Modifying only C1 can roughly adjust the temperature to a close temperature. If there is no special forced need, just modify C1.
Modifying C1 on a profile that has many coefficients (thermocouple or PTC), may produce highly unexpected results, since it is a non-linear polynomial. Just bear in mind. My intention as these coefficients to be changed and saved in windows software, but... as I said several times, I am very short of time and I don't know when this will happen.
I thought JBC use transformers with several secondaries to have a seperate low voltage rail for MCU etc.
They use it for both several voltages to the iron, and separate voltage for the MCU/etc. There are 5 - 7 secondaries on their transformers. Maybe some of them center-tapped - I don't know. This is alright if you order 100 or 1000 transformers, and sell stations with their prices, but no good for DIY.
In fact in my country I don't have a problem ordering any transofrmer (or at least 5-6 years ago), for a price 2-3 times this of decent of-the-shelf single or double secondary toroidal, but that's not the case everywhere, at least not for low enough price.
...with the right combination of heating and hard vibration...
This is not a hard vibration bua a very ligh one. And normal 120-150W transformer on this controller normally does not heat up beyond 50 degrees celsius. Cheap ones heat's up not because of the windings, but because of the core losses - they heat up even without any load on them.
Don't get me wrong - a good transformer should not emit more noise than a noise cockroach walking on a wood floor emits. Everything beyond this speaks about transformer's quality. Only, IMHO, the noise itself won't be the reason for them to fail, even when they are low quality ones. If they fail, they will fail even without load, if their quality is that low.
You will have to wait a VERY long time for varnish to fail beacuse of this. Every electric motor in the universe has MANY times more stress on the windings and varnish, and despite this they last almost forever and when they fail, they fail because of overheat/overload bad bearings or similar stuff. The controller has no way of compromising the transformer, no matter how bad quality it is. The only way is DC offset generation, but even this is hard enough even when firmware fails - is must do switching synchronous to the mains in order to be able to ram up a DC current on the transformer, and constantly switch uneven portions between positive and negative mains half-periods.
Moreover, even the highest quality dead silent transformer has some constant vibrations on it. They are not audible just because the much lower mains frequency needs considerably larger amplitude in order to be audible.
it might have something to do with DC offsets created by chopper modes (less than full power). Toroidal transformers are VERY sensitive to DC offsets as they can saturate the iron core very rapidly.
No DC offsets are created. It is switched at the same time on both positibe and negative half-period. It is not switched at the zero cross, so there is a sudden current bump at the moment of switching, then some voltage ringing because of it. This creates the noise, and every transofmer will have noise, more or less. The better the windings are wound and impregnated/isolated, the less the noise will be. There's some noise even from the core of some (most of them non-toroidal) transformers.
Even most DC suppies are producing some noise on sudden load changes.
This can be avoided with some snubbers and fine tuning of the switching speed, but the lower the switching speed, the more suffering from the MOSFETs there will be. And snubbers for 50Hz are HUGE. And all these optimisations are strongly transformer-dependant, so they even can make things worse on different transformer than the one they are made for.
JBC are using transformers with several secondaries to avoid this. But transformer with more than 2 secondaries is no good for DIY project, IMHO. The price will be higher also, because solid state relays will be needed fo smooth switching.
It's exactly at less than full power that the issue with noise from my toroidal transformer occurs. I might have gotten a dud then.
There will be noise from ANY transformer, even if epoxy impregnated. In fact, because the toroidal cannot be wound with windings toghtly spaced to each other, tha toroidal may even produce more audible noise. I am talking about electrical noises/peaks, which are mainly a result of leakage inductance. And the toroidals excell in this respect.
My toroidals have some sudible noise too, but it is low enough not to bother me in any way, despite the fact that they are not impregnated in any way (or at least it seems so). You won't be able to defeat this noise, no matter what you do. And even if you defeat it on the transformer, the electronics will have some noise too, and even some soldering tips have some audible noise, when the power not is switched at exact zero cross point, and this is exactly what happens on 1/2 power and below.
I'm not worried about using block transformers at all. JBC themselves are using them just fine.
As I said several times already, the controller works with pretty nasty/low quality transformers. I do have one that has pretty large leakage inductance, producing pretty nasty inductive peaks. But the controller works with it wothout a problem. I spent pretty big amount of time tunning the firmware to withstand this without problem. I also tested it chinese with non-toroidal power power transformer - nothing dramatic either, more than
But still, IMHO, a toroidal transformer with decent quality is the best power source for this project. One look at the oscillograms around power switching, especially on 1/2, 1/4 and 1/8 power tells pretty much everything about this.
Moth my controllers at the moment are from tindie, and the last 5.2B was also from tindie. No problems so far. The only thing is the cheaper (not Microchip) opamps tend to give a bit different offsets form channel to channel, but nothing too dramatic.
For example, I never had an issue with the darlington transistor, both on mine and on tindie boards. And before 5.2 there were 2 versions already with the same darlington. And I definitely NEVER had an issue with the MOSFETs, straight form v0.1 - this is maybe 10 boards made from me alone, and at least 5 from tindie. Honestly, I don't know where these issues come from - maybe either bad soldering or bad parts, but the soldering on my tindie boards looks great so far.
The profile parameter for this is c1. c0 is the resistance at 0 degrees celsius, c1 is degrees per ohm. It is floating point so can be tunet very precisely - you/ve got to lower it a bit.
T= c0 + R*c1 + R*R*c2 + R*R*R*c3 and so on.
About sensitivity - it depends on the iron. On very small irons sometimes a small change makes big difference.
I thing you must play with OVSGain (if I remember well). This comes from "Overshoot protection gain" and "DGain", which is the "prediction" parameter of the PID.
The most basic PID algorithm guide is - slow iron = slow PID (which is mainly low "I" and "D" coefficients). The rest is hous of playing with all parameters and a cup of water/isopropyl alcohol. Slow iron means more hours, because everything happens slowly.
Just don't look at wire colors, because they might be different than on WSP80. Sensor positive and sensor negative are interchangeable, since this is resistive sensor, same goes for heater negative and positive, since these are izolated from enything else.
However I tend to connect the brighter/warmer colors as positive, and the darker/colder colors as negative.
So, from what I see on the picture, even the wire colors are the same as on WSP80, so there's no questions about the connection, so just connecti it this way.
You may also look what's inside the handle, because some (Weller) models have electronics there.
I do have a stron desire to write some kind of manual for connections and creating profiles, but I am very short of time these days, and this tend to be pretty time consuming.
Then, at 0 degrees celsius it should be around 20 ohms. Then c0 should be -20.
c1 = 1/0.077 = 12.98, make it 13.0
The currentA shoud be around 1.8mA, then CBandA shoud be 0 instead of 1, and CurrentA should be 37.
Offset is a bit harder. 1.8mA * 60ohms * 27 = 2.916V after the first amplifier, which means the second one's gain should be around 1, which in terms of iron profile is 256/27 = 9.218 = 9.0, which is too low.
Now instead of using 37 for the current, we use 24, and the current becomes 24*(1.225/100/256) = 1.148mA
Then 60*1.148ma*27.77 = 1,913V after the first amplifier, but this includes also the offset from the 20ohms at 0 degrees. If the offset was not there, the foltage is 40*1.148mA*27.77 = 1,275V. These 1,275V represents 500 degrees at the sensor, and shoud be amplified to 3.0V by the second amplifier, so the gain should be 3/1,275 = 2,353, which in terms of iron profile gain is 2.353/(27/256)=22.
Now the offset in the profile shoule be 20 ohms * 0.001148 * 27.77 * 27*(22/256) * 1024/3 = 505.12 = 505.