+12V not fully operational

[neilh10]

@s-hicks2 made a comment "issue related to the current-limiting circuitry"
for "the wiper on the Yosemitech Y511A sometimes causes issues if you’re powering the sensor directly from the Mayfly 1.0 revA3 (current release) onboard 12v boost regulator. I’ve come up with a workaround, but will have to apply it to the next manufacturing run of boards, and also make a couple minor changes to the Modbus adapter half-shield design"

https://www.envirodiy.org/new-mayfly-data-logger-and-accessories/#comment-492

This is to track what is the operational limits of +12V Mayfly 1.0 revA3
Its very typical that high risk new circuits may not reach full design criterar, but they should be characterized so that purchasers can make decisions.

[neilh10]

For the Mayfly 1.0 revA3, the +12V boost is implemented with a nice chip called MAX17250ATD+.
https://datasheets.maximintegrated.com/en/ds/MAX17250.pdf
This uses a 2.2uH inductor for the boost, and requires a specifically designed shielded inductor and capacitor selection .
The Ipeak of the inductor is selected at 3.5A with ISET=V. Other options are 2.7A or 1.85A. The inductors need to be designed for switching and the fast changing magnetic field, that contains the energy switching from V_BATT to +12V.
It appears that the inductor may be Murata DFE201610E-2R2M=P2
https://search.murata.co.jp/Ceramy/image/img/P02/J(E)TE243A-0001_DFE201610E_reference.pdf
https://www.murata.com/en-us/products/productdetail?partno=DFE201610E-2R2M%23
which doesn’t identify the construction.
https://www.digikey.com/en/products/detail/murata-electronics/DFE201610E-2R2M-P2/7430537
Identifies it as shielded and current saturation is 2.4A, rating it at 1.7A

For the Mayfly 1.0 revA3 cutting the short trace ISET from U10/Max17250 pin 5 to C40, would lower the Iswitch to 1.85A.

Then power switched at V_BATT of 3.6V*1.85/2 = 3.24W, or at +12V 0.27A

For the +5V boost that uses the same inductor with Max17225 that has a peak inductor current of 1A. The 3.3V_SW is derived from an XC6220B331/1A LDO.

There may be less risk and more manageable guaranteed uniformity in production in selecting a part that is specifically designed for switched power supplies,

[s-hicks2]

We've found that the MAX17250 doesn't work properly with half of the dozen Y511A wipered turbidity sensors we tried, and it seems like the serial number of the sensors indicates whether or not it will work, so I'm guessing there's something slightly different between the two batches of sensors. Three other Mayfly v1.0 revision boards were tested earlier this year, each with different 12v boost circuitry, but on each subsequent production run, semiconductor shortage issues cause us to have to redesign significant portions of the board each time to use parts that were currently available. The MAX17250 is no longer available now either, so the next production run will use a different circuit for the 12v boost portion.

[aufdenkampe]

@s-hicks2, thanks for adding the capabilities to run YosemiTech sensors from the new Mayfly v1 with your new Modbus Wing!

@neilh10, thanks for bringing up this issue!

It is really important to the EnviroDIY community that already uses YosemiTech sensors that ALL of these sensors from all historic lots can run flawlessly on the new Mayfly + Modbus wing.

After quite a few tests with a wide range of YosemiTech sensors, we discovered that there are indeed variations among lots of sensors, and that to get them to all work consistently at their full voltage range we needed at least:

• CAP ALUM 220UF 20% 10V RADIAL, for the lowest power demand sensors.
• CAP 470 UF 20% 10 V, which was the minimum size that would power a YosemiTech Y511 Turbidity sensor with a brush.
• CAP ALUM 680UF 20% 6.3V T/H, which was barely the minimum size that would power a YosemiTech Y4000 Sonde with a brush. We usually used two of these, or one of these plus a 470 uF capacitor.

@neilh10 used an even larger capacitor on his knh002-MayflyWingShield.

For more info, see:

• https://github.com/EnviroDIY/YosemitechModbus#power-supply
• https://github.com/EnviroDIY/Mayfly-Modbus-Wing/tree/master/Modbus-Mayfly_WingShield#modbus-mafly-wingshield

[neilh10]

@s-hicks2 brings up the challenge of 2021/22 - supply shortages on devices, so annoying.

Seems to me though that what is needed, is to define the minimum power rating to support, and then design to that as a guaranteed power requirement for all future revisions, whatever the circuit.

From a spec sheet https://www.isweek.com/product/online-self-cleaning-turbidity-sensor-y511-a_2460.html I read it as saying that that it should take less than 50mA at 12V.
So adding a comfortable margin it could that there is requirement of minimum power of 2.4W or 200mA @ 12V - this is just a guess right now, based on what should be easily achievable. Update: from characterizations specs suggests it needs 290mA at 12V - https://github.com/EnviroDIY/YosemitechModbus#power-supply.

Testing for a minimum power is about having some variable power resistors available, putting them on as a load, and then checking when the output voltage starts dropping. Will post some pics when I do it next.

The power boost circuitry then needs to support this at the lowest LiIon voltage, which is when the power boost current is being drawn from the battery - and could be 3.4V but say for design safety would be 3.0V - so 2.4W is then 0.8A@3.0V.
The boost circuitry works by switching on/off the current through the inductor, so typically that would mean the peak current is at least 1.6A ~ 2*0.8A @ 3.0V. - so the inductor needs to not saturate on the max peak current

A possible article (though older 2003) describes the challenge and complexity of switching https://www.powerelectronics.com/content/article/21861300/guide-to-selecting-inductors-for-switching-regulators
The MAX17250ATD is really nice device. However seems to be missing the special design issues for selecting the Inductor, Other IC designs have more information.
If the Murata DFE201610E-2R2M=P2 is the inductor, it doesn't give a lot of information about the way its built and shielded.
It would be nice if the BOM used to build the rev1.0A3 could be posted so part selection is clear.
So I wonder if reducing the peak current has been tried - cutting the short trace ISET from U10/Max17250 pin 5 to C40, would lower the Iswitch to 1.85A.

A source for selecting other ICs might be
https://www.ti.com/power-management/non-isolated-dc-dc-switching-regulators/step-up-boost/boost-converters-integrated-switch/products.html
Choosing one of them - https://www.ti.com/lit/ds/symlink/tlv61048.pdf it has more information about inductors, and over specifying the inductors current rating.
It suggests an inductor family XAL4030 of which there is a XAL4020-222ME but it is overacted for this at 5A peak, and possibly a bit costly. Available https://www.mouser.com/ProductDetail/Coilcraft/XAL4020-222MEC?qs=%2Fha2pyFaduiAj9SbPgJuHx9wdZ8mU14d%2F9NDuD%252B8vf0eCL%252BjV9OT8g%3D%3D

Depending on the circuit, it may not be needed to have a large capacitor on the output. I was reacting to the need for a capacitor on that circuit, and once I had something working and tested I did choose a LOW ESR Electrolytic, based on the peak current flow into it, and then selected on cost, which resulted in the large value uF that fell out.

So just some quick thoughts on the scope of the challenge.

[neilh10]

Using Mayfly Sn21260 + Flat 2.5Ahr battery.
Some characterization readings, using simple blinkPower that turns on power 10sec and off 2sec,
with a Resistive load that consists of variable 75W/Power Resistor+ increments of 20Ohm/5W
Using Seeed D10-11 – Setting Jp2 to +5V
75+20ohms=5.010V ~ 52mA
~40+20ohms=4.990 ~ 83mA
~9+20ohms=4.900V ~ 169mA

Using Seeed D10-11 – Setting Jp2 to +12V
With no load, gets to +12V boost OK
Attempting +12V with a load of 40+75ohms, only gets to 6.9V and sometime lower and unstable
75+40ohms 6.9V ~ 60mA
75+60ohms 7.1V ~53mA
75+80ohms 7.9V ~ 51mA
75+100ohms=8.3V ~ 69mA
(only have 5 x 20ohm R, so only went to 75+100ohms.)