I need to modify design for a circuit to switch 3 heating elements in a
3-phase design. The original design has 3 power relays to switch each heating element on/off . The heating elements are 2kw each . My questions are . Can i use triacs to switch the elements instead of relays ?. What are the cooling requirements then ?. Can i use pwm to adjust heating power ? Since i will be switching large loads on/off (with pwm control) what regulation problems will i encounter and can these be solved ? Has anybody experience and is willing to work as a consultant for this design ? Are there any application notes with pwm control for power loads ?
In message , dated Fri, 25 Aug 2006, mike theodore writes
Yes.
Read the triac data sheets. Triacs are cheap; use a generously rated one, both for voltage and current. For Europe, a 600 V, 12 A part might be good. Or even 800 V.
You CAN, but your product won't meet EN 61000-3-2; it's specifically not allowed to use PWM for heating above 200 W, unless the Class D limits can be met, and I doubt that you can do that. See clause 6.1 of EN
61000-3-2.
You almost certainly can't use PWM, and you need to consider what happens when you switch the loads manually.
2 kW/phase isn't all that high. Look at EN 61000-3-3 for voltage-change limits your product must meet.
I can help you with the standards issues. The e-mail address works.
I haven't looked on the web recently, but TI used to have some.
--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
2006 is YMMVI- Your mileage may vary immensely.
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
That depends on the on voltage of the triacs and the current through them. Use a heatsink that has a low enough thermal resistance to keep the junction temp. lower than the spec.
Well, kind of - for this kind of app, we usually use "phase control", which is very much like PWM, but is per half-cycle - it starts with "dead time", then after a delay, you trigger the triac, and it conducts for the rest of the half-cycle. Excellent way of regulating, especially heaters.
Another technique is called "integral-cycle" control, where your control circuitry fires the triac for a whole cycle, 360 degrees, but picks and chooses _which_ cycle it triggers on. This cuts down on EMI and power factor problems considerably.
If you design a good enough controller, there will be no problems of any consequence, and any that you do encounter are easy to solve.
Been there, done that. :-)
Yes - just concatenate my first and last name and send it to the Yahoo server. :-)
I don't know yet - why don't we search google concurrently? :-)
Yes. But they exhibit a different set of pathologies so be careful.
Needs to be calculated but for 2KW it's not going to be outrageous because triacs are essentially switches.
Your utility might not like it at all if you are planning to turn them on only part of the phase. Above a certain load level they may require muffling and power factor correction (PFC). Also, some control gear might "go on the fritz" if you'd PWM without any filtering.
It is good to realize early on when it's better to let a pro take over. I have to congratulate you for that, many potential clients only realize that after a mid-size disaster. Or not at all and then they give up.
I am pretty booked up right now but it'll help to let folks in this NG know which company or at least where it is. With this kind of stuff there'll likely be quite a bit of on-site work.
The load level isn't so important because the techniques are all similar. Only the device sizes and driver methods differ, really. Unitrode has great app notes, some of which can now be found under TI because they bought the company.
If this is just for a one-off installation and not a series I'd consider contacting a local control gear place that sells and installs industrial electronics. For example heater controllers for breweries, bakeries etc.
"mike theodore" schreef in bericht news:ecl42i$mmv$ snipped-for-privacy@ulysses.noc.ntua.gr...
Yes, it can be done, I ever build a one phase version. Supposing you've
220/240Vac resistive heaters, I'd go for solid state relays of 800V/16A which zero cross switching. Derating you know. You can control them with a pulse of 1s repetition time and a duty cycle 0-100%. So your control goes from 0-100% in steps of 20W. (More or less will not be problem.) You'll have to adapt the old relay drivers to drive the solid states and you will need some heatsinks as the solid state relays will produce some heat. Numbers to be calculated from the datasheets of the components involved. Regulations very. Not only from country to country but even from town to town but FAIK (half) cycle control for this relative small heaters will give no problems.
Yes. You will probably have to consider the triacs as expendable devices, since they will tend to fail when the heaters fail unless they are well fused and very much overrated. If they fail, for whatever reason short of massive over-current, they will tend to fail "on". Consider this in your safety evaluation of the entire system and whatever limit controls you deem necessary to assure no injury or property damage is likely to occur over the life of the product.
Roughly 1W per ampere per triac. That's a fair bit of heat to get rid of using natural convection. Fans (and liquid cooling) have other problems.
Yes, zero-voltage-switched PWM with a cycle time in the seconds is preferable for control of most constructions of electric heating elements (it's not practical for some that respond very quickly). One downside is that it can cause annoying visible flicker in lights that operate on the same electrical circuit. Another is possible thermal fatigue induced failures of the triacs (die bond fractures, die overheats, die fails on, heaters stay on at 100% power, other damage may then occur).
You have not mentioned what you are trying to control. If it is a closed-loop temperature control system then you have no particular problem. Variations in line voltage result in proportional variations in output voltage and the resulting output 'actuator' is quite linear.
I might be able to help, but for a one-off it will be cheaper to buy a packaged SCR power pack or several SSRs operated by the controller.
Yes, more or less, see the power device manufacturers. You may not find your exact application described, but it's just a combination of the zero-voltage switching and the control.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
give me a drawing / sketch and the proper solution can be found.
We have already experimented with the switching switching many types of loads.The other important factor is the switching cycle - how often will the load be switched on and off.
Triacs is by far better suited to switch loads than a relay.
Altough i'm quite experienced in some areas of electronics these high power applications are not my field and i don't want to spend too much time until i'm on the right track. Since it's a production device i want to have clear all aspects before i start any design . My last question is if somebody has worked with board traces carrying these currents and what types of connectors to use . Are special pcb's used ?. As i see the solutions are terminal barriers or terminal blocks . Any opinions ?
I did something like that 15 years ago. The goal was to regulate the temperature of circulating water in a loop at just shy of boiling at atmospheric pressure. 6 KW element, about 6 minutes to circulate the whole loop (back to the heater). They already had temperature sensors (and chart recorders) on inlet and outlet that I could tap into.
I built a small PWM circuit that monitored both inlet and outlet. When the returning water was 90 degrees C, I dropped out one relay and went to one heater (one and 1/4 heaters effectively).
The total mass of water was about 20 liters - overshooting would be disastrous because the system was closed, and there was too much heating element for too small a volume. The purpose was to sanitize a sterile water delivery system to rooms in a pharmaceutical environment.
I used solid state relays to switch power, to a set of three mercury relays to power to the elements - Today I'd just use SSR's to switch the power directly. The system has been in service for all this time with one failure - a solid state relay failed by not passing power to the control relays.
Only safety was a large mechanical thermostat inside the heater (with too slow a response to be an effective safety in my opinion). In theory that thermostat was supposed to regulate the temperature but wouldn't meet their specification - 98 C plus or minus 1/2.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
formatting link
The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----
You may want to use '2oz' ( 70 um ) copper but there's nothing unusual about pcbs carrying these currents. The relevant IEC / EN standard wil give you guidance about trace separations.
The highest I did was 100 amps although it was designed with plenty of safety margin (medical electronics). It was a backplane. We used nickel plating to mitigate any contact issues and be compatible with the lugs. The lugs were then screwed onto the board with very heavy hardware, IIRC it was 16mm bolts. Could have been 5/8" hardware but that's almost the same. Strong spring washers made sure this would not come loose but we also had sense wires that would have tripped off the power supply if it ever did. In our case there was a lot of rough air shipment so stuff could come unglued upon a rough landing etc.
We used six planes for the power traces and they were 2-3cm or so wide. The land area for the cable lugs was about 3cm by 4cm. It was "peppered" with vias to make sure all the planes were well connected. These vias were done without thermal reliefs to minimize resistance. If you do that you need to check the Gerbers right before board fab because it has happened that people thought this was an error and wanted to edit thermal reliefs back in. Mostly I make sure a technical note is sent along explaining this to the fab house so nobody will get the idea to "improve" things.
On a side note, the board to chassis contact areas must also be able to handle the full worst case load current (short) in case something goes wrong. You don't want to have a FR-4 board go up in flames. IIRC we used nickel plating there as well and the chassis was also nickel plated at those contact areas.
Jim hi Thanks for the reply One of the proposals I got is to use integral mode pwm so I fire the triac the whole 50 hz cycle leaving some cycles off . I think this can accommodate with the harmonics produced when adjusting the phase angle since the triac will be activated at zero crossings every time. What is your opinion on this ?
In message , dated Fri, 25 Aug 2006, mike theodore writes
There is a very long story about this an IEC/EN 61000-3-2, together with the specification of the measuring instrument for harmonics, IEC/EN
61000-4-7. The latter was changed, to increase the measurement bandwidth so as to include interharmonics (frequencies not exact multiple of the supply frequency). Before this change (which has been suspended until the problems are solved), the arrangement you propose was permitted, as being 'symmetrical control'. But in fact control of this type may not be symmetrical, and even if it is, it produces interharmonic currents that the new measuring instrument takes as being even harmonics. The limits for even harmonics are very low, so the equipment is very likely to fail.
People have been working on this problem for about a year, and the full resolution may be achieved in November. However, since there isn't yet full agreement on measured values on what appears to be the same circuit (it's not a round-robin test, though), there may be a delay.
The thing to do is to study the latest editions of EN 61000-3-2 and EN
61000-4-7, disregarding clause 7 of that, so that the wider bandwidth is invoked, and make sure that the switching that you choose will allow the product to meet the requirements of EN 61000-3-2.
--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
2006 is YMMVI- Your mileage may vary immensely.
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.