Electrical engineers and electricians assemble! brew controller question

[hoppyscotty]

My 3 vessel brew system project has been working flawlessly for about 10 brews or so since mid last year. at the heart of the system is my DIY controller, which not being an electronics or electrical engineer I made as simple as possible with no fancy features and was feeling pretty smug about myself on the basis I've managed not to electrocute myself or burn down the house so had chalked that up as a success.

However, pride comes before a fall, on my last brew there was a smell of electrical burning and a whisp of white smoke emerging from the controller so clearly something was not quite right. Upon further inspection I've found the negative terminal block has clearly overheated and melted, and there are signs of an SSR overheating with distortion signifying its overheated.

I'm trying to uderstand what has gone wrong and if its a fundamental problem with the way I've wired it up or just a failure of a component.

Firstly the SSR has loads of capacity with a 75a rating for a 24a load...is this likely to be a primary failure of the SSR? I know these cheap Chinese ones can be a bit unreliable or not particularly long lasting.

Secondly, since I'm sharing the same power supply with my HLT and Kettle elements I have a switch to switch between the elements, so either sending power to the HLT or the Kettle. The way I've wired it the live output of the SSR go through the switch and then diverted to either the HLT or Kettle element. And the negative/nutral wires are all connected up together in a large terminal block. Does this sound like a good/correct/safe way to wire this up through the switch? This terminal is close in proximity to a live terminal block...could it have been arcing between the two?

Given all this, are these two failures linked? i.e. the SSR failure causing the negative terminal block melting, or visa versa or two seperate coincidental things.

Thanks.

 

[jof]

Heating usually occurs due to resistance.
This can happen when you don't get a large enough area of contact between the wire and the terminal block, especially if there is/was any tarnishing on them.

Another common place is if you get arcing in a switch each time it goes on/off. That can lead to degrading the switch contacts by oxidation and pitting. (Think about how we had to change the points on cars 30 years ago)
But I would only expect that if you keep repeatedly switching the switch on/off like a small child with a light bulb.

Also I hope you are not using the SSR to do pulse width modulation power control. I.e it's not switching on/off every few seconds.
There might be something in the technical documentation about duty cycle that tells you how fast it can be switched on/off.

 

[The-Engineer-That-Brews]

Two frequently encountered issues with SSRs:

1. If an SSR was bought from Amazon, eBay etc it will almost invariably been falsely re-labelled with an exaggerated capacity rating. The biggest target for this activity is FOTEK (see https://www.eevblog.com/forum/reviews/warning-dont-buy-fotek-solid-state-relays/). Often the relabelled ones will hold up for a while but then fail. Personally I only ever buy SSRs from reputable sellers such as RS, Farnell, DigiKey etc.
2. It's often misunderstood how important it is to mount an SSR on a good heatsink, which MUST be in a well-ventilated space (no heatsink can ever work properly if it's in a closed box with no circulation of external air).

The reason the second point is important is that when people use SSRs they think they are just an electrically controlled switch. Well, they are... sort of... but unlike a switch, they have a bit of a voltage drop across the "contacts" when they are switched on (usually about 1.5v). This doesn't sound particularly important, but you have to remember that power = voltage times current. So if you're passing 13A through the thing then it's going to dissipate 13 x 1.5 = 19.5 Watts (i.e. about 20 Joules per second of heat energy). TLDR: they get hot, baby!

So if you're going to use an SSR to control a 3kW boiler that's going to be on for about an hour, you need a decent heatsink on the outside of the box:

 

[The-Engineer-That-Brews]

Heating usually occurs due to resistance.
This can happen when you don't get a large enough area of contact between the wire and the terminal block, especially if there is/was any tarnishing on them.

Another common place is if you get arcing in a switch each time it goes on/off. That can lead to degrading the switch contacts by oxidation and pitting. (Think about how we had to change the points on cars 30 years ago)
But I would only expect that if you keep repeatedly switching the switch on/off like a small child with a light bulb.

Also I hope you are not using the SSR to do pulse width modulation power control. I.e it's not switching on/off every few seconds.
There might be something in the technical documentation about duty cycle that tells you how fast it can be switched on/off.

There's no problem at all using an SSR to do power width modulation by turning it on and off frequently, although in practice you should really only do so by turning it on/off for complete half cycles (in fact to control a big load like a kettle, you should probably only use SSRs of the 'zero crossing' type, which will automatically only turn on/off when the mains voltage goes through 0v every 0.01 sec)

Note that SSRs don't have mechanical contacts. There is no oxidation or pitting.

 

[The-Engineer-That-Brews]

Note that kosher SSRs from RS/Farnell/Digikey etc are significantly more expensive than similarly rated ones on 't interweb. This is a bit of a clue as to what is going on. Dodgy people with no interest in your house burning down buy them up, re-label them, flog them cheap and laugh.

 

[hoppyscotty]

Yeah I've always been a bit concerned about SSR cooling even though they are on heat sinks and heat paste...I am planning a new controller incorporating my 'wish I'd done it this way' features, including an external heat sink or having the SSRs mounted on a heat sink in a separate enclosure with a PC fan to pull air through for forced convection, but since this one had been working apparently flawlessly I had not been prioritising this new project.

Though the SSR looks a bit heat distorted, there are no signs of burning so I don't think the SSR is the source of the burning/smoke as it worked flawlessly, so maybe I've caught it in the process of failing. The main issue is the melting of the nutral terminal block. It also bothered me it was the negative/neutral side that overheated, but a bit more research suggests that melting negative terminals can also be a sign of a failing heating element so that could also be a cause. I can do a resistance check on the heating elements, which to be fair were inherited in a box of parts from the chap I bought the vessels off along with the FOTEK SSR's, so I've no idea how old they are.

Ultimately I'm happy if it's just a failing component...they can be replaced...but more concerned I've done something dangerous with my wiring and connections. Everything is over engineered in terms of component spec (cheap Chinese SSR's aside) and wire gauges etc. so taken all the precautions I can on that side.

I have a spare SSR I can use until such time I invest in some better ones.

And yes, I am using the SSR in conjunction with a PID controller so is being used in pulse width modulation mode. But thought that is exactly what they're for and are commonly used for this purpose and a great variety of applications.

 

[The-Engineer-That-Brews]

@jof is correct with high current loads, making a really solid connection to all terminals is vitally important otherwise you can get severe overheating.

This is equally important in your mains plug, your connection to the SSR terminals, and the actual kettle.

A common cause of failure wherever there are screw terminals (e.g. in a plug) is that the connection is initially fine but over time the effects of vibration and thermal expansion cause the screw to work loose. Once that happens you're in a dangerous and often invisible downward spiral, because extra heat builds up and makes the connection get rapidly worse.

With the mains plug for something like a 3kW load that's going to be on for extended periods of time (e.g. a kettle) you really want to try to get something like one of the 'MK' plugs with a proper screw terminal post that make a really solid connection and can't easily work loose (stock image - ignore the outdated wire colours)

 

[The-Engineer-That-Brews]

For the connection to the SSR itself you should if at all possible be using crimped and/or soldered ring terminations on the cables that are a good fit to the terminal posts on the SSR.

Do yourself a favour and bin the FOTEK SSRs altogether: it's just not worth the risk.

Get yourself one of these instead: https://uk.rs-online.com/web/p/solid-state-relays/9224996
Note that the control voltage is actually anything in the range 3-32 Vdc.

 

[hoppyscotty]

Yep, knew the Fotek ones were not great and on the list for replacement, but never seemed to get around to it. Being bumped up the priority list now though. The connections to the SSR are all crimped so no issues there in principle. But to be fair the FOTEK SSR's were inherited off the previous user who was brewing with them for several years albeit incorporated in a different control system, and I got a good ten or so brews out of them plus countless heating cycles for heating up water for cleaning etc. so they've lasted some time.

As for the terminal block connections, they are all screw type via a DIN rail component that is rated sufficiently, but Its always possible the connection could have backed off.

For now I've replaced the DIN terminal block connector for a 32A WAGO connector so I can position it well away from anything else and replaced the FOTEK SSR with a spare I had just to get me though this weekends brew. Will order replacement SSR's from RS or something.

Will also investigate the heating element to see if that is failing.

 

[The-Engineer-That-Brews]

@hoppyscotty nice one - sounds like you made a good solid job of it

melting negative terminals can also be a sign of a failing heating element

Hmmm not too sure about that: you get the same current through the live terminal as through the neutral one?

 

[hoppyscotty]

Yep sounds about plausible to me too but I was concerned that it was the nutral terminal block that was the item that has overheated and melted causing the smoke and not the SSr which has bulged but no signs of burning. One hypothesis is the nutral terminal block was very close to the live so may have had some arcing. We’ll see.

Just heating some water now to clean some kegs so will see how it goes.

Thanks all for ideas/suggestions and other snippets.

 

[MashBag]

I have seen this before. A few times.

Do you have an SSR bypass?

 

[hoppyscotty]

I have seen this before. A few times.

Do you have an SSR bypass?

Not knowingly. Just connected up pretty conventionally from what I can tell. Input side directly connected to the PID and the output connected to the live of the heating element.

My only worry about the way I've connected things up is that since I have a switch to enable switching the power to either the HLT or Kettle element the output of the SSR goes through a 3-way rotary switch before the elements (HLT/Kettle/OFF). The neutrals are all tied together back to the supply neutral. Clearly only one element can be live at once so no way to be pulling more load than intended by accidentally having both elements running simultaneously.

 

[The-Engineer-That-Brews]

Not knowingly. Just connected up pretty conventionally from what I can tell. Input side directly connected to the PID and the output connected to the live of the heating element.

My only worry about the way I've connected things up is that since I have a switch to enable switching the power to either the HLT or Kettle element the output of the SSR goes through a 3-way rotary switch before the elements (HLT/Kettle/OFF). The neutrals are all tied together back to the supply neutral. Clearly only one element can be live at once so no way to be pulling more load than intended by accidentally having both elements running simultaneously.

Blimey that must be quite a meaty rotary switch, to handle 13A

 

[Vinylwasp]

This is a very interesting thread as like you @hoppyscotty I built my own electrics using opensource software and a proprietary brewery controller, and about 12 brews in I burnt out a 16A rated toggle switch rather than a terminal block as in your case

On digging into this I eventually got this advice from a friend who has an electrical engineering background and following this thread I'd be interested in hearing @The-Engineer-That-Brews view on this:

switches (terminal blocks) are rated in a number of different ways, absolute current it can handle, absolute breakdown voltage, and power it can transfer (V * I), etc' which seems logical and explains why switches have so many ratings

In terms of the current I was drawing when brewing he went on to explain that 'heating elements can generate significant voltage and current spikes when switched on while they heat up and reach maximum impedance'.

So what he was saying is that as the PID is switching the SSR through multiple cycles per minute sometimes, the current being drawn through all the components may have peaks that far exceed the 16A rating of the switches I used multiple times while the water hits the target temp (100's even). The construction of those rocker switches was cheap as well with contact points were not perfectly at 0 degrees (planar) either, so that was an inbuilt design weakness as well.

I upgraded to 20amp Rotary switches after that, but now I'm thinking that I should upgrade my terminal blocks and move my SSR/Heatsinks as well. Thanks @The-Engineer-That-Brews

 

[hoppyscotty]

Blimey that must be quite a meaty rotary switch, to handle 13A

think its rated for about 60a. Like something you'd see in Frankenstein's lab

 

[hoppyscotty]

This is a very interesting thread as like you @hoppyscotty I built my own electrics using opensource software and a proprietary brewery controller, and about 12 brews in I burnt out a 16A rated toggle switch rather than a terminal block as in your case

On digging into this I eventually got this advice from a friend who has an electrical engineering background and following this thread I'd be interested in hearing @The-Engineer-That-Brews view on this:

switches (terminal blocks) are rated in a number of different ways, absolute current it can handle, absolute breakdown voltage, and power it can transfer (V * I), etc' which seems logical and explains why switches have so many ratings

In terms of the current I was drawing when brewing he went on to explain that 'heating elements can generate significant voltage and current spikes when switched on while they heat up and reach maximum impedance'.

So what he was saying is that as the PID is switching the SSR through multiple cycles per minute sometimes, the current being drawn through all the components may have peaks that far exceed the 16A rating of the switches I used multiple times while the water hits the target temp (100's even). The construction of those rocker switches was cheap as well with contact points were not perfectly at 0 degrees (planar) either, so that was an inbuilt design weakness as well.

I upgraded to 20amp Rotary switches after that, but now I'm thinking that I should upgrade my terminal blocks and move my SSR/Heatsinks as well. Thanks @The-Engineer-That-Brews

Ah interesting. I'm running 5.5kw elements so drawing about 24a so rated everything for a minimum of 32a and most things 40a+, but who knows what it might be spiking upto if that is the case.

The rotary switches you have linked to are the same type as mine but mine are rated to about 60a I think. They are pretty hefty things.

Defo upgrading the SSR's so maybe that would help as I would guess/hope that one of the benefits of better quality SSR's would be better control of those spike?

 

[The-Engineer-That-Brews]

Hi @Vinylwasp yes, with the wide availability of bits and bobs off the shelf it feels like electrical engineering is a bit like Lego... but unfortunately when you're switching a few kW of power things can get a bit more complicated.

I'll answer your last point first.

Your friend is right that the electrical resistance of a heating element generally increases as it gets hotter, so there can be a bit of an in-rush current. However in the cases of our brewing kettle elements the element shouldn't be getting so hot that it makes a big difference (it's a different story for something like an electric fire element or hairdryer, where the element gets hot enough to glow red).

Amusingly however, to put some numbers on it I've just done a quick check on the "3kW" elements in (a) my kitchen kettle and (b) my 30L Burco.

• at room temp they are both 25 Ohms
• at boiling water temp they come down to 22 Ohms

Well that'll teach me to jump to conclusions eh... it's the opposite way round to what I expected!

I suspect that's because the manufacturers have designed in a feature to increase the resistance when cold, precisely to limit the in-rush current. I guess that goes to prove what I was saying about it being not just like Lego.

Anyhow the bottom line is that with this kind of heating element you don't get an excessive in-rush current even with a cold element. Furthermore there's another more subtle effect at work (warning: slightly geeky bit ahead).

If you're sensible and controlling your kettle element with a zero-crossing mode SSR (which is a darn good idea for many reasons) then it's not like the kind of phase-angle control that you get with something like an electric light dimmer switch:

For that kind of circuit, the overall power delivered into the load (light bulb) is controlled by "turning on" the switching element part way through the mains cycle - as shown in the picture above. This is called a 'phase angle' dimmer and it uses what's called a "random turn-on" SSR. In this kind of circuit, when the SSR switches on, the voltage across the bulb suddenly goes from zero to (e.g.) 100v or even 270v. This definitely CAN cause an in-rush current... as well as other effects like bad RF interference, although in reality even light-switch dimmer switches use 'snubber' circuits to limit this.

Controlling a heavy load like a 3kW heating element with that kind of approach is definitely not recommended (although some people do it, by using cheap "drill controllers" ). Apart from anything else it's simply not necessary: the element and 20L of water ain't going to change temperature very quickly (!) so there is simply no need to control the power on every cycle like you have to to stop a lightbulb flickering. You can just turn it on for, say, seven mains cycles in every ten (which gives you 70% of full power).

Instead the circuitry in a zero-crossing SSR, which is almost certainly the type of SSR you are using, is pretty clever.

The clever thing about a zero-crossing SSR is that it always waits to turn on until the next time the mains voltage goes through 0v (which it does 100 times a second). This little extra bit of delay guarantees that instead of your heating element seeing a voltage that suddenly goes from zero to 100v (or 270v, even) it just sees the smooth sinusoidal increase of the next "incoming wave" of the next mains cycle:

You can buy both types of SSR, but the 'normal' type is usually zero-crossing.
You need to be careful to check the spec though, and make sure you're buying the type you expect - and don't believe the label of any SSR that you buy off Amazon or eBay etc.

OK - enough geekiness for now. Bottom line is you probably don't have to worry about in-rush current.

 

[Vinylwasp]

Bottom line is you probably don't have to worry about in-rush current.

Thanks for the detailed reply, I now understand what Zero-Crossing means and will make sure I choose the correct SSRs going forwards. Interesting about the inverse impedance values of the elements, but that's why we have smart engineers. Cheers

 

[The-Engineer-That-Brews]

In case it helps, here's a closer look inside one of the TETB power controllers.

First and most importantly, ALL external metalwork (heatsinks, case, control knobs, etc) MUST be properly bonded to earth.

This isn't "elf an safety gorn mad" ... it's just bloody sensible. You're dealing with high power mains in an environment where there's water sloshing around and a lot of bits of earthed metal. Sometimes things go wrong, and this stuff really can kill you.​

Which leads to:

1. Use a plastic box with a rubber sealing strip. External junction boxes (which is what I'm using below) are a good cheap option.
2. Insulate the bare mains connections inside the box too. Water might get in, and sooner or later you'll accidentally leave it plugged in when you take the lid off. Either use electricians tape or (as here) heat-shrink sleeving.
3. Use the correct colours of cables - you might not be the only person who ever uses this box, and someone might assume "it's safe to cut the blue wire because it must be neutral".
4. Use thick enough cable - personally I prefer cable with 2.5mm sq of copper. It's not strictly necessary for the current capacity, but it makes much better connections.
5. Wherever possible use securely crimped ring connectors at the terminal posts - as in the first picture, which shows the connections to the SSR and in the background the safely earth for the heatsink. Avoid like the plague the kind of terminals where a screw terminal pushes down on the conductor (a.k.a 'chocolate block' connector strip). With stranded cable these are a prime source of connections that come loose over time and cause overheating. If you have to use this kind of terminal (e.g. inside the mains plug) then your best bet is to crimp a 'thimble' (thin metal sleeve) over the cable first. Do NOT 'tin' cables going into screw terminals with solder: it makes them even more susceptible to working loose.
6. Use proper 'heater' cable for the mains lead - search for cable meeting specification "3093Y". This is slightly stiffer, but the main point with it is that the outside jacket won't melt if you accidentally leave it lying touching the outside of your kettle...
7. When you need an internal connector, use a proper 'Barrier Block' (the black rectangular thing in the middle of the 2nd photo) and DON'T put too many wires into each strip. The terminals on these must be a flat metal plate that tightens down onto the cable - NOT a screw that digs into it.
8. As mentioned in previous posts above, the heatsink must be big enough, and MUST be in the fresh air - either by it being on the outside of the box, or by forced ventilation of EXTERNAL air with a fan (but in the latter case be very careful to make sure that the air vents can't be blocked or easily sprayed with water...). Note that decent SSR heatsinks are not cheap, and often cost more than the SSR itself. To control a 3kW kettle element over a 1hr boil in a hot brew-shed you want one with a thermal capacity of no more than 2-3 degrees per Watt, to keep the SSR at a sensible temperature (e.g. RS 171-9070 - currently £13.88)
9. As also mentioned above it's highly recommended only to buy SSRs from reputable suppliers (RS, Farnell, DigiKey and so on). I have never had one from tAmazon, eBay, etc that hasn't been falsely re-labelled with the wrong capacity (see here for details of how to spot them: https://www.eevblog.com/forum/reviews/warning-dont-buy-fotek-solid-state-relays/ )

@Chippy_Tea given the safely implications, any chance of making this post "sticky" at the top of the thread?