Last modified: 2026-10-27 21:41:19
< 2026-10-26 2026-10-28 >We have a "Charles Austen ET 60 AO Linear Air Pump". We only got it last year because the previous one had failed, but now this one has stopped working. I didn't investigate last time, I just let Mark Perry replace it.
For a while it seemed to work intermittently, which I wasn't going to worry about, but lately it seems the red light is always on and the pump never makes any noise.
I took it apart and found that one of the diaphragms is not correctly seated, apparently having been assembled incorrectly.
And also there is red dust everywhere, I think from the shuttle rubbing against the electromagnet because the one diaphragm is not holding it correctly centred.
The red light is apparently just supposed to indicate that the output pressure is low, i.e. there is some kind of fault, and not interrupt the operation. So the fact that it no longer makes any noise suggests there is another problem.
With it mostly taken apart I put the shuttle back in and then tried turning it on in all of run/mute/test modes, and it never showed any inclination towards making a noise or producing movement.
So one possibility is that all the rubbing has actually rubbed through something that needs to not get rubbed through. Another possibility is that it caused it to overheat and there is a latching thermal cutout or something that I just need to reset.
I tested with a multimeter and the electromagnet seems to be open-circuit.
Inside the control box I noticed this relay or fuse or whatever with black soot marks coming out of it as if it has been arcing:
Possibly what happened is the motor drew too much current due to insulation getting rubbed off or something and shorting out, which then overpowered this relay (?) and eventually burnt out the windings and now it is an open circuit.
Rather than keep finding broken parts that I need to repair I am tempted to keep this one as a source of spare parts and buy a new pump. Although it is disappointing that this one failed after about a year. Potentially it would have been OK if I had investigated at the first sign of intermittent failure instead of waiting for complete failure.
According to ChatGPT the "relay (?)" component is just a buzzer to let you know there is a fault, which evidently I have simply ignored. I never noticed one but if it is just a 50Hz buzzer then it will presumably make broadly the same noise as the pump itself so why would you notice it, especially inside a box with lots of sound insulation.
I don't know how long the previous one worked for before failing. It could have been 15+ years for all I know. So I'm just going to buy another and see how we get on.
"ET 60 AO" means:
I've ordered a replacement for £125 from https://direct-drainage.co.uk/
And as part of this I just ran into my first instance of "agent SEO required". ChatGPT told me that they were lowest price but were out of stock, I went to look myself and it was in stock, but they have an "OUT OF STOCK" label in html that is just switched off by CSS or whatever for human users but which ChatGPT still reads, ouch.
I have the X axis end pieces printed now, plus some other sundries, enough to wire it up and test it:
And it worked perfectly first time! X axis drives back and forth from FluidNC web UI. I couldn't conveniently video it because I was also operating FluidNC with the phone because the laptop charger seems to have died.
I drilled out the mounting holes on the motors to 6mm and mounted them with M6 bolts and nuts, instead of using the threaded inserts, because the holes for the threaded inserts were too large and they pulled out.
Other errors include:
Oh! Actually. Instead of having the bolt on the outside and projecting into the adjuster block, obviously the adjuster block should have a captive bolt and then it can stick out as far as we want, and then adjustment travel doesn't have to all fit inside the block, and the block can get shorter so provide more travel. Yeah, obvious improvement, should do this.
Anyway good that it worked first time, no obvious belt slipping etc., may get worse at higher speed/acceleration, currently the speed is capped by whatever random config Codex came up with when we set up FluidNC, it is just random steps/mm etc. but it does move which is good.
< 2026-10-26 2026-10-28 >