Grounding scheme for DIY Soldering Station

[MooshooMatt]

Hello everyone!

I am currently working on designing my own soldering iron controller for the JBC C245 handle. One thing I am confused about, and can't seem to find too much information about, is grounding for devices like these.

My plan is to use a dual 24VAC secondary transformer to power the logic and iron. One secondary will be rectified and regulated to 3.3V for the STM32F3 microcontroller and other logic. The other secondary will just be fused and switched to the iron. However, I also need to sample the thermocouple that is built into the iron cartridge.

I know that when I actually build the thing, the AC input has Earth, neutral, and hot. My question is, what is the safest way to configure the earth/ground/chassis connections for safety and noise performance? I am assuming that you definitely connect the chassis to earth...

One secondary will be connected (switched) directly to the C245 handle. Assuming the attached schematic is correct for the connections for the soldering cartridges, I can then just connect one side of the secondary to ground and then the other to the middle contact of the cartridge to drive the heater. Then, the thermocouple is also ground referenced... Or do I connect one side of the secondary to Earth?

The other secondary is rectified and regulated to 3.3V for the microcontroller and other logic. I am simply grounding at the output of the full bridge rectifier... However, since the heater is connected via a completely separate secondary, I am not sure what implications this has for reading the thermocouple. How can I guarantee that I do not exceed the common mode range of my instrumentation amplifier?

Do I connect earth/chassis to ground via some high impedance? That is what I often see in other designs.

Sorry for so many questions. As you can see I am quite confused

Any pointers would be greatly appreciated.

[S. Petrukhin]

You can combine the negative wire of both power supplies, because you will need to make measurements on the same source that you will heat.
It is more convenient not to even make two windings, but to make a 3.3V step-down converter from 24V - this is completely normal.
The filter housing and capacitors are rigidly grounded to PE. Minus does not need to be grounded at all - it will be flying. And the sting, I believe, should not be grounded rigidly, it is better to use 1-2MOm resistor, this is quite enough to remove the charge, but you will never risk a soldering iron sting to make a short circuit to the ground (PE).

[Miti]

Check out this project:

http://dangerousprototypes.com/forum/index.php?topic=7218.0

It would show you how Sparky did it. And it also supports c245... and many others.
I built it and I’m using it with Hakko T15. This is my go to soldering station for awhile now.
Mine has a 19V DC power supply from an Intel NUC, modified for 24V.

Edit: This is what mine looks like.  https://www.eevblog.com/forum/projects/unisolder-5-2-universal-soldering-station/msg2341236/#msg2341236

[MooshooMatt]

Thanks for the replies! So it sounds like you want to connect the outer shell of the tip to earth/chassis, and then connect earth/chassis together with some high value resistor...  However, how then can we measure the thermocouple? It is now isolated from the logic ground so I would have to account for all sorts of common mode input ranges...

Unfortunately, I don't see anything in the Unisolder schematics about case grounding...

Sounds like I may have been overthinking and can just use both secondary windings in parallel...

Anyone know the C245 handle connector pinout?

[S. Petrukhin]

Thanks for the replies! So it sounds like you want to connect the outer shell of the tip to earth/chassis, and then connect earth/chassis together with some high value resistor...  However, how then can we measure the thermocouple? It is now isolated from the logic ground so I would have to account for all sorts of common mode input ranges...

Unfortunately, I don't see anything in the Unisolder schematics about case grounding...

Sounds like I may have been overthinking and can just use both secondary windings in parallel...

Anyone know the C245 handle connector pinout?

If you use a soldering iron only in production, for not yet live boards, then hard grounding will not be a hindrance at all.

But I didn't quite understand the structure of the tip in your drawing.
As far as I know, the heating element is connected in series with a thermocouple and the leads of this circuit to contacts C2, C3.
Contact C1 has only the outer shell of the tip.

You can use, for example, C2 as a common wire (minus): for heating, for a button (which will be connected to the circuit by a separate wire), for a motion sensor (which will also be connected to the circuit by its own wire), and for an external posistor (I'm not sure that it is useful).

Use C3 for heating by applying a voltage of +24V relative to C2 and measuring the voltage of the thermocouple during heating breaks relative to C2.

About C1 will be connected to the PE through a high-resistance resistor.

The device housing and the 230V input filter capacitors are also connected to the PE. Do not ground low-voltage wires to PE.

You need to understand that there are two lines that are called the ground: earth and gnd. Earth - is located in the power outlet as PE wire and is connected to the ground loop in the earth of the planet. And GND is the fictional ground of the circuit, the common wire, most often it is the power minus after the power supply.

[Miti]

Unfortunately, I don't see anything in the Unisolder schematics about case grounding...

Sounds like I may have been overthinking and can just use both secondary windings in parallel...

Anyone know the C245 handle connector pinout?

Look at the first post to find all the connection diagrams for all the irons supported by this controller, then see how they relate to the schematic. If I were you, I would just build this project and be done with it. There’s a lot of engineering that goes into such project. Why reinvent?

[MooshooMatt]

Look at the first post to find all the connection diagrams for all the irons supported by this controller, then see how they relate to the schematic. If I were you, I would just build this project and be done with it. There’s a lot of engineering that goes into such project. Why reinvent?

I thought there was nothing wrong with wanting to do something myself and learning along the way... but apparently not...

Anyways... People seem to disagree with the pinout given on the first page of the dangerous prototypes post. It hasn't been updated in three years and since then people have made the case that current pinout is actually passing current through the thermocouple when its not needed.

I think I will not go with an AC transformer, and instead get an off the shelf isolated AC-24VDC converter. It's cheaper anyways and then I don't have to worry about separating grounds...

And will probably just wait for the tip to arrive and measure for myself... If anyone knows which contacts correspond to which color wire and which pin on the connector, that information would still be very helpful!

Thanks

Matt

[Miti]

I thought there was nothing wrong with wanting to do something myself and learning along the way... but apparently not...

Nothing wrong with learning, I do that all the time and I didn't realize that's what you were doing.
I just measured my genuine Hakko FX-951 and the outer shell is solid to earth ground. IIRC Sparky mentioned in one of his posts that he tried to measure the thermocouple without disconnecting both ends but he couldn't get a meaningful reading due to noise. Your experience may be different.

[rfmerrill]

I've heard (but not experienced first hand) that some soldering irons actually have a switch that lets the iron tip be floating, hard grounded, or grounded through 1M.

[MooshooMatt]

Nothing wrong with learning, I do that all the time and I didn't realize that's what you were doing.
I just measured my genuine Hakko FX-951 and the outer shell is solid to earth ground. IIRC Sparky mentioned in one of his posts that he tried to measure the thermocouple without disconnecting both ends but he couldn't get a meaningful reading due to noise. Your experience may be different.

No worries mate

What do you mean "without disconnecting both ends?"

[Miti]

If you look at the Unisolder schematic, Q2, Q8, Q10, Q11 disconnect both ends of the heater/thermocouple from power during thermocouple measurement. Four transistors because it is a dual iron system.

[MooshooMatt]

If you look at the Unisolder schematic, Q2, Q8, Q10, Q11 disconnect both ends of the heater/thermocouple from power during thermocouple measurement. Four transistors because it is a dual iron system.

Ah... I wasn't aware that they were dual-purposing the heater-select (low-side switches) to disconnect the iron from ground during measurement. It's confusing because in their implementation, one heater must always be connected to ground via the low side switch...

I was planning on doing something similar in my implementation anyways!

However, if you completely isolate the cartridge from the power supply during temperature reading, then how do you guarantee that you do not exceed the common mode rating of the amplifier?

[Miti]

If you look at the Unisolder schematic, Q2, Q8, Q10, Q11 disconnect both ends of the heater/thermocouple from power during thermocouple measurement. Four transistors because it is a dual iron system.

Ah... I wasn't aware that they were dual-purposing the heater-select (low-side switches) to disconnect the iron from ground during measurement. It's confusing because in their implementation, one heater must always be connected to ground via the low side switch...

I was planning on doing something similar in my implementation anyways!

However, if you completely isolate the cartridge from the power supply during temperature reading, then how do you guarantee that you do not exceed the common mode rating of the amplifier?

Looking again through the schematics after many years, you are right, my memory doesn't serve me well. 
Everything is referenced to SGND when measuring the temperature so the low side Q10 (or Q11) must be on.