Not quite but it won't work though.
Of course you won't feed me the answer directly, where would the learn effect be? right?
(or at least not that of a big one)
Well since you asked to make a diagram, and looking at it, I more and more had the feeling I have drain / source wrong. Because ledsync said 10V goes to drain. And source to control.
In my wiring I have it reversed, correct?
(I suspect I did the mistake, and not ledsyn)
This circuit is safe.
The PWM should be around 500Hz.
You may not be able to achieve full speed control with PWM. Anythig below 20% may be zero and anything above 80% may be full speed
Oh wow. I stared at that graphic pretty long, did not expect that. So the german post, that used a optocoupler was right all along. Also now I understand stitechs post better. It was the diagram of the optocoupler.
Yesterday I wanted to do the optocoupler approach, not sure why I switched back to the mosfet "solution".
So the mosfet "solution" is/was wrong in general? could you explain that a bit?
The lower PWM freq. is because the 4N35 could get too hot with 1k, since it's a bit on the smaller side I learned yesterday.
Oh this is VERY important. From pc 12v case fans, I learned they are more precise to control, than with DC control.
I guess with these bigger fans it's the opposite, since they internally run on 0-10V control anyway, and NOT real pwm, as I understood.
but for me, precise control is important. Especially to go as low as possible with the RPM.
Of course in the software I need to start the fan faster, then reduce it's speed, to overcome the initial start-resistance.
What solution would be best to achive the lowest RPMs?
a 0-10V Signal?
So actually these pwm to 0-10V converter things are the best?
There is no 1K in my schematic.
What is low as possible?
It's not specified how slow the fan can go whether you use 0-10v or PWM.
They are built for that purpose.
But you need separate >12V power supply.
I meant to use 500hz instead of 1khz.
Well which technique or "solution" is normally better suited for that? If these pwm to 0-10V converters are really the "best" for that, then I buy one and wait.
But would also be nice to make something myself. I will try jim-p solution, but curious what solution is "probably" best to go as low as possible.
Well, one could built something for that purpose as well, not really an argument. But looking at it's complexity, there is much more going on than a optocoupler or a mosfet.
This is probably what you mean.
Yes.
It's not fair comparison between converter board and simple optocoupler+resistor.
Both can do what you are looking for, one might be better, another much simpler.
Stop insisting with 10V PWM and mosfets and give Jim-p circuit a try.
The 12K resistor slows down the response time of the 4N35. To speed it up you would need a smaller resistor but that would draw to much current from the 10V pin.
Well which technique or "solution" is normally better suited for that?
Depends on the fan and in your case that is quite unknown as the datasheet give no information on the relationship between speed and the control signal input. For all I know the fan may be completely off for any voltage below 1V and at full speed for any voltage above 9V.
Can't emphasize how greatful I am too you.
Not the first time you helped me. 
I finally got some sleep.
I did not answer yet, since I don't want to waste your time with 1000 more questions, I can find out elsewhere.
I will try your solution anyway, but not rush it.(Will later check the wiring diagram of the german posts(when I got access to the pics) that used optos as well.
So I will have a working solution.
But regarding the "optimal" solution: (primarily low RMP support, secondarily: Heat)
Well I told you this was the only control info from systemair,....
well....hold your horses Mr. P, you won't believe it(Or probably saw it anyway 
Of course systemair, has wiring solutions.(Me so stupido )
And Not just one...
The biggest question for me here is, if you can see any info, that helps tackle the question, how to best approach controlling that thing. I read most threads about the systemair, and the german discussion, was so far the best one. (But noone referenced the manual)
Can you now make a more sophisticated guess, what the "best" approach would be?
So much I would love to "make" something myself, can the PWM to 0-10V Converters actually give more control around the low RPMs?
MTP10/20 is their own external potentiometer:
"Control unit with a 10kOhm potentiometer for controlling devices with 10V output and 0...10 V control input and a switching contact."
page 26 onwards:
INSTALLATION__OPERATION_AND_MAINTENANCE_INSTRUCTION_K__KV__PRIO__EN_010.PDF
- Wiringdiagrams for speedcontrollers for EC motors
- Wiringdiagrams for ON/OFF controls for EC motors
- Wiringdiagrams for demandcontrol for EC motors
....
I think my last post was NOT an answer to your post. See my last post please.(cannot change it anymore)
As I said there is no way of knowing from the info in the datasheet. Even for the common 4 wire fan, there is no general specification for how the PWM relates to speed.
Well, the fan does have a tach output. You could monitor that with a scope while you vary the PWM and see what happens.
From this point on you wou have to experiment to see what works best.
1 Like
Thx for your help, will mark the solution as right, when it's done. Even though I am sure it's correct.
I had a little confusion about Drain and Source.(As you already expected from the start 
)
I thought the naming of drain and source would change with N/P-channel. It made a knot in my puny brain.
Here is the revised version by me.
A flyback Diode could be added, but likely it's not needed.
I know you are very critical towards AI, and with all due respect, in a deep conversation(which we don't have time for here) I would be able to change your opinion.
Somehow I have a good feeling, when to trust AI and when not, proven by the fact, that after ledsyn gave his solution, confirmed by two strong AIs, I hesitated.
But for learning it's actually extremely good.
I was totally lost until yesterday, but now things start to click together.
I knew what a transistor was, but all the reset, where and why was a mystery. Still is of course, but I went to AI school a bit.
- 1 Why does it work?
- 2 What would happen if you place the pull-down not between gate-source, but drain-source?
a) What happens if resistance is lower/higher than the mosfet?
b) what are ALL the single states, and how does the voltage vary, and how much? (just roughly, no actual math)
I asked ALL this question, AND answered them BEFORE AI gave the solution.
It said I got it all right, and one of it is actually called a "bleed resistor".
You see I try to do a bit of homework Mr P., and try not to bother humans too much, but one thing I would love to hear from YOU.
Havn't even asked AI yet.
If you are not already asleep from all the text. 
I think you had your good reasons NOT to answer the obvious question in the room.
Maybe it changed? 
Why is the mosfet approach not ideal/safe??
I tried to put myself in the place of the manufacturer of a fan. I know many potential customers won't need fancy features raising the sales price, so I want to offer a complete package as cheap as possible.
On the other hand, I might put in a nice feature — if it doesn't cost too much. Adding one resistor and one capacitor allows for both DC and PWM control, a feature that will do well in advertising my product, without significant rise of the production cost.
If the customer wants something fancy, it’s his job to provide the circuitry. As for post #40, there are several mistakes in one circuit.
The IRLB8721 might just work. Now you’ve got 4 components. To make the fan go slow, use analogWrite(pin, 254); and to go fast analogWrite(pin, 1);
[Edit: symbol corrected]
I think the MOSFET is drawn upside down (source should be grounded) and the body diode is shown backwards.
I think you meant his graphic, not mine.
Well 24 hours ago I would not have understood what you even mean. But today I feel cocky and try to attempt to answer that.
I can give you my full 100% guarantee, that I cannot guarantee for anything .
but:
First the diode is correctly drawn for a N-channel mosfet. both of them are correct.
Orientation of the mosfet I don't know, but since the circuit software lets me flip and rotate a mosfet, it's likely "legal".
The source is basically the ground. So not sure what you mean.
Well you have not yet posted a schematic that would work or is safe.
Why is the mosfet approach not ideal/safe??
Ask your AI friend
1 Like