Home Artists Posts Import Register

Content

I bought this generic floor-standing HEPA air filter unit purely so we could take it to bits.

It's actually a bit too big, even for the second bench that is intended for larger items.
The fetching pink background is a silicone bench mat that was sent by Scott of the  Defpom channel.

By far the most exciting bit of circuitry was the bit that goes "ping!"

https://www.youtube.com/watch?v=Xi_X-erjacU

Files

Teardown of a big HEPA air filter unit (with schematic)

I got a big floor-standing air filter unit so we could take it apart and see what's inside. These were being sold at a very jacked up price during the pandemic, and then the price crashed when the residual stock had to be cleared. There's a good chance that you'll find these stuffed into cupboards of public spaces, or just thrown out after they have served their purpose of ticking the pandemic air quality box. In some cases the units will have been used without even taking the plastic wrapping off the filter first. (I came across two like that.) The curse of these units is that they are sold in the same way as inkjet printers, where you are tempted by a cheap unit - but then pay the price later with the cost of replacement filters. Or worse still, the seller just stops selling the new filters and you are left with a useless unit. Pink background is a silicone bench mat courtesy of Scott (Defpom's Electronic Repair channel). The function is very simple. Air is pulled in through a multi-layer filter and then blown out the top past the obligatory ioniser. They don't need a circuit board or touch buttons. You could literally have an on/off switch and an optional high/low speed selector switch. Because of the multi-tap motor winding, only one of the speed connections should be powered at any given time. Powering more than one would effectively add a shorted winding section. If hot-wiring the motor the unused taps must be insulated. Motor wiring:- Black - common (I'd expect there to be a thermal fuse on this connection.) Red - high power White - medium power Yellow - low power The buck regulator number is LN8K05 The processor is an SC92F83A2M The triacs are JKT1D8W I'd guess these units are mass produced in China with labelling and pricing to suit the company selling them. The design is very cost optimised, and much more interesting as a result. They've literally squeezed every feature they could out of a cheap processor. The filter construction is unusual. The honeycomb frame with loose carbon granules is quite neat. The bit that goes "PING" took me a moment to get my head around. It's genius. Basically a hardware envelope generator for a bell noise, where the processor just needs to provide a square wave at the desired frequency. If you enjoy these videos you can help support the channel with a dollar for coffee, cookies and random gadgets for disassembly at:- http://www.bigclive.com/coffee.htm This also keeps the channel independent of YouTube's algorithm quirks, allowing it to be a bit more dangerous and naughty. #ElectronicsCreators

Comments

Anonymous

Love the "Oxidise Karens and Destroy Moon"

Anonymous

“Double Death ☠️ Beam” Capacitors! 🤣🤣 Love it, Clive 😜

Anonymous

I haven't even watched this yet, but I had to be the first one to write "Ah! I see you have the machine that goes PING!"

NN Thomas

Had a look for that component you couldn't find a datasheet for. First go there was nothing but then changed the 'o' to a '0' (LN8K05) and found it on alldatasheetdotcom

Anonymous

Damn, was expecting a teardown of a High Energy Particle Accelerator ;)

Anonymous

An Oxidise Karens machine is definitely the future! The figuring out of that pinger circuitry was a masterpiece of detective work! I’d never have got that one from what you could see there. My hat is off to you sir!!

Bill Donnelly

Hi Clive, what are your thoughts on HEPA filter for 3D Printing? YouTube is awash with "solutions" . . .is there actually a problem? I have mt Prusa Mini+ in an enclosure, print daily next to my work bench. On warmer days (I'm in Oz so that is often ;-) I do sense the plastic and occasionally on large prints feel some eye irritation. Like most things my intuition is that 'dose' is the main factor. I'd like to build an extraction system (mainly just an excuse to design and build), but am aware of so many fake filters I may not. Thoughts? Thnx for the great videos, you've even got me watching the streams now . . . in the background while I'm pottering in my garage/workbench . . . it's like the random mate that pops in for a chat while you are trying to concentrate . . . except with your streams I don't feel guilty ;-) Thnx - Billd

Mike Page

I'm intrigued by the touch switches. I use something similar on my testers as an earth reference - but this (^^^) is minimal. I had hoped to use something similar as a simple switch but there's a snag: if the capacitive current isn't enough to overcome gate leakage, you'll never get a transition. Accurate leakage specs are notoriously hard to come by. They usually say 1uA over temperature and you only have 50nA or so capacitive current. It might well work in real life, then you'll get a dirty board and it won't. I can't charge money for something that works by luck. So I end up over-engineering it.

Mike Page

That piezo - oh it's envelope modulation! You can do things with PWM duty cycle but it's a lot more work and changes the harmonic distribution. This is neat.

bigclive

After I rebuilt this unit I did actually put it in the same room as the 3D printer. Not sure if it's needed. I use PLA for almost everything which I think may be low risk.

bigclive

I'm wondering if they were using comparator inputs to detect a lower voltage differential. Or maybe the chip does have touch input circuitry. As well as the usual supply decoupling capacitor there was one other one connected to a pin, hinting at an internal regulator or reference voltage.

bigclive

It's very cool. Means the software just has to create a simple square wave, and gets a high resolution volume fade for the ping noise.

Mike Page

You don't generally get that many comparators on a chip. But if you have that many comparator inputs you could switch around the pin functions in real time. I'm not sure it solves the problem; but generally analogue pins are better specced for leakage than straight digital. In fact the pins on my preferred small target - LPC824 - are specced 10nA. But I wanted some positive feedback to set a threshold. So I had an external SOT23 OR gate, and that was specced 1uA. I just wanted too much, didn't I? At some point I had to draw a line and do what works. I've not used the "touch switch" microcontrollers (eg LPC845 and Microchip have a bunch of them) but I think they work on mutual capacitance. They don't need the chip elevated to Live.

bigclive

I probably still have the hideously unpleasant prototype touch controller I designed decades ago, that used a microcontroller with directly exposed main referenced touch pads. I had issues with the input impedance and bias resistors meaning you could really feel the tingle when you touched the pads while grounded. It might be worth revisiting to see if capacitive coupling would work with a modern microcontroller.

Mike Page

Mine are capacitive. But I use a CMOS opamp. Like a voltstick, really. If you have a PDL310 or DL420 you have it.

bigclive

I do indeed have a PDL310 (and its sibling the PDRC380). And a SOK36 with capacitive sensing too.

Zeedijk Mike

Active Carbon is (or was) used in gasmasks, so it removes more than just bad odors.

Mike Page

The PDL310 can be turned into a DL420 with a little jiggery. You gain a high current L-N test. IMHO the replacements with rotary switch are much nicer to use. The PDL310 "Hamal" label would crack and the buttons would go. Also the LCD design was by moi. Then I told the client to take it seriously and he never looked back. There are some things I should not do.