Conformal coating for REF volatge

[mikeblankenship]

I am currently working on a project where the Voltage Reference is very critical over Temperature.
The Vref feeds into a 20Bit DAC. I am building a programmable voltage source partially based on a design from Ian Johnston

I am using a MAX6325 2.5V reference with the TRIM option. I am able to trim it to within 50uV
of 2.5V (2.500 048). From 10C to 60C the output for the 1st 2 weeks was +/- 4mV of this value
going in 5.0C increments and letting settle for 30 minutes each step ( I made about 4 complete runs within 2 weeks).

I went on a 1 week vacation .... I came back and I set my poor-mans Thermal chamber up to do another run
( I purchased a MyTemp incubator, Dave modified one of these in one of his videos) and my readings now
are between 2.500 125 and 2.500 145 for the last week. It still stays within ~ 5uV over temp when it settles in
at the 2.500 125 or the 2.500 145 level. The TRIM option I am using is a 1024 step DIGI pot, I have the range
set in EEPROM and have not changed it for the last week.

I have heard from this Blog and others that conformal coating is good for moisture and keeping dirt and grime from the contacts. But
how about the thermal characteristics for using it? Has anyone tried using C-Coating to try and minimize thermal drift with a reference?

I have just purchased ( 1 week wait time here in California) both Silicon and Acrylic C-Cotaing in a spray can to experiment with.

https://www.amazon.com/gp/product/B008O9YGQI/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1
https://www.amazon.com/gp/product/B008OA7178/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1
https://www.amazon.com/gp/product/B008OA77NG/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1



Thanks

mike

[Tomorokoshi]

I have a few questions about the overall architecture of the system:

1. Are there significant sources of heat on the board anywhere near the reference?

2. Is there a heat sink around the reference?

3. Is the source of temperature change from environmental conditions, or from circuits being powered on and off?

4. Is the concern about conformal coating due to the thermal resistance of the thin layer, reduction in convection, etc.?

5. What thermal resistance specifications are listed for the package of the reference you are using?

6. Are there thermal resistance specifications listed for the various conformal coating options?

[mikeblankenship]

1. Are there significant sources of heat on the board anywhere near the reference?
    I have a cover over the PCB when testing is done. Refer to the attached picture of the layout, the MAX6325
    is "U4".
2. Is there a heat sink around the reference?
    No there in not a heat sink around the ref.
3. Is the source of temperature change from environmental conditions, or from circuits being powered on and off?
    The source of Temperature chnage is from me adjusting the front panle of the Thermal chamber. I do hear circuitry
    from the Peltier element going ON/OFF. I can get within .5 C of set value but this value will drift about +/- .5C
4. Is the concern about conformal coating due to the thermal resistance of the thin layer, reduction in convection, etc.?
    As I am somewhat new to precision References, my train of thought was to try and prevent "thermal air drift" from
    affecting the reference. I hear fans go on/off in my chamber
5. What thermal resistance specifications are listed for the package of the reference you are using?
    MAX6325(CSA) ppm/C = MAX 1PPM
    Long term stability = 30 ppm/100 hours
    I do not see a Thermal resistance specification, what would be another name for this?
6. Are there thermal resistance specifications listed for the various conformal coating options?
    I will look into this

Thanks

mike


     

[IanJ]

Hi,

I did consider coating the pcb's in my voltage source but in the end didn't want to take the chance of any negative effects springing into life. I think on the offset it may seem like a good idea, but for me there are questions I never really answered (never got around to it) in respect to vref's and associated circuits. I've used conformal coatings lots in the past for pcb's going offshore, but never on a vref circuit.

With a vref/DAC stability down in the uV's:-

- How permeable is the coating, i.e. it won't be zero so there's a change it could absorb moisture to some degree.

- What, if any, mechanical stresses might it impose on the pcb. As the temperature varies will the coating pull/push the pcb and affect the vref?

- Any contamination like finger print moisture will be locked in making it hard to rectify later.

Good info here:-
http://www.electrolube.com/technical-articles/conformal-coating-of-printed-circuit-boards/

PS. I did use the MAX6350 in my early prototype (that Dave got in his mailbag), but when I went into production I changed to a LM399AH.......much more stable in respect to temco.

Have fun!

Ian.

[mikeblankenship]

Hi Ian,
Thanks for your reply, and all your videos on your voltage source.

Here is some initial data I have been keeping :
https://docs.google.com/spreadsheets/d/192j8yuMogPcjxVxv44RQzKNSMyxzqDqMaO5j2jcY-m0/edit?usp=sharing

The most current data in the STEPS tab starting at line 169. I have it connected to a 1024 step digital POT so all the other data is when
I was changing the TRIM values.

If I trim the Vref to within ~ 25uV of the set value(2.500 025) I was not able to reach the full scale 10V (At 50C). So I added ~ another 100uV.

I just sent out 12 PCB's to OSH park so I can do some formal coating testing on them in my quasi thermal chamber.
I am modelling a 3D printed enclosure now ~ 1 inch square to house the board.

Let me know if your interested in playing around with it when I get them back.
I have some extra MAX6325 and MAX6350 I can send also.

If you ever decide to use a 20 Bit DAC I have some arduino code (DAC1220) that has been working pretty good,
https://github.com/GarageProto
Im working on the Internal CAL function in the code now, I can post if your interested.

Thanks again for your videos and the inspiration for this hair pulling adventure!

mike

[Kleinstein]

The conformal coating makes later repairs and tests rather difficult. Usually such a coating in needed in really dirty environment only.

So I would prefer a closed case to keep out dirt, even if not hermetically sealed. For humidity it helps a lot if the circuit runs at slightly above environment temperature. This reduces the relative humidity be a certain factor (e.g. about 2 at 10 K). This way one can avoid the critical range of about more than 50%RH where a surface water film is formed.

[alex-sh]

@mikeblankenship

Hi. Did you manage to complete this project please? What was your conclusion on coating?