USA today ... I wonder what caused that ? 😄

Thank you!

John

[eKretz]

Fraser, I have a question for you. When a thermal cam is calibrated, does the entire field of view need to be filled by the radiating surface?

[bap2703]

If you want the entire field of view properly calibrated, yes !
If really desperate you can probably radiometrycally calibrate only part of the sensor and extrapolate to the uncalibrated area. That should work provided that you do that on top of an existing non uniformity correction.

[eKretz]

Interesting. I'm just wondering how they go about this with something like the Flir EXX series. As I recall, the photos posted from their cal lab seemed to show black body radiators only around 4" diameter. That shouldn't fill the field of view at minimum focus distance. I was under the impression that they were calibrating the cameras with these, but perhaps they were only testing calibration?

[Fraser]

The pictures I posted are of a calibration test jig and not the factory camera calibration process. It tests accuracy and uses the centre pixels onl to measure the thermal references. They do not create the NUC tables with that jug. At the factory the cameras are two point calibrated, or more points if needed. They are positioned with their lens inside an aperture that contains a black body emitter. The emitter fills the FOV.

Fraser

[eKretz]

So does the camera need to be focused on the radiator or does that not matter?  Can it be closer than minimum focus distance if it fills the field of view? I was under the impression that the camera needed to be focused for an accurate temperature reading.

[bap2703]

As I recall, the photos posted from their cal lab seemed to show black body radiators only around 4" diameter. That shouldn't fill the field of view at minimum focus distance.

The key here is that there is no need to focus. Since you are looking at a supposedly uniform target adding any blur should'nt change anything to what you are seeing.
Hence a calibrator large enough to fill the field of view can be at any distance... So probably quite close and quite small.

I hope master Fraser agrees with that theory

Errata, beware of your optics, if it has a large range of focusing depth or even a zoom.

[eKretz]

Gotcha. Thanks for the info.

[Devov]

Let me know if you guys need any help determining thermal stability of the paints you're using.  I have a PerkinElmer TGA in our lab and I'd be happy to run some samples.

[Bill W]

The key here is that there is no need to focus. Since you are looking at a supposedly uniform target adding any blur should'nt change anything to what you are seeing.
Hence a calibrator large enough to fill the field of view can be at any distance... So probably quite close and quite small.

Correct

Bill

[ksilabs]

On black bodies -- does anybody know anything about those cheap Electro Optical B1804D and B1804D/S7 sources that are on Ebay now? What are the temperature ranges and other parameters? I don't mind making my own controllers for those but there is ABSOLUTELY NO information on them at all so I don't even know if they are worth buying even that cheap...

Electro Optical site only shows new ones, no information on anything older at all like they never existed. I don't even mention manuals; there are no even rudimentary datasheets anywhere...

I bought a cheap WS152 source just to look what it is and it turned out to be a cavity unit i.e. a high-temp one but what is its temperature range? Having it in my hands didn't make this matter any clearer

I do have a very nice IR-160 wide area (12"x12") 350 degree Celsius source from Infrared Systems Development as well as their hi-temp cavity one, IR-563 that goes up to 1050 degrees Celsius but only have one IR-301 controller and one cable for both and no manuals but those are current production units so there is at least some information on the manufacturer's website and I'll try to fish manuals and other info from them in a week or two when all other urgent tasks are done.

But there is absolutely nothing on those Electro Optical units -- zero, nada, zilch...

BTW, Flir asks $625 for Exx calibration PER LENS so it is just mere $1,875 for 3-lens calibration And they recommend to spend that pocket change on calibration each and every year...

With just IR-160/301 I can do such calibration even every week for free and I spent on those less that just one Flir factory calibration. I would like to extend my calibration capabilities into negative temperature range using some older source as a base for Peltier cooled/heated one but it would be very nice to have at least some information to find out which source to buy if at all...

[eKretz]

Are you already doing calibrations on your Exx? I and I'm sure others here would be very interested in finding out the workflow/procedure. A how-to thread would be welcome. I have done a little poking around in the service menu but don't have a black body yet, so haven't got that far into it.

[Bill W]

Those are around the 1990 era, EOI have been sold since then and certainly dropped factory support.  However a few ex-EOI local service centres might be able to help, but even there as people retire the knowledge is going with them.

Also there were generally differential target units for MDTD / MRTD rather than a flat field calibrator.

regards
Bill

[ksilabs]

I've just got my blackbodies, not setup/tested them yet. Sure I will share the entire procedure but you guys should be aware that I have a REAL extended surface blackbody, IR-160, that is specifically designed for such calibration and has 1 square foot (12" x 12") uniform radiating surface. It is really big and fills the entire FOV as it is supposed to.

Please let me 2 to 3 weeks to post a report -- there are too many more urgent task now.

Then, my camera is 64xxxxx E40 with 3.18.0 firmware so it is missing the entire service menu as well as some files from E60. I'm going to try parts from 49xxxxx firmware this weekend but not sure it will work fully although it might if done properly. Various missing E60 parts would be nice to have but not all that critical.

The highly desirable part is that 3.12.7 firmware update that's been mentioned several times and somebody even told he found it on his PC and it had all the service menu parts. It would've been absolutely best thing to have if it really does have those parts.

64xxxx E60 files would've been also nice if somebody cared to share those...

[eKretz]

Even if I don't end up being able to do it, I would still be interested in learning the procedure. And I have a set of config files for E60 but only the 49xxxxx version, don't have the 64xxxxx one.

[ksilabs]

Ah, config files are of no interest, most of the FULL configuration is already there It is used when camera is switched into service mode -- full resolution, no noise and other things... It's switched back to the original crippled configuration when it is out of service mode but nobody forbids reading that service mode config and edit your working one accordingly...

Reading strings from 2.27.16 prodapp.exe now (strings prodapp.exe | less) -- quite a fascinating reading indeed

Next step is reworking web server files to match 64xxxx Exx configuration and attempt to get a fully working service menu on my E40 with 3.18.0 firmware. But that will probably come tomorrow.

[ksilabs]

OK, it looks like prodapp.exe binary from any 64xxxx camera is required. The one I got from 2.27.16 won't start on my 3.18.0 64xxxxx camera. No error messages, just won't start.

Other binaries (e.g. i2c.exe, fvd.exe) from 2.27.26 run just fine.

It might be due to hardware type -- gethwtype.exe from 2.27.16 returns F1 while original 3.18.0 one returns F2.

[mahony]

There is a very nice description how to calibrate TEC-less and shutterless uncooled microbolometer cameras:
https://www.j-sens-sens-syst.net/5/9/2016/jsss-5-9-2016.pdf

The procedure is pretty straight forward but requires 2 blackbody sources and one climate chamber for the camera...
2 BBs - checked
camera with different lenses - checked
climate chamber - tbd

;-)

[ksilabs]

There is a very nice description how to calibrate TEC-less and shutterless uncooled microbolometer cameras:
https://www.j-sens-sens-syst.net/5/9/2016/jsss-5-9-2016.pdf

The procedure is pretty straight forward but requires 2 blackbody sources and one climate chamber for the camera...
2 BBs - checked
camera with different lenses - checked
climate chamber - tbd

;-)

There is a calibration routine in prodapp.exe so it makes life much easier. The only requirements is to get that prodapp.exe run on newer Exx F2 cameras with 3.xx.x/4.xx.x firmware

Climate chamber is nice but not absolutely required -- it is possible to do this without it. The more blackbodies the better

[ksilabs]

No luck getting 3.12.7 or so firmware or at least prodapp.exe binary from anything 3.xx.x earlier than 3.18.0. I would even pay a reasonable price for that if somebody doesn't want to give it up for free.