AVR Freaks

AVR Microcontrollers, megaAVR and tinyAVR

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

Posted by Lennart: Sun. Mar 4, 2007 - 09:08 PM

I need advice concerning microprocessor control of a Triac.
My goal is to control a 1-phase motor (24V AC/50 Hz). The motor is used to control a mechanical 3-way shunt in an heating system.
Water temperature for radiators are regulated by closing or opening the shunt. The motor is now controlled by an analog circuit not built by me. The motor is controlled by two Triacs (TIC206D). According to datasheet they need 5 mA to turn on.

I have already built a MCU-controlled system including a number of termistors that control ON/OFF for a circulation-pump and display temperatures on LCD.
I am pretty experienced in programming MCU:s but have no experience around Triacs.
I have tried to get information on the web about Triacs, but still feel uncertain.

So far I have written software that detect zero-crossing and trigger an external interrupt on falling edge. This interrupt start a timer that will interrupt and restart every 0.2 ms. A full wave is 20 ms since my mains is 50 Hz. So the timer do 99 cycles and then a new external interrupt restart program flow. On every timer interrupt I can control the state of the pin that trigger the Triac. I have looked at the signal on my scope and it is very well syncronised with the AC-wave.So far so good.... 8)

The motor is supposed to run at full speed for one second and then halt for 20-30 seconds. The direction of the motor is controlled by a termistor that measure water-temperature. MCU compute if temperature need to be higher or lower.

Finally arrived to my questions…. :P
Where in the AC-wave should the Triac be triggered.?
Where in the AC-wave should triggering stop?
Should there be one or several trigger pulses during the AC-wave?
Since my Triac only need 5 mA to turn on, can it be connected directly to the MCU-trigger pin with only a current-limiting resistor in between?
Any other important things to consider?
I only have one motor, specially made to fit with the mechanical shunt. It is rather expensive so I’d rather not destroy it by missing some basic rules when designing
AC motor-control with Triacs.
Grateful for any advice….

Richard 8)

Tags:

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

Posted by Kartman: Sun. Mar 4, 2007 - 11:59 PM

Finally arrived to my questions….
Where in the AC-wave should the Triac be triggered.?

Just past zero crossing. The triac needs a certain current flowing through it for it to latch on. At zero crossing, it is most likely this current is zero.

Where in the AC-wave should triggering stop?
Should there be one or several trigger pulses during the AC-wave?
Usually one would keep the trigger to the triac the whole cycle. Technically, you only need to give the triac one pulse and assuming the current through the device exceeds the holding current, the triac will stay triggered for the half cycle. In many cases, one would use an optocoupled triac with zero cross detection in it and avoid having your cpu worry about it. If you were doing phase control, yes, the cpu would need the zero cross reference. For on/off - let hardware take care of it.

Since my Triac only need 5 mA to turn on, can it be connected directly to the MCU-trigger pin with only a current-limiting resistor in between?

Yes. But be careful to read the specs - the trigger current varies with temperature and maybe load. If in doubt, use a transistor.

Any other important things to consider?

Current rating of the triac! Is your motor 24VAC or 240VAC? Be aware that triacs frequently fail in 'half wave' mode - only half the triac works and the other half is turned on giving you a half wave. Motors don't like this and tend to go up in smoke. Triacs also have a voltage drop of around 1-2V - at 24VAC this may be an issue, at 240vac it is insignificant.
At a pinch , I'd suggest you use relays - they're more robust or use commercially available solid state relays and avoid the whole zero crossing issue.

I only have one motor, specially made to fit with the mechanical shunt. It is rather expensive so I’d rather not destroy it by missing some basic rules when designing
AC motor-control with Triacs.
Grateful for any advice….

Level: Raving Lunatic

Joined: Thu. Sep 25, 2003

Posts: 2026 View posts

Location: Los Angeles, USA

Posted by mneary: Mon. Mar 5, 2007 - 12:30 AM

Lennart wrote:

Hi

Finally arrived to my questions…. :P
Where in the AC-wave should the Triac be triggered.?
Where in the AC-wave should triggering stop?
Should there be one or several trigger pulses during the AC-wave?
Since my Triac only need 5 mA to turn on, can it be connected directly to the MCU-trigger pin with only a current-limiting resistor in between?
Any other important things to consider?
I only have one motor, specially made to fit with the mechanical shunt. It is rather expensive so I’d rather not destroy it by missing some basic rules when designing
AC motor-control with Triacs.
Grateful for any advice….

Richard 8)

For a motor (inductive) load, the triac should be triggered somewhere near the peak of the wave. Triggering at zero-cross starts the motor with a DC bias that diminishes but never goes away. (Triggering at zero cross minimizes noise, but that's IMO less important.) Your 5 mA trigger is for only three quadrants. If your AC is negative and your trigger is positive, you need 10 mA.

You need a snubber circuit to provide the 30 mA turn-on ("latching") current to the triac. This also gives the motor current time to rise to the "holding" level of 15 mA. The triac gate drive should be maintained as long as you need to have the motor turned on.

You can turn the gate drive off almost anytime you want during the final cycle, since the triac will stay triggered until the next zero (current) crossing. Again, a good snubber is required because the zero (current) crossing may be at a point where the voltage is high. Snubber prevents a dV/dT re-trigger. The TIC206D has a critical dV/dT of only 1V/uS, so the snubber is NOT optional.

The gate probably shouldn't be connected directly to an MCU pin. When it triggers, some of the load current flows out of the gate. Even with a current limit resistor, this is potentially harmful to the MCU. This current never appears on a data sheet, but trust me it will bite you.

Quote:

I only have one motor,

Test with a different motor of similar size.

Last Edited: Mon. Mar 5, 2007 - 12:31 AM

GBaars

Level: Hangaround

Joined: Fri. Jul 23, 2004

Posts: 143 View posts

Posted by GBaars: Mon. Mar 5, 2007 - 12:30 AM

Simplest is chosing optotriacs here.
These can be switched on directly
by an IO port and some trigger on
zero-cross only by their selves.

Level: Raving Lunatic

Joined: Thu. Sep 25, 2003

Posts: 2026 View posts

Location: Los Angeles, USA

Posted by mneary: Mon. Mar 5, 2007 - 12:48 AM

A couple more comments. If your motor is 240VAC (339V peak), then the TIC240D at only 400V is a poor choice. This is especially true if the motor is capacitor-start and you're reversing it by energizing alternate ends of the start capacitor.

You can eliminate most of these problems by using a solid state relay. A 240V version of Magnecraft & Struthers-Dunn# 70S2-04-B-03-V is a good cheap example.

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

Posted by Lennart: Mon. Mar 5, 2007 - 01:29 AM

Thank you for your answers. Since I never tried to control an AC motor I was very uncertain...
The motor is 24VAC, not 240VAC. Specifications says it will draw 150 mA at full torque. I will go for using a relay as suggested unless someone else have serious opinions how this could harm the motor. I wasn't sure that relays was possible for inductive AC-loads, the electronic box that came with the motor used TIC206D Triacs so my plan was I would be safe to use same ones but control them with an MCU. But as far as I understand from the answers I get I can use relays instead. The motor is encapsuled in a box but I think I can spot one end of a large capacitor in a drawing that explain how to change the motors moving angle from 90 to 180 degrees...
Anyway, I have learned a hell lot more about Triacs than before and You kindly filled in some blanks.
Who knows, maybe I need to control the speed in some other project and now I feel more confident what's involved.

Richard

Level: Raving Lunatic

Joined: Thu. Sep 25, 2003

Posts: 2026 View posts

Location: Los Angeles, USA

Posted by mneary: Mon. Mar 5, 2007 - 03:06 AM

I agree that the relay is best for your situation. 150 mA is barely enough to hold that triac on at the edges of the cycle, much less trigger it.

A couple of opto-triacx from the MOC30xx family would also have been a reasonable choice, since you're only switching a few watts.

Be sure to put some arc suppression across the contacts. For a 150 mA motor, I would start with 470 ohms in series with 0.1 micro farad (100V or more, non polar).

pdvorak

Level: Wannabe

Joined: Thu. May 10, 2001

Posts: 69 View posts

Posted by pdvorak: Mon. Mar 5, 2007 - 03:22 AM

Lennart

For 150mA AC load I would use AC rated optomos SSR and drive it with 5mA DC.

Peter

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

Posted by Kartman: Mon. Mar 5, 2007 - 07:47 AM

Most AC motors don't like to be speed controlled other than altering the frequency - and that's another ball game.

Using a standard relay to switch the motor is a no-brainer and is economical. The failure modes of relays are quite simple - usually open or closed! They also take a bit of punishment. Add the snubber/arc supression across the contacts like mneary suggests - it should have a long a fruitful life.

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#10

Posted by Lennart: Mon. Mar 5, 2007 - 11:19 AM

Just a little remark around using triacs to control this motor. The motor and the electronic box controlling it was bought by me in 1991. During the cold part of the year which is almost six months since I live in Sweden, the motor have been triggered by TIC206D triacs appx 500 times in each direction 180 days for 16 years.

The control box is still sold by the company who made it. My conclusion is that it can't be terribly wrong to use these triacs for this purpose. There are a transistor sitting close to each triac and a couple of resistors and caps are also on the board. The design seems rather simple and elegant.

The only reason I want to swap it for a circuit built by me is that I have substituted all other parts of this control box, i.e. a mechanical 7-day clock with my own MCU-controlled circuit and added features like LCD displaying temperatures and on/off for a circulation pump.

Also because I love pöaying around with AVR:s I started this project just for the fun of taking control of the electronics in my house.

Since I'm no good at analog electronics I like to ask about the resistor and cap that was suggested to be used in series with the relay. Can I use smd parts where resistor is 1/4 Watt or would it blow up?

markus_b

Level: Posting Freak

Joined: Tue. Mar 20, 2001

Posts: 1510 View posts

Location: Switzerland

#11

Posted by markus_b: Mon. Mar 5, 2007 - 11:32 AM

Lennard,

you may want to sit down with a pen and pencil and draw the schematics of the existing circuit board. It might appeal to you most to recreate the circuit with the same or similar components except the interface to the AVR instead.

I don't really understand why you want to substitute this part of the electronics as you seem to have already replaced the mechanical clock with an AVR.

Markus

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#12

Posted by Lennart: Mon. Mar 5, 2007 - 01:38 PM

Hi Marcus

My original plan was to do what you suggest.
To figure out the circuit by drawing it on a paper and then use only the part that needs to go with the MCU trigger signal.

The circuit involves five analog potentiometers and two termistors to set and measure desired water temperature. The sum of these components are fed to an 324N op-amp that trigger the triac part of the circuit.

Since all potentiometers would be replaced with a possibility to set and measure these values from MCU, LCD and a rotary switch I started to look for application notes on how to trigger triacs from an microcontroller. All were related to how to control speed of a motor or brightness of a lamp, so I started this thread to find out how to safely trigger an AC motor at full speed for a short while from an MCU.

Since the system is working there is no desperate need to replace it, just a wish make a smarter and smaller design. Most of similar controller boxes on the market today are controlled by a microcontroller and since I've played with AVR:s for many years I wanted to use my knowledge to make a more up to date version of the controlling box. Also I would feel confident that I never would need to send it away for repair if something fails, since I would be familiar with every part of the circuit...

Regards,
Richard

gintaras_bar

Level: Wannabe

Joined: Mon. May 28, 2001

Posts: 84 View posts

Location: Lithuania

#13

Posted by gintaras_bar: Mon. Mar 5, 2007 - 05:39 PM

Useful application notes you can find here:
http://www.littelfuse.com/cgi-bi...

Especially AN1007.pdf

Regards,
Gintaras

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#14

Posted by Kartman: Mon. Mar 5, 2007 - 11:20 PM

Like most things in life, its the small details that make the big difference. I suggested a relay, as it is a simple solution versus using a triac, as the triac requires a bit more knowlege to apply it correctly. Obviously the triac works in the application but there is probably a few small things in the original circuit that makes it so. Reverse engineering the original circuit is always good for an education and finding out how and why things work. You can apply this knowlege for your modifications.

You can use surface mount components - I think a 1206 resistor is rated at 0.25w.

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#15

Posted by Lennart: Tue. Mar 6, 2007 - 12:18 AM

Hello again Kartman

I agree to 100%...
A great reason why I want to do my own version of the control box is for the fun of getting more knowledge about a new side of electronics I've never tried to conquer before.

My worry is that such education often involves at least ONE catastrophic mistake from where you learn a lot....
Since the motor involved is rather expensive and also keep me warm this season of the year I am a bit scared to be educated...

Still feels like a bit of a defeat to use relays and not get educated.
Would it be possible to use something else as a load while I test the circuit? Would the triacs be destroyed or just don't work if I send AC through them without a load and look at the signal in a scope?
I don't have to many 24VAC motors lying around back home...

Richard

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#16

Posted by Kartman: Tue. Mar 6, 2007 - 03:23 AM

Richard, you can use a lightbulb (24V of course!) as a load but it is not the same as the motor!

You might find with your investigations into the original circuit that they simply didn't worry about zero crossing detection - they probably just turn the triac on and off as they need to. The problem with triac circuits is that there is lots of 'gotchas' that can cause problems. Relays may be 'low tech', but they are easy to use and difficult to destroy.

I'm doing a design at the moment where I'm using
relays simply because they are more robust than using a handful of mosfets - also you get isolation for free.

gintaras_bar

Level: Wannabe

Joined: Mon. May 28, 2001

Posts: 84 View posts

Location: Lithuania

#17

Posted by gintaras_bar: Tue. Mar 6, 2007 - 09:55 AM

Dear Kartman,

why are you scarying people with triacs usage :) I'm working with triacs/thyristors for 20 years. They are nice things to work with, especially when loads aren't high.

On the other hand, relays are very robust. But... they can't be switched on and off immediately. They are bouncing and making arcs/sparks. This is not good for inductive load (motor). And they are requiring a lot of power (in comparison to optotriac/triac solution) to be driven.

If I would be involved in such project, I would use optotriac/triac pair. If current control is based on triacs and works well, why it should fail after modification (if it is done correctly of course).

But debugging is another question. You can destroy motor using relays either. I would recommend to try with lightbulbs first, as Kartman suggested, and then switching to motors, connect PTC thermistors (or lightbulbs - if something is going wrong, you will notice it very easily) in series with your motor.

Good luck!

Gintaras

emilevanderlaan

Level: Wannabe

Joined: Mon. Apr 12, 2004

Posts: 79 View posts

Location: Amstelveen/Amsterdam

#18

Posted by emilevanderlaan: Tue. Mar 6, 2007 - 11:08 AM

Lennart,
I use to work for a company that made electronic for central heating systems,
The 3-way valve that you use is probably a Honeywell or some other vendor.
We did always controlled the 3-way valve by a triac as it is the best price / performers.
But a relay will also work, I recommend to use a 10A or 16A relay as the chances that it will stick are less. You will have once in a million times that it will not switch but that is probably not a big problem.
On the Triac, I recommend to use a triac that is over spec, 10A and 800V as the dV/dT is better and you do not need so much of spark suppression. And it will not really cost a lot more, if you are not making a lot of them.
Triggering the triac, do that on the zero cross interrupt, just set a bit in ram if you want to have the motor on and test that bit in the interrupt routine.
Have a fuse in your circuit that will blow if you have a short circuit in your 3-way valve this will also help in wrongly installation.

markus_b

Level: Posting Freak

Joined: Tue. Mar 20, 2001

Posts: 1510 View posts

Location: Switzerland

#19

Posted by markus_b: Tue. Mar 6, 2007 - 07:47 PM

- How do you interface the triac to the AVR ?
- Can it be sufficient just to connect the port pin with a resistor ?
- Will the triac trigger on the positive and negative wave, even if the voltage at the gate is always positive ?

Markus
I'm thinking about building a timer for the 230V motor of my coffee machine and would like a simple circuit.

Markus

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#20

Posted by Lennart: Tue. Mar 6, 2007 - 09:05 PM

Hello Markus

I have not connected my AVR to a triac yet. I started this thread to get advice how to do this in a proper way.
If you read the answers I have received so far you might get some information. On the other hand people have different opinions how to do. I have seen several application notes where the AVR is connected to the triac with just a small resistor in between.
On the other hand I have been adviced not to do so when getting answers to my thread. As far as I understand you can use the same positive gate signal for the negative wave but if I undestood one the answers I got right it will draw more current in the negative wave. Don't trust my answer!!!

I have spent the most of the day to figure out the schematics for the analog control electronics that so far control my motor. Sad to say I get even more confused what part of the schematic I could "throw away" and where to interface my AVR trigger signal.
In case somebody have energy to help me figure out this I can post the schematic. As for now I am rather confused and about to give up using a triac...

Richard

markus_b

Level: Posting Freak

Joined: Tue. Mar 20, 2001

Posts: 1510 View posts

Location: Switzerland

#21

Posted by markus_b: Tue. Mar 6, 2007 - 09:15 PM

Hi Richard,

First, I apologize for stepping in your thread, I've had the impression that our goals are somewhat similar.

However, posting the schematic might bring you a good step forward. There are many knowledgeable and friendly folks around in this group !

For you particular problem, picking a relay or an electronic relay with built-in zero crossing might be the easiest solution.

Markus

zauberer

Level: Resident

Joined: Tue. Mar 6, 2001

Posts: 544 View posts

Location: Moscow

#22

Posted by zauberer: Tue. Mar 6, 2007 - 09:38 PM

...How to connect triac to AVR

Attachment(s):

triac.jpg

gintaras_bar

Level: Wannabe

Joined: Mon. May 28, 2001

Posts: 84 View posts

Location: Lithuania

#23

Posted by gintaras_bar: Tue. Mar 6, 2007 - 10:14 PM

Lennart,
look at this Data Sheet:
http://www.littelfuse.com/data/e...
we are using schematics from this Data Sheet in our commercial projects for 230V:
Figure AN1007.8 where motor is phase controlled
Optocoupler MO3021
Triac TIC206M
Resistors 330Om

AN1007.10 where motor is switched during Zero Crossing
Optocoupler MO3063
Triac TIC206M; Z0103MA
Resistor 330Om

Both versions work wery reliably. But in your case for 24V, may be lower resistance is needed - try 100Om, if triac is reliably switched on, leave it, if not - use lower 47-33Om.

So don't be confused, these are examples from "real life", copy this and you will be happy with such design.

Gintaras

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#24

Posted by Kartman: Wed. Mar 7, 2007 - 12:30 AM

Gintaras_bar -I too have worked with triacs over many years in high power phase controlled circuits. I also work with circuits critical for safety. I would say I'm pretty aware of the failure modes of both triacs and relays. Was I trying to scare Richard? That wasn't my intention. With limited experience, it is easier to apply a relay vs a triac. Relays don't have issues with dv/dt, holding current or the like.

Last week a collegue brought in his air conditioning controller. As to be expected, the triac had failed and caused the motor to burn out. The relays were all working fine. Then you have the optotriacs which are probably more unreliable as they age etc.

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#25

Posted by Lennart: Wed. Mar 7, 2007 - 01:35 AM

Don't worry, I'm not scared.
I am confused...
Obviously using triacs is a bit tricky since I have received so many answers. They all seem pretty reliable for me as a newbie at triacs but often contradict each other...
When I reverse-engineered the control-circuit used today I got even more confused. There are no fancy stuff there, just a BC547B NPN transistor to trigger each TIC206D. Base of the transistor connected to an LM324N op-amp. Really cheap stuff, have all of this at home. No snubber circuit or optotriacs. Still it has worked flawlessly for 16 years and is still on market...
There is a lot of resistors and a couple of pot's and diodes cross-connected to the different op-amps but as far as I understand they are used to sense when to fire any of the two triacs.
All of that sensing is what I want to replace with software.
Is there a good reason why I could't use transistor and op-amp together with the uC?

Richard

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#26

Posted by Kartman: Wed. Mar 7, 2007 - 02:39 AM

The design of triac circuits depends on the type of load. If you know exactly what load you're having, then the design can be simpler. In this instance, it is 24VAC from a known source, so we don't need to worry too much about transients etc like we would on a mains based circuit. Also the motor characteristics are known, so we know what currents are involved and how much inductive energy is involved. If many of these items are not known, you have to err on the side of caution and add snubbers etc. Anyway, we have a known circuit that works and has worksed for many years, so we can be pretty sure it was well designed. So I would suggest just using the BC547 and the TIC206D is probably all that is needed and not to worry about any synchronisation with zero crossing (assuming the original circuit didn't worry about it). Simply turn it on or off. I dare say if it gets turned on during the cycle, the peak current is minimal and the motor can tolerate this due to the low duty cycle of on/off.

Make sure you have capacitors on your voltage regulator to the AVR - keep the DC clean.

Also make sure your software has hysteresis in the the control so that the motor is not opened/closed if the temperature is near the setpoint. The original circuit probably has a 'window comparator' so that there is hysteresis.

Level: Raving Lunatic

Joined: Sun. Oct 29, 2006

Posts: 2757 View posts

Location: Sweden

#27

Posted by Lennart: Wed. Mar 7, 2007 - 03:19 AM

Hi Kartman

I really liked your last post. :D
Present design had one of the op-amps take care of hysteresis, and of course I will do the same in my software. Actually it's already written and waiting for the hardware to be sorted out.
If you think it's a good idea I could use a led and an optotransistor to isolate my AVR.
There would also be a possibility to use zero-crossing detection to turn on transistor at a desired time in AC-wave if this would make my motor happier. Already written and tested the software for this also since I thought it was mandatory.
Yes, yes now I feel much less confused... 8)
Thanks for your guidance

Richard

Level: Raving Lunatic

Joined: Thu. Sep 25, 2003

Posts: 2026 View posts

Location: Los Angeles, USA

#28

Posted by mneary: Wed. Mar 7, 2007 - 04:01 AM

If the circuit already works with a BC547B, just copy that section. I would guess that it's providing a generous gate drive, to make it work with the low current and voltage.

I wouldn't bother with an opto since it's only 24V. Try to copy the general grounding scheme.

Many of my comments were based on the assumption that you were working with 240V at about 2A. Now that you've told us otherwise, it's different.

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#29

Posted by Kartman: Wed. Mar 7, 2007 - 06:33 AM

Using an optocoupler might be warranted if the AVR is powered from a remote source. If powered from the 24VAC, it isn't buying you much if anything.
If the original controller did't worry about zero crossing I would suggest you don't worry about it either. Less code = less things to go wrong.

If you were going to be a little smarter, having the AVR controller networked via RS485 so you can find out what the temperature is and what the controller is doing then you might want to look carefully at how the device is powered and possible isolation. Again, the relay looks promising! I posted some assembler code a while back for a tachometer that implemented a Modbus slave device that would suit you application nicely. Modbus is an instrumentation protocol that can use RS485 and have multiple slaves on the one pair of wires.

tpappano

Level: Raving Lunatic

Joined: Mon. Dec 1, 2003

Posts: 2963 View posts

#30

Posted by tpappano: Wed. Mar 7, 2007 - 07:20 AM

Just a couple words of encouragement to the OP.
Triacs are robust devices and will serve you well
if you follow the very basic guidelines in the Littlefuse/Teccor application notes. There is
nothing strange about controlling AC motors with
triacs, and FWIW, AC motors can be speed controlled
with phase controlled triacs, but the available
torque goes down as the voltage to the motor is
reduced. Because of this, phase control methods
are most useful on loads such as fans, blowers,
and centrifugal pumps. For the last few years,
I've been using triacs ("alternistors", actually)
to speed control 480 volt motors up to 1.5 hp.
I have actually never used a mechanical relay to
control a "power" AC load, and so far none of my
triac equipped devices have failed. Triacs are
worth a little study and experimentation to
understand, and are excellent tools to add to your
"bag of tricks"

Tom Pappano
Tulsa, Oklahoma

Level: 10k+ Postman

Joined: Thu. Dec 30, 2004

Posts: 23819 View posts

Location: Melbourne,Australia

#31

Posted by Kartman: Wed. Mar 7, 2007 - 11:30 AM

Tom - how do you protect your alternistors? This is the critical thing to ensure their survival. With light dimmers - if a bulb blows there's a fair chance it will take out the triac - even a 60W lamp with a 40A triac. Get any reasonable surge - dead triac. Maybe the power supply downunder is worse than yours? I usually used SCRs - much more robust and expensive, however they rarely fail. The optos are more likely to fail. The SCRs also withstand dead shorts across 3phase feeds whilst the circuit breaker trips. Unfortunately industry would counter your claims, the sales of contactors (relays) are still strong worldwide and are the number one choice of motor switching in industry. Fast surpassing triac control is variable speed drives - using phase control for induction motors can be risky as the slip angle can get too large and cause the motor to overheat.