Replacing an SSR in a kiln? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Replacing an SSR in a kiln?

1 2 next

Move Thread

LAN_403

GreySoul

Tue Sept 04 2007, 01:53AM

Registered Member #546 Joined: Fri Feb 23 2007, 11:43PM
Location: Albuquerque, NM
Posts: 239

So I'm buying a glass annealing oven (or for anyone outside the glass blowing world, a kiln) that uses a digital controller with a low voltage signal to a solid state relay. The relays are 2 x 220v 30A. I have heard that the relay they use tends to fail closed, which can ruin glass work in short order.

So I'm looking for solutions. My first thought is to replace the cheap SSRs with some rather stupidly large IGBTs I have on hand. My question, I suppose, is would I eventually have the same problem, with an IGBT failing in the ON state? Or do they always fail off? Are there any other solid state devices you all can think of I might use? I've considered a mercury relay, but replacement costs are high on those, and mercury around a hot kiln doesn't seem too good of an idea anyways.

I've asked a few glass blowers and the kiln manufacturer, but they've all been pretty clueless (no one even knew what an IGBT was)

The kiln itself has an over temp protection circuit, but it can't be set below 2300F and by then my glass is a corrosive puddle on the kiln floor. I could rig up my own over temp circuit, but by the time it triggers, even if i set it 5-10 degrees over my set point, the residual heat would cause problems, if not outright ruination.

.... anyways, I'm getting frustrated by the lack of knowledge from the people building the kilns.... I hope you all have some better suggestions than "replace the relay every 6 months before it has a chance to fail" (their advice). I asked if they guarantee the relays for 6 months, they said 30 days.

Thanks in advance

-Doug

Bjørn

Tue Sept 04 2007, 02:31AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

I have seen the result of such a failure and it boiled what was placed inside and destroyed the heating element.

What about a detector that compares the output of the of the switch with the control signal and if there is a discrepancy it uses a mechanical relay to disconnect everything?

Carbon_Rod

Tue Sept 04 2007, 02:40AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

Infrared radiation can damage color vision - use Orton protective glasses.

What kind/make of kiln?
Gas
Electric (assumed this is what you meant)
Microwave

Solid state relays are usually just a TRIAC with built in opto-coupler thats CSA/UL approved.
(About 1/10 the price of IGBTs of the same rating.)

Thermocouples without ceramic sleeves, pyrometers, or relays generally malfunction before other parts.

Hazmatt_(The Underdog)

Tue Sept 04 2007, 05:56AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

Okay, first of all if you're replacing a SSR you will be replacing it with a suitably rated TRIAC, not an IGBT. The TRIAC is a fullwave AC switch for phase control and its just an enormous lamp dimmer circuit that can handle an inductive load, your heating elements. What you will want to do is have the controller a couple of feet away from the kiln so the heat doesn't affect it. You know that if the kiln is at 1800*F and the controller is right next to it, the controller will probably be seeing 150*F from a thermal gradient external to the kiln.

The next issue is control. The Basic Stamp has us build a hatchery oven controller using a PID algorithm, which is going to be very handy to you in the future. For $150 investment you can get the programming, controller, code, and enough information to start development on a good ramping-soaking controller with good hands on experience.

I even turned the basic stamp into a voltage ramp control board for anodize process. It's very versitile.

I was going to build an oven for my glassblowing but I haven't gotten around to that just yet. I might be able to help you out in the future, but if you want my help i'm sorta gonna force you to learn a bit about control systems since that's what my area is.

Matt.

Steve Conner

Tue Sept 04 2007, 08:39AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I'd suggest replacing each SSR with two SSRs of twice the rating, in series. And relocate them in a box away from the heat of the oven, or provide fan cooling (with dual redundant fans, obviously) Put an indicator lamp across each of your (now four) SSRs so you will know when any one of them has failed short. (all four lamps should light when the SSRs open) Keep spares around to increase the odds of your being bothered to replace failed units promptly. Add surge protectors to the line, since SSRs can be damaged by spikes.

The combined effect of all these measures should bump the odds of damage to your work down the bell curve by a few orders of magnitude. Of course, the meaures cost money, and depending on the value of your work, you may find it a better solution to just let the oven fail and trash the occasional piece.

The optoisolated inputs should probably be connected in series too.

A home-made IGBT circuit is unlikely to be a good idea, as the odds of your homebrew isolated driver circuit failing short will probably be higher than the odds of an off-the-shelf SSR failing.

thedatastream

Tue Sept 04 2007, 10:26AM

Registered Member #505 Joined: Sun Nov 19 2006, 06:42PM
Location: Yorkshire!
Posts: 329

Another option is to use normally open (NO) relays so that they always fail safe if there is a lack of power. You would have to use a bipolar transistor to turn them on and off with a diode across the relay coil to catch the back EMF.

Installing your own thermal cutout would be a good idea too if the current one doesn't go down to the temperature you require.

It all depends on just how much protection you want...

Sulaiman

Tue Sept 04 2007, 12:26PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

As far as I know, all common semiconductors fail short-circuit.
If there's enough energy/current available the short-circuit can cause bonding wires etc. to fuse
so sometimes it appears as if a semiconductor has failed open-circuit
but it's the connecting wires etc. The actual semiconductor device will be a short circuit.

The result is that you need to adopt a scheme as proposed by Steve Connor.
Series redundant semiconductors.

Mechanical switches sometimes fail open circuit, usually due to mechanical failure,
sometimes the contacts become high resistance due to pitting/piling and oxidation,
sometimes the contacts weld together and fail short circuit.

Personally, I would just use a higher rated SSR
I've maintained equipment that frequently had failed SSRs
replacing with the specified part just resulted in a later failure
increasing the rating 'cured' the problem permanently.
Also ensure that the SSRs are not overheating
from memory an SSR drops about 1.6V, multiplied by 30A that's 50W
if the ambient temperature is high, then removing 50W needs quite a large heatsink,
or a moderate heatsink with a fan. Then you need a 'fan failure' alarm!

sparky

Wed Sept 05 2007, 02:06AM

Registered Member #530 Joined: Sat Feb 17 2007, 07:56AM
Location: Victoria BC, Canada
Posts: 178

Some really old kilns use a pyrometric cone that melts when the kiln gets up to a high temperature. Some of the kilns - digital controlled ones use a protected temperature relay system that fires the main circuit when the temperature is selected. Pretty simple....

IGBTs? Why?

GreySoul

Wed Sept 05 2007, 06:19AM

Registered Member #546 Joined: Fri Feb 23 2007, 11:43PM
Location: Albuquerque, NM
Posts: 239

Thanks all, you've pointed me in the right direction.

Bjorn: I like the idea of comparing the signal with a kill relay - that would probably be the most straight forward approach, and something I could easily figure out on my own I think.

Carbon_rod: Eyes? who needs eyes? Actually I wear Aura Lens AUR-286 specs when I work around kilns, but thatnks for thinking of my eyes :) And, sorry I'm not being overly specific about the manufacturer.... I have reasons. It's a large front loading oven that's got some known problems.

Matt: I'm all for putting $150 into some controller package - I used to be pretty good with BASIC, so I'm sure I could pick it back up. Do you have a particular kit you recommend? If this project could spawn a commercially viable kiln controller for the artistic glass market, I could sell them at every show I go to. What's available now is based on ceramic controllers that have programs for "cones" instead of proper annealing of glassware.

Steve: Again another great suggestion, and something I'll probably do if I have to rewire the entire control panel.

Datastream: A mechanical relay defeats the purpose here. And the work inside the kiln is sometimes worth more than the kiln, so letting it slide isn't really an options. Also, molten glass is corrosive, and would destroy the kiln floor, and damage the elements. A thermal protection circuit wouldn't be able to trigger until it's too late.... like I said. The striking point and the liquidus point on boro are about 20F apart.... the kiln can bump 30 on a simple overshoot if you're not careful.... even with a good controller. by the time it got high enough to trigger a thermal overload, it would be too late - square one.

Sulaiman: The relays are 220x30A but the elements are only around 22A each once heated up. I don't have specs for the starting current - but that could be what kills the SSRs. the kiln is designed with 2 fans (in and out) with the control circuitry in a well insulated / isolated cabinet below the actual oven. The air moving through the control panel area was room temp at all times.... and the SSRs are on over-kill heatsinks - good for me, if i upgrade the SSRs. I have some old Crydom's that say 220-45A.... think I should look for something bigger?

Sparky: This is a brand new digitally controlled glass annealing kiln. Orton cones don't apply. The controller I'm getting is a 64 segment fuzzy logic ramp/soak controller with various pre-sets and holds.... they use a millivolt signal to run one or more SSRs.... up to 7 or 8 on this model, I believe. The controller reads the temp with a K type thermocouple, and runs the relays at 60Hz to minimize "drift" as the elements heat up...instead of a long on/off cycle of old analog and cheap digital kilns - sometimes a multi-minute cycle, this one switches 60 times a second. A mechanical really wouldn't hold up to that abuse for long. And IGBTs because I have a grip of them already, and my SSR stocks are kinda low. but I'm no longer considering IBGTs. Maybe for a motor controller for the glass lathe tho.... I have a pair of nice H-bridge IGBTs that would make a nice DC drive controller I think.

Anyways....

Solution:

Short term: beef up the stock SSRs, buy extras, add monitor lamps, possible make a comparator doo-hicky.

Long-term: Build my own custom controller (with Matts help, I hope)

Thanks again,

-Doug

Wolfram

Wed Sept 05 2007, 08:24AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

If you're making your own controller, you should consider using a common microcontroller instead of a basic stamp. Basic stamps are expensive, slow, and not very capable compared to most common two-dollar microcontrollers. It's not going to be more complex, most microcontrollers just require a crystal and a pullup resistor on the reset line. Many new ones (at least PICs) have built-in oscillators too (far from as accurate as crystals, but usable for most stuff), so you don't neccessarily need the crystal.

Making a programmer can be as simple as a couple of resistors on the serial port.

If you want to program it in a high-level language, compilers are available.

Programming PICs in assembly is not too hard, it can be learned in a couple of hours. Assembly it's best for simpler stuff or stuff requiring high timing precision or high execution speed. I am a big assembly fan myself, but I have to admit that if you are writing a program with lots of complex math, not requiring the absolutely fastest possible execution speed, you're probably better off with C.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob