Beginner seeking advice on a design to make a desoldering station.

Thread Starter

The_Boots

Joined Feb 3, 2020
8
Apologies for my ignorance. I'm still in the early stages... I wanted to ask for advice on a project I was working on to make a desoldering station. I thought it would be a fun challenge to come up with something that I could use to experiment with a few components. I'm trying not to use too many specific ICs, so I tried to set up some parts to do amplification and temperature controlled switching, but I thought it would be a bad idea to just throw stuff together without at least asking for advice. I'd appreciate if anyone could look over this and tell me the obvious errors (things that will explode, things that literally won't work, etc). Pointing me in the right direction to find places to learn WHY they're obvious errors would be appreciated, too! I've never really tried to design a circuit, especially using software like this, so please be gentle!
Thanks for your time!desoldering.png
 

ronsimpson

Joined Oct 7, 2019
2,988
How fast/often will you turn on/off the Heating Element? Asking because Q2 will not do what you think it should. The ESP32 outputs 0/3.3V. Q3 pulls down well but Q2 only pulls up to 2.7 volts. Many ways to fix this but if you turn on/off once every 5 minutes or you "PWM" Q1 at 50,000 times a second require different ways to drive the Gare of Q1.

I don't know the ESP32 well. Input pins D4 and D34, are they "analog" inputs?
 

dl324

Joined Mar 30, 2015
16,845
Welcome to AAC!
Apologies for my ignorance. I'm still in the early stages...
Circuits are easier to read when you use logic symbols instead of boxes with pins in IC order.

We also prefer for the "flow" to be predominantly left to right and top to bottom, with no unnecessary wire bends, scenic routing, or extraneous connection dots. Color coded schematics also look funny. My schematic editor color codes, so I print to B&W PDF to post clips.

This is what the opamp portion of your schematic would look like when drawn more conventionally. Circuit intent is much easier to understand.
1580745301450.png
EDIT: Added missing ground connection on T-.

The circuit doesn't make sense. IC1A is an inverting amplifier that feeds a voltage follower that has it's output clamped. But you don't have a negative supply, so the outputs of the opamps can't go below ground.

You don't have a current limiting resistor for D1, but the opamp can only source about 25mA and it can survive a short indefinitely.
 
Last edited:

Thread Starter

The_Boots

Joined Feb 3, 2020
8
Welcome to AAC!

Circuits are easier to read when you use logic symbols instead of boxes with pins in IC order.

This is what the opamp portion of your schematic would look like when drawn more conventionally. Circuit intent is much easier to understand.
View attachment 198208

The circuit doesn't make sense. IC1A is an inverting amplifier that feeds a voltage follower that has it's output clamped. But you don't have a negative supply, so the outputs of the opamps can't go below ground.

You don't have a current limiting resistor for D1, but the opamp can only source about 25mA and it can survive a short indefinitely.
Thank you all for your help! I couldn't figure out a way to get EasyEDA (the software I was using to make this diagram) to put in both opamps in the package, so I picked one that at least had the correct number of pins. Do you have a suggestion for a good layout program to start with?

I think my error was mislabeling the thermocouple +/-. I also added in a 4.7 Ohm resistor to limit the diode current. I wanted to make sure not to fry the ESP with 5V, even though there appears to be some argument over whether they're 5V tolerant or not.
 

dl324

Joined Mar 30, 2015
16,845
Do you have a suggestion for a good layout program to start with?
I use an old version of Eagle. Some on this site say it's too difficult to learn; I didn't find that to be the case. I tried a couple others (ExpressPCB and another with a similar name) and found both lacking in functionality or ease of use. Once I tried Eagle, I stopped looking.
I wanted to make sure not to fry the ESP with 5V, even though there appears to be some argument over whether they're 5V tolerant or not.
According to the datasheet, the maximum input voltage is VDD+0.3V.
1580750414796.png
1580750452245.png
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
How fast/often will you turn on/off the Heating Element? Asking because Q2 will not do what you think it should. The ESP32 outputs 0/3.3V. Q3 pulls down well but Q2 only pulls up to 2.7 volts. Many ways to fix this but if you turn on/off once every 5 minutes or you "PWM" Q1 at 50,000 times a second require different ways to drive the Gare of Q1.

I don't know the ESP32 well. Input pins D4 and D34, are they "analog" inputs?
I'm shooting for PWM, but it doesn't need to be stupendously fast. Do you have any suggestions of where to look to learn more about gate driver circuits? I wasn't having a ton of luck beyond places explaining why you need them (which I get) and not what the specific variables are in designing them. I think most places I found assumed you would use an off-the-shelf IC.
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
OK. I tried to remake it in Eagle. No promises I didn't screw up a new pin, though! I put in a few more resistors and rearranged things. I'm not too worried about the microcontroller pins. I can adjust those as needed (plus, nearly all are multiplexed and can be changed in the program).
Some of the new resistors don't have values yet, but I wanted to make sure the rest of the circuit looks at least a little improved! Thanks again for all the help you've given me!
desolder.png
 

AlbertHall

Joined Jun 4, 2014
12,345
If the ESP32 is running on 3.3V then the MOSFET will not get as high a gate voltage as would be desirable. You will get about 2.7V on the MOSFET gate. The RDSon is specified at 5V on the gate so the MOSFET resistance will be higher and that means more heat.

I am not keen on the D1/R6 circuit.
I would choose circuit 3 from the link below, but you won't need the capacitor (though it will do no harm) and the resistor could be higher, say 1k, and if the ESP32 is running from 3.3V the zener should be 3.3V
http://www.thebox.myzen.co.uk/Tutorial/Protection.html
 

swr999

Joined Mar 30, 2011
38
Consider reselecting R8 to be equal to the resistance of the thermistor at your temperature setpoint. This will put the Vout from the R8-X1 voltage divider at about half scale (3.3/2) when you're at the temperature setpoint.
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
Consider reselecting R8 to be equal to the resistance of the thermistor at your temperature setpoint. This will put the Vout from the R8-X1 voltage divider at about half scale (3.3/2) when you're at the temperature setpoint.
R8 is intended to be for cold junction compensation for the thermocouple. It ideally will be around room temperature (or maybe slightly higher from the other electronics) so I figure it should sit around 100k.
Of course, I guess the actual cold junction will be somewhere inside the desoldering gun, now that I think about it. You have to measure the temp at the place where the thermistor wires transition to copper, right?
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
If the ESP32 is running on 3.3V then the MOSFET will not get as high a gate voltage as would be desirable. You will get about 2.7V on the MOSFET gate. The RDSon is specified at 5V on the gate so the MOSFET resistance will be higher and that means more heat.

I am not keen on the D1/R6 circuit.
I would choose circuit 3 from the link below, but you won't need the capacitor (though it will do no harm) and the resistor could be higher, say 1k, and if the ESP32 is running from 3.3V the zener should be 3.3V
http://www.thebox.myzen.co.uk/Tutorial/Protection.html
OK. I think I understand. It can't pull anymore than the V(be) minus a diode drop, right? So I need to make sure that I get the base up closer to 12V to make sure it's properly opening the mosfet gate.
With that in mind, I went back to looking for a driver circuit and came up with this. I threw it into a simulator just to make sure that it appeared to work. I might have the wrong resistor value in a place or two, but watching it the simulator really helped visualize what was going on.
I also tried to tweak the zener clamp to mimic the diagram you suggested. It makes sense in retrospect, so thanks for pointing me in that direction. I wanted to put something there because I figured that there was a possible chance that the voltage may go over 3.3V, but given the opamp is being powered by 5V, that doesn't seem too likely, does it? The hardest it should be able to push would only be 3.5V to begin with. If what I said isn't totally wrong, do I really need the clamp, or is it maybe still a good idea, just in case?
Thanks for helping me learn! All of your explanations and suggestions have helped immensely.
How does it look now?desolder.png
 

AlbertHall

Joined Jun 4, 2014
12,345
D1 should be from the micro pin to ground and the resistor, R6, can be much higher, 4.7k perhaps.
R10 should be between Q4 collector and ground so Q2 can pull the MOSFET gate smartly down (ot needs some base current to do that).
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
D1 should be from the micro pin to ground and the resistor, R6, can be much higher, 4.7k perhaps.
R10 should be between Q4 collector and ground so Q2 can pull the MOSFET gate smartly down (ot needs some base current to do that).
Doesn't Q4's collector connect to ground through R10 and Q5? Why is another resistor needed there? Wouldn't it just pull the voltage down when the gate is high? I suppose not enough to matter, but still.

I'm sorry if I'm asking dumb questions, I'm just trying to watch the simulator and reconcile that with what you're saying. Obviously the things in the simulator are all ideal, so it doesn't really account for everything in the real world.

I didn't have time last night, but I want to breadboard this and hook it up to my scope to get a real sense for what's going on. I only have one bench supply though, so I'm either going to have to risk exploding something, or get creative with my voltages.
I have a power supply waiting with all the voltages needed, but it has no current limiting, and I don't want to blow everything up just yet.
 

AlbertHall

Joined Jun 4, 2014
12,345
Doesn't Q4's collector connect to ground through R10 and Q5? Why is another resistor needed there? Wouldn't it just pull the voltage down when the gate is high? I suppose not enough to matter, but still.
When Q5 is on it turns Q4 on and that pulls the bases of Q2 and Q3 high.
When Q5 is off so is Q4 and R1 and R9 pull the collector of Q5 high and so R11 is pulling Q2 and Q3 bases high not low.

[Edit] I am suggesting moving R11 not adding another resistor.
 

Thread Starter

The_Boots

Joined Feb 3, 2020
8
I decided to try a simpler circuit to drive the mosfet, since I was having trouble understanding what was really happening in the more complex one. How does this one look?
driver.PNG
 
Top