---

A first good news... from the second oscilloscope screen shot, seems that the ZC signal is TC4 compatible, as shape, logic (polarity) and voltage (5V)
So apparently you can use the Orange wire as is, without any conditioning, directly connecting to the INT_PIN of the TC4.

For triacs control I need more, unfortunately... a bit of schematic of circuits around them.
Or, the circuitry related to Yellow - heater command and Brown - fan wires.

---

Thank you for digging into this! I will open up my machine and try to take some photos and get information regarding the power board in the next day or two.

---

Out of curiosity I popped open my SR800 and took a look at the power control circuit.

---

Now that is an interesting idea.

One would have to be careful, with the control lines exposed to a connector, someone could for instance short the heater command to ground without the fan on and smoke the heater. With the control logic inside the housing, there are fewer chances for mistakes like that.

Also, I would definitely have the 1k resistor in series with the TRIAC driver inside the SR800. There are two options, tap into the wiring between the CPU board and the power board with the 1k resistors between the tap and the connector, or you could tap onto the CPU board at the NPNs and use the 1k on the CPU board. Personally, if I go this route, I will use the first approach.

I don't know anything about the TC4/TC4+ firmware, but anything I use will have logic to prevent heater energization without the fan set to some minimum speed.

Ideally you would have access to a temperature sensor that is tightly thermally coupled to the heater. That way you could detect heater overtemperature conditions in case of a fan failure or airflow restriction due to clogged chaff collector and shut down the heater.

Perhaps one of these days I will open the unit back up and check out the built-in temperature sensor.

---

This looks directly compatible with OT1/2 outputs of TC4 shield, so again no need to change any bit, except the resistor, but lower value because we have 5V instead 12V.
Typical Arduino schematics are using 220-330 Ohms for this task, to ensure a minimum of 10 mA for the optocoupler to fire. With 1kO the current will be about 3 mA, not enough.
Except this, just handle the optocoupler inputs as a SSR in the TC4 schematic.

---

Why 5V? The optocoupler anode is connected to 12V on the power board. Unless I am totally missing how you are proposing making the connection.

The way I see it you have gnd and the optocoupler cathode available on the cable between the CPU board and the power board.

You would connect gnd, and OT1 / OT2 P1 to the cathode via the 1k resistor. The 5V on the TC4 would not be involved.

Unless you are trying to avoid connecting the gnds together? But in that case, you would need to gain access to the optocoupler anode and disconnect it from the 12V on the power board.

---

The TC4 software logic handles only cases where the operator reduce the fan speed under a preset value, and in this case the heater power is cut at all.
There is no code to monitor unexpected rapid rise of heat without any explanation, maybe because in this case the detection of FC event could be compromised, I guess.
Such protections can be added very easy in code, for example in my TC4ESP fork I introduced a high temperature value alarm for ET, preset at 260C.
Also an alarm based on rise more than ... twice... faster than the average rise of last minute, could be effective and do not disturb FC detection. The later was just an idea coming in my mind while typing the above.
Anyway, what I mean is that personally I don't feel the need for another separate temperature sensor and monitoring circuits.
Coffee roasting is not an unattended activity, and you should have mains circuit breaker handy for all scenarios.

---

If still interest for such protection with separate sensor circuitry, @CK did such job, check here:
https://homeroast...post_71738

---

You are right, if keeping the 12V source the resistor value of 1kO is good.
As I wrote, my image was to handle the optocoupler as a SSR and connect both pins to TC4.
Ignored the 12V supply at all when thinking about this, supposing even it can be removed at all with this mod.
Btw, what other circuits remains to be powered inside if TC4 rules them all ?

---

Fair enough! And after looking at the pictures in the thread you linked, it appears that there is a thermal fuse and a bimetallic thermostat on the heater, so unless there is no airflow at all, one of those items should protect against catastrophic damage.

---

Nothing else is required. There is a temperature sensor of some sort that measures air temperature entering the roast chamber or possibly the temperature of the metal the roast chamber sits on. I am not totally sure what it is measuring. But I don't feel is necessary if there is a probe in the roast chamber.

The power board and the CPU board are powered up if the unit is plugged in. The beauty of tapping into the optocoupler control signals is that TC4 can fire the TRIACs or the built in CPU can fire the TRIACs. So, you can use the roaster with either front panel control or computer control. No functionality is lost.

---

So if I am reading corrrectly......
On the hardware side

OT1 goes to the yellow heater command wire with a 220-330ohm resistor in series

OT2 goes to the brown fan command wire with a 220-330ohm resistor in series

Is a connection the orange zero cross wire to IO2 required? with a resistor?

Then on the artisanq firmware

CONFIG_PAC2 // phase angle control on OT1 (heater) and OT2 (fan); IO2 used to read the ZCD

---

Orange to IO2 is mandatory, not required... without ZC there is no phase angle control.

220-330ohm only if you are using tc TC4 5v power for the optocoupler anode, that actually is connected to 12V in the schematic from post #23.
If you are using only the yellow/brown wires and keep the 12v source, as in that schematic, then use 1kO.

The comment about CONFIG_PAC2 above is a bit outdated, the right and current definition is below:
https://github.co...user.h#L24

This not affect your fan control plans, the difference is about the heater control.

---

Houston, we have a problem. On the schematic, I show a ground symbol. I drew it as such because the board labels it as Gnd, but it is in fact not ground. It is line neutral. As such, we can not connect to it. It is not touch safe. The only grounded component is the metal tray the roasting chamber sits on.

We could disconnect the TRIAC driver from the 12V and use the 5V from the TC4, but we still have the issue that the ZC signal is referenced to the line neutral. Disconnecting the TRIAC driver from the 12V also means the cpu board would no longer control fan or heat.

I think I will add an extra layer of optoisolators. Two for the control and one for the zero crossing detector.

---

Wouldn't the gnd cable on the five pin connector suffice as ground or is this the ground you are referencing that is to neutral?

Also, Tom was kind enough to post some additional circuit drawings.

http://cholla.mmt...k/circuit/

---

That is correct, the gnd on the five-pin cable is actually connected to neutral of the 120V AC via a bridge rectifier. Normally neutral is within a few volts of actual ground, but you can't rely on that. And the bridge complicates it even more. There will actually be a 12Vp-p waveform there that varies between 0.7V and -12.7V. And if the house wiring is messed up .... Well let's just say it can be a lot worse.

In the stock SR800, you can't touch any of the circuitry. Everything is hidden behind plastic.

In our situation we want to connect to other electronics, including electronics that may actually be grounded.

I will check out the new information Tom Posted

---

OK I took a peek at Tom's new circuit page. There is an error in the opamp circuit for U3_B. The drawing is missing a diode between U3_A output and the parallel RC to gnd. U3_B is a peak detected version of U3_A minus one diode drop. I think it must be connected to one of the analog inputs of the PIC. It can be used to measure the line voltage. Perhaps shut things down if there is a brown out or poor wiring to the SR800.

Getting the Orange wire into a format we can use will take a little more work. but it is doable. It is essentially a scaled half wave rectified version of the input voltage. We need to drive an optocoupler with it. Probably squaring it up first. Like Tom measured on the internal version, the duty cycle won't be 50%, but we can use different delays for the positive and negative transitions.

---

Indeed... little problem... a look at the dimmer schematic will tell the right wiring of everything.
This is the most minimal and functional way to do these things working.
Then try to equate existing SR parts with the dimmer schematic, adding the missing puzzle parts.

For example the SR existing bridge can be equated to the dimmer ZC circuit bridge, just route the bridge outputs to the optocoupler diode inputs, in series with a current limiting resistor of 1 kO, and use the optoisolator transistor output to drive INT pin IO2.
No more using the orange wire at all... gone ...

Also, you should be aware that these should be two different types of opto...something.
The ZC side should be with transistor output, while the control side they are opto triacs.
Just a headsup... to not be tempted to use for ZC the same optocoupler as for control.

---

I measured the SR800 "GND" this morning and it is exactly as expected. So yes, we need more optocouplers than originally assumed.

---

Just one more

Or, simpler to do, imo... as wiring I mean... get a robotdyn dimmer and use it for ZC and fan. Keep only the heater circuit with original SR components, fed from OT1.