Here's another illustration of the danger of letting The Stupid Fat Bloke loose in the workshop. As well as being the kind of idiot who knows it all and rushes in to do shit without thinking things through, he's also prone to simply forgetting simple things. He should be called The Stupid Old Fat Bloke.
The issue with the Tree installation isn't anything to do with the hydraulic pump motor or indeed a failure in any of the other mains powered components. It's simply because said SFB forgot to plug in the mains isolation transformer.
The issue is that I have a series of converters and inverters and EMC filters all leaking current into the chassis. Mainly, this is through their Y caps but also through the insulation in the motors, solenoids etc. Given that the consumer units in the house are mandated to have 30mA RCDs, there's a limit to how much cumulative Y caps can be tolerated before nuisance tripping becomes a major issue.
The mains in the UK seems to be climbing ever higher these days. I'm measuring 256Vrms on several DVMs and short of examining the actual waveform with a scope, I have no reason not to believe that's what we are seeing. This raises the leakage current from the nominal, which I think (from memory) is specified at nominal voltage.
The current is a Y cap with 230Vac across it is
I = 2 πfCV (where f = 50Hz & V = 230Vrms)
So, if we were to run leakage of 30mA entirely down to the Y capacitance, we'd be able to tolerate a value of ~415nF. Of course, that would be pretty stupid as it would allow no additional leakage anywhere downstream of the workshop consumer unit RCD.
Here's the Schaffner EMC filter that comes with the spindle VFD, which is a 4kW Yaskawa GA500:
They estimate ~5mA leakage for this filter. There's also some Y cap in the VFD itself. In the manual for the GA500, it says "If you turn on the internal EMC filter, the leakage current of the drive will be more than 3.5 mA"
"Use a high frequency RCM/RCD at the power input side of the drive and make sure that each drive has a minimum cumulative sensitivity amperage of 30 mA. The specialized breaker detects only the leakage current from frequency bands that are dangerous to humans.
If a device does not have protection against high frequencies, high frequency leakage currents can cause the device to malfunction. If you have a malfunction on a device that is not protected, decrease the carrier frequency of the drive, switch to a better breaker, or use an RCM/RCD with a minimum cumulative sensitivity amperage of 200 mA for each drive."
There's no mention of filter compt values or leakage currents in the Lichuan servo drives. The Meanwell SMPS PSUs claim <2mA each at 240V (there are 2 of them).
Bottom line - measuring the capacitance between live and ground suggests a total of around 850-900nF, which will deliver a leakage current that is well above that nominal 30mA threshold. A measurement from neutral to ground wouldn't be relevant, as neutral is held at ground, so shouldn't normally be causing any significant leakage.
My solution to this was to run the lathe from an 230:230V isolation ("site") transformer - with a strong earth bond of course. Then, any leakage within the installation would be isolated from the workshop consumer unit.
What Fat Boy forgot to do was install that isolation transformer between the wall socket and the machine. It got misplaced when the machine was last moved a few months ago. This realisation struck me during the night (as it does) and today I've gone about convincing myself if / that this is what the issue was. After some buggering about with insulation and continuity testers, it's clearly what the problem was. With the transformer back in circuit, everything is fine.
What now, Fatty?
So now, I have to reconnect the hydraulic pump assembly and and refit it into the machine. No harm done (hopefully I haven't nadgered any of the venerable hoses) but at least I know the thing appears to be in a reasonable state - and understand what I'm looking at.
No comments:
Post a Comment