AC Fan Speed Controll - RC Snubber "problems"

Thread Starter

jrcs

Joined Nov 9, 2016
17
Hi all,


I'm doing a project where I must constroll the speed of an AC motor. I'm using a random phase triac driver (MOC3020) to fire a triac (BT136-600) and change the speed as I want. All ok here.

I started by using the snubber circuit recommended in the MOC3020 datasheet for inductive loads:

upload_2016-11-9_19-35-30.png


The fan I'm controlling is a 57W, 230Vac motor. The grid here is 230V, 50Hz.

I'm using an arduino to trigger the MOC and I confirmed in the oscilloscope that my zero cross detector was equivalent to a zerocross in the grid and also confirmed that the arduino was triggering the MOC in the expected time. The problem was that if I increased the trigger angle, the speed of the fan was not changing linearly; Increasing the trigger just a little bit would get my motor full speed or super slow or even stopped it.

I solve this problem by changing the capacitor from 0.1uF to 0.33uF. I read someone that did this and worked out... I did the same and worked out aswell. That was ok but WHY did this solve the problem? I can't figure it out and it is bothering me not knowing why a larger capacitor solved this. On the internet I read a lot of recommended snubber circuits and they get 33ohms and 0.01uF.. Why do I need such a large values for my RC snubber if I'm triggering such a "small" motor?

Another "problem" are the resistores (180ohms and 2.4kohms). They need to be power resistores but this isn't mentioned in any place.. I've got 5W resistores and measured the voltage across them. I get 68V across the 2.4k resistor and 20V across the 180ohm resistor.. This is a lot of heat! They heat up until about 80ºC (2.4k) and 100ºC (180 ohms). I know they are suposed to burn power and transform it into heat but.. so much? Are this temperatures safe? Can I lower them easily?

I can get the snubber circuit off by using a snubberless triac like a T405-600H and I saw this configuration:
upload_2016-11-9_20-7-41.png


I have the moc already installed so.. can I do this?
upload_2016-11-9_20-6-11.png
(I don't actually need the MOC.. I could use another optocoupler?

Thanks in advance for your help.

Regards
 

MaxHeadRoom

Joined Jul 18, 2013
28,576
Not sure why you should be drawing that much current through the resitors, certainly not normal.
I'm also assuming this is a simple shaded pole motor?
Max.
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
It's a crossflow blower like this one
I only know his power (57W) and the nominal voltage 230V, 50Hz.

upload_2016-11-9_21-13-57.png

Simple shaded pole motor? I guess it is the same as an induction motor.

This one doesn't need a capacitor do start up.

Sorry for my bad written english.. It's been awhile since my last classes.

Regards.
 

MaxHeadRoom

Joined Jul 18, 2013
28,576
The copper ring shows it to be a shaded pole (induction motor).
The 2.4k res is not normally that high.
Check out the Fairchild app note AN-3006.
Max.
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
But the MOC3020 datasheet uses a 2.4k resistor. Now that I saw the voltage across it and the heat it "blows", i'm afrait do put a lower resistor instead, like the 180ohms shown on the AN-3006 app note.



Regards
 

Alec_t

Joined Sep 17, 2013
14,263
But the MOC3020 datasheet uses a 2.4k resistor.
That is for a sensitive gate triac. The BT136 you are using is not a sensitive gate type. Its gate may need as much as 35mA (worst case, per datasheet) to trigger the triac. If that were the case, the resistor would be dissipating nearly 3W at the trigger point.
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
That is for a sensitive gate triac. The BT136 you are using is not a sensitive gate type. Its gate may need as much as 35mA (worst case, per datasheet) to trigger the triac. If that were the case, the resistor would be dissipating nearly 3W at the trigger point.
Hi,

Thank you for your alert.
I saw a typical current being 5mA and I thought it was enough to consider it a sensitive gate triac. If I change the 2.4k to a 1.2k resistor wouldn't I increase the current and increase the power dissipated by the resistors?

Another question: I'm measuring a 20 V across the 180 res wich give about 11mA into the gate. Could the 2.4k res (being a higher than recommended value for a sensitive gate) cause the problems when using a lower value cap?

Could you explain me how you get to those 3W at trigger point?

Thank you
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
Ok, so I just changed the snubber to 1.2k resistor and "0.33uF" (don't have any 0.2uF here) as advised by the 3020 datasheet for non-sensitive gate triacs. The heat comming from the resistors indeed got lower (50ºC max) but the fan got erratic behavior just like I had in the first place. Changed the cap to 0.1uF won't get better either...

Any thoughts about my inicial questions?

Regards
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
Ok, so I just changed the snubber to 1.2k resistor and "0.33uF" (don't have any 0.2uF here) as advised by the 3020 datasheet for non-sensitive gate triacs. The heat comming from the resistors indeed got lower (50ºC max) but the fan got erratic behavior just like I had in the first place. Changed the cap to 0.1uF won't get better either...

Any thoughts about my inicial questions?

Regards
(don't know how to edit my posts, sorry)

Increasing the resistor from 1.2k to 2.4k improved the results with 0.1uF but still not close to what I can get with 0.47uF.
 

Alec_t

Joined Sep 17, 2013
14,263
Sorry, I've lost track of your 'initial questions'. Which particular ones?
As for editing your posts, there's an Edit button at the bottom left of the post.
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
In my first post:
I solve this problem by changing the capacitor from 0.1uF to 0.33uF. I read someone that did this and worked out... I did the same and worked out aswell. That was ok but WHY did this solve the problem? I can't figure it out and it is bothering me not knowing why a larger capacitor solved this. On the internet I read a lot of recommended snubber circuits and they get 33ohms and 0.01uF.. Why do I need such a large values for my RC snubber if I'm triggering such a "small" motor?

Another "problem" are the resistores (180ohms and 2.4kohms). They need to be power resistores but this isn't mentioned in any place.. I've got 5W resistores and measured the voltage across them. I get 68V across the 2.4k resistor and 20V across the 180ohm resistor.. This is a lot of heat! They heat up until about 80ºC (2.4k) and 100ºC (180 ohms). I know they are suposed to burn power and transform it into heat but.. so much? Are this temperatures safe? Can I lower them easily?
And then the questions about the use of a snuberless triac like the schematics in the first post aswell.

Thanks in advance.

Regards

Just saw the edit button, thanks.

the resistor would be dissipating nearly 3W at the trigger point.
How did you get to this value?
 
Last edited:

Alec_t

Joined Sep 17, 2013
14,263
Switching off current through an inductor causes voltage spikes, which in turn can cause erratic triggering of a triac. Spike amplitude depends on the inductance value, which varies from motor to motor. A snubber damps the spikes. The higher the capacitance of the snubber's capacitor, the lower is its AC reactance, hence in general the more effective it is at damping the spikes.
Resistor peak watts =I^2 x R = 35mA x 35mA x 2.4kOhms = 2.94W.
 

Thread Starter

jrcs

Joined Nov 9, 2016
17
Hi,

I've been reading and I think I might just found the "problem". The snubber it's not supposed to "burn" so much heat.

I changed the cap to 0.1uF and the resistors to 180ohms and 2.2k. Normally this method gives-me a lot of heat and problems in the speed variation. I changed MT2 for MT1 and the problems are gone. Now I measure 7v across the 180ohm resistor and 21 across the 2.2k one. The speed change looks ok now!

Looks like I was trying to run the triac in the 4th quadrant and this requires much more current. That's maybe why I needed such a large capacitor to get speed changing linearly. Changing it to trigger in 1st and 3rd quadrant seems to patch this problem.

Could this be an explanation?

Hope you understand my post.
Regards
 
Top