Tripping mcb and rccb

[Ndagger]

Hi guys.

Just want some advice or thoughts from the pro's.

So I'm an engineer who works on espresso machines. Going back a few months ago now I got called out to a machine tripping the power, in my line of work this generally means one loop on the heating element has shorted to ground or one of the 240v solenoid coils has got wet.
I got to site and no fault was found, trips stayed up and machine functioned ok. Nothing visual to be found so customer waved us away and said he'd call back if anything happened again.
2 weeks later same thing, mcb and rccb both tripped so I took a replacement machine and swapped his out assuming intermittent fault.

Wired his machine into our workshop circuit and tested everything, kept it on test for 2 weeks and put through it's paces, nothing happened that shouldn't have. Earth leakage on machine tested to be 0.018mA with element and pump running, coils all tested >999megohm with insulation tester. The machine we installed on his site never tripped his board either. After a few weeks on test he agreed to have his machine back and see what happens.

Machine was installed and fine for 1 month and now has tripped it's mcb and the rccb again so yet again I've collected and brought back to the workshop.
Customer is reluctant to have his wiring checked by an electrician based on the fact the machine we loaned him never tripped so it must be a fault on his machine.
I am seriously at a loss of what is going on and wondering if any of you pro's have suggestions.

I clamped the machine and at peak it pulled 26a. He has it on a 32A type B mcb on a split board with rccb

Any suggestions guys?

 

[shaun1]

Perhaps you could reach an agreement that he will pay if a fault is found, and if not you will pick up the bill. (Although not finding the fault doesnt guarentee it isnt there, but it should give you an idea of what is good and what isnt)

The sequence he has discribed (if accurate) would seem to suggest the fault lies in the machine. Perhaps have another could look for signs of arcing/burning in the machine whilst an electrician is testing the circuit?

 

[Ndagger]

Yeah I think that's what I'll suggest.
Admittedly I do agree it sounds like a fault with the machine and if I was an impartial observer I would side with him and say it must be the machine too based on the evidence he's provided.
I've spent a month off and on pulling the machine apart, have visually inspected everything, pcb / coils / rocker switch / motor / run capacitor / heating elements / screens / earthing points / cup warming element and there are zero signs of arcing, water ingress, no swollen coils or capacitors and the machine has been put under heavy use in our place for a month. Nothing unusual (even makes a fantastic espresso!)

Very baffled

 

[Lucien Nunes]

as soon as he switched coffee machine rocker switch on the rccb and coffee machine mcb both dropped instantly.

This is pretty convincing if it is true. It would take a very roundabout route to justify how an MCB trip that occurs immediately on closing a switch on the machine could be caused by a wiring fault. An RCD-only trip would be easy to explain: An N-E short will sometimes only cause a trip when there is load on the circuit to cause voltage drop along the N conductor and drive leakage through the fault. The short is always there, but the RCD trips the moment the load comes on. But the MCB.... something must be going bang.

Welcome to the Society of Frustrated Fault-Finders. You might be here a while; my longest search that actually found a fault was about 3 years, by which time the machine was obsolete.

 

[marconi]

Ndagger: Could you post a link to the machine's wiring diagram please?

 

[marconi]

Ndagger: At the moment, every time the machine trips the breakers no information is captured. My idea is to split the machine up into its sub-systems and include a temporary in-line fuse into the line supply to each sub-system to provide discrimination with the MCB. A blown fuse to a sub-system will then leave a 'tell-tale' message to you to point out which sub-system had experienced an over-current event. The client could also perhaps discover which sub-system was not energised and tell you so you could bring along a spare or spares for it. If you send me the wiring diagram I can suggest where to put the tell-tale fuses and their sizes.

A suitable in-line fuseholder which come with leads is:

SCI R3-48 WHITE 20mm In Line Fuseholder 5A 250V - https://www.rapidonline.com/sci-r3-48-white-20mm-in-line-fuseholder-5a-250v-26-1586

which would cover sub-systems up to 5A. (I will have to look for another fuse holder for those sub-systems which might draw more than 5A - I have not yet but will.)

You could use the same connectors as used in the machine eg. Spade on each end of these fuseholders to insert and remove them easily from the line to each sub-system.

What do you think?

 

[Lucien Nunes]

Without ever having seen the machine, I think an element is the most likely cause, in which case the fuse might have to be large enough that it won't reliably co-ordinate with the MCB.

One strategy that bypasses further troubleshooting and might or might not be economical would be to change the elements and any other likely troublemakers and give the machine back to the customer in the knowledge that there's a 75% chance you've removed the fault, despite having no proof of which part it was in. A lot of amateurs think they can fix vintage electronics this way - replace all 100 capacitors and hope - although it tends to add more faults than it removes. In the hands of someone who knows the machines, a shotgun repair can be worth a try when normal faultfinding runs out of enthusiasm.

 

[marconi]

Without ever having seen the machine, I think an element is the most likely cause, in which case the fuse might have to be large enough that it won't reliably co-ordinate with the MCB.

So one could put low ampage fuses which would discriminate with the 32B mcb in all the sub-systems' lines bar the one(s) with high ampage elements......

 

[marconi]

Ndagger: What instrument have you used (make model no) to measure the maximum current please?

 

[DPG]

I like the temporary fuse idea.

 

[Ndagger]

Marconi - I really like that idea, it's brilliant to be fair.
A lot of other manufacturers use a similar method themselves factory standard, the pcb / motor have inline waterproof glass fuse holders. I'm a bit embarrassed the idea has never come to me myself in the past!

I've measured peak current with my fluke 325.
The machine itself isn't anywhere near vintage, it's about 3 years old and is the flagship model of that brand, think it was around 12k when new. Finding a wiring diagram might take me a few days, they don't even have a UK office apparently so I'll have to reach out to one of their distributors to get the schematic.
[automerge]1570640489[/automerge]
Lucien nunes - if it was a run of the mill low end machine I would start swapping out likely culprit components as I have done in the past with intermittent electrical issues, reason I can't with this job is that the customer will almost definitely not want to spend money on bits that just my be faulty, it's such a niche machine that prices are horrendous, there are 15 solenoids inside that are 3 figures each, not priced an element or pcb for it but at a guess I'd say 250 for an element and maybe a grand for the pcb, if he's reluctant to pay a spark to check his supply I don't think he's going to green light the best guess method

 

[marconi]

Does a maximum current of 26A tie in with the total wattage of the heating elements? 26 x 240 = 6.240kW?

What does the specification plate say for power, current and operating voltage?

Ask the client to tell you what the mcbs are for and their ratings to the left and right of the machine's mcb. Is the CU which contains the mcb in a place with a hot ambient temperature with no air flow? Is it a plastic CU? Mcbs often require de-rating for temperature and 'grouping' (- the previous question about what is left and right.)

Are there times in the cycle of producing an espresso when the heating element(s) and a compressor motor are started at the same time?

Is the cup warmer element on constantly or cycled by a thermostat?

Does the machine contain an electromagnetic interference filter at the point the power cord enters?

Does the machine have a work light for the barista to see what he is doing? Is this light fused or only protected by the external mcb? What type of lamp is used - cfl, LV LED, mains LED, LV halogen, LV filament, mains filament?

 

[GBDamo]

Without ever having seen the machine, I think an element is the most likely cause, in which case the fuse might have to be large enough that it won't reliably co-ordinate with the MCB.

One strategy that bypasses further troubleshooting and might or might not be economical would be to change the elements and any other likely troublemakers and give the machine back to the customer in the knowledge that there's a 75% chance you've removed the fault, despite having no proof of which part it was in. A lot of amateurs think they can fix vintage electronics this way - replace all 100 capacitors and hope - although it tends to add more faults than it removes. In the hands of someone who knows the machines, a shotgun repair can be worth a try when normal faultfinding runs out of enthusiasm.

Was thinking similar, could be an element thats intermittently faulting to earth.

Cant see there being much else there that could handle the fault current and not show signs of damage.

Could it be the bench testing never stressed the elements in the same way real service duties do?

 

[Ndagger]

The 26A falls within what the machine spec documents. Funny thing with the coffee industry is dataplates on the equipment always tend to have a generous margin of error, a machine badged up at 6kw may only work out on paper too say 5.2kw at full whack however the spec sheet for this machine states 5300 - 7000w.

We have a training room in our place where our baristas hold training sessions for our customers and their staff, the machine has been installed in there with numerous training sessions being done on it where coffees are made continuously for hours on end. I did this to make as real world use as possible due to the issue of bench testing not being as strenuous as real world use.

Essentially the way this machine work is this ...
There are 2 handles where you can make coffee, we call these 'group's. Each group has it's own small boiler with an element and temp stat, so each group has independent temp control, eg left group can be 95C and right can be 92C (don't ask why, it's a coffee thing!) There is also one large boiler with an element in to generate steam and hot water for tea, this is about 70% water and 30% gap for steam to fill, a temp stat will watch the temp of the water, if we ser the machine to produce 1bar of steam it uses a PID system to watch the temp of the water (under steam pressure) and calculate how hot to keep it to maintain steam pressure (~120C under pressure will generate 1bar)

All temperatures are polled constantly and a triac on the pcb will pulse the elements constantly to maintain temperatures.

When a coffee is being made, the push of the button will power a couple of coils on solenoids to allow water to flow out of boiler and through the coffee, the pump also kicks in and boosts water pressure to 9bar.

Since the elements are always pulsing, the pump always runs in conjunction with the elements and one of more powered coil.

The pcb is always powered to run the machine screens and circuitry.

This machine has 3 touch screen panels which are similar to that on your phone, it also has tons of pretty lights, all are LED.

The cup warming element is 240v and also has a temp stat, off / 25% / 50% / 75% and 100% are selectable on the touch screen and this is also monitored by the board and pulsed with a triac

 

[DPG]

I feel like I'm on a training course

 

[Ndagger]

If you want learn more you can come to the workshop and see the machine in person, just bring all your test equipment, I'll even make you a few coffees