# Suction testing & evaluation

In order to find a well fitting motor / filter / dust-exactror combination, we need to keep some aspects in mind.

- **Suction power relative to ambient** (well-working vacuums reach around 25 kPa)
- **Suction power through a filter-cyclone-combination** (especially with actual dust)
- **Suction power vs. power consumtion** (Useful if we are using a battery)
- **Suction power vs. overheating** (safety testing)

## Next Steps

- [x] Find a proper negative pressure sensor (e.g. https://www.adafruit.com/product/3965)
- [ ] Find a simple 10K NTC thermistor
- [x] Design and 3D print a really simple universal connector for a vacuum hose
- [ ] Write a super basic script (ESP32/Arduino) for reading and calculating suction power

## Test subjects and setups

### Raw suction power

**Why test:** Continuously comparing our setup to market products (e.g. Xiaomi or Dyson) improves component selection and ensures a well-working device later one.

**How to test**: Use a custom-made negative pressure test-bench.

**Quick example:**
- Use an absolute negative pressure sensor ([Link](https://www.adafruit.com/product/3965))
- Attach to the end of the tubing (either a turbine to test or a market product)
- Measure once without applied suction (to get the current ambient pressure)
- Turn on vacuum, measure again to get the absolute pressure
- Substract the reading from the measures ambient pressure to get a reliable value

---

### In-use suction power

**Why test:** Dust will clog up the filter and will reduce the suction power. Our setup will need enough power reserves to be still usable.

**How to test**: Use the method above in between cleaning sessions or implement a negative pressure sensor directly into the osh-vacuum prototype to measure suction power over time and during use (might not be needed in the final concept).

---

### Power consumption vs. suction power

**Why test:** Suction power has a direct relation to the motor's rpm setting. Higher RPM needs more current and thus increases the power consumption (and reduces battery life). We need to find a good balance (either fixed RPM), user changable settings or automatic adaption to different situations (nice feature, more complicated).

According to Xiaomi, their G9 vacuum will have the following battery life:

- Setting 3 (Full power, for carpets): **8 Minutes**
- Setting 2 (Medium power, general use): **28 Minutes**
- Setting 1 (Energy saving, light cleaning): **60 Minutes**

**How to test**: Run either from a battery or through a lab power supply. Imitate scenarios: e.g. putting a carpet in front of the hose, use a full dust filter etc. Calculate runtime by power consumption or do a long-time testing until the battery runs out.

---

### Suction power vs. overheating

**Why test:** We need to make sure that our vacuum runs safely without any risks. Vacuums are using airflow to cool internal components (motor, esc, motherboards, sometimes even batteries). We need to make sure that our electrical components, airflow and vent construction are keeping the motor from overheating during continuous use.

**How to test**: Use a motor with an included thermistor or include a 10K NTC thermistor in our early prototypes. Do long-term testing while measuring the temperature near any crititcal compoments. If the setup still tends to overheat after some time, we could implement automatic power reduction once a temperature threshold is reached into our final version (low complexity).

# Test-Bench

The current test-setup includes a mount/adapter that converts each motors air intake into the same diameter connector. This way we can attach different motors to the same hoses/cyclones and dust extractors later on.


We included a filter housing that utilizes one of the cheap filters supplied with the AliExpress vacuum machine. This is not for actual cleaning, but to protect the motors during testing from objects being sucked into them.

At this stage, there is no digital sensoring (yet).

### 3D-Files

- Onshape: [Public URL](https://cad.onshape.com/documents/5de0260cb3418d7c1923936b/w/3fbc0c31d3f577acbdce6c5d/e/b330385d27266245abf7c604)
- PLA

# Test-Bench 2.0

For comparable and valid results, we are upgrading the test bench to take the following measurements:

- Static negative pressure (fully covered air inlet)
- Air flow / speed inside of a 40 mm tube with different sized orifice plates
- Negative pressure in the same 40 mm tube different sized orifice plates

This allows us to calculate air-watts. It is **not** calibrated, but this should give us a great suction profile to compare vacuums and motors to each other.


3D-Data on [OnShape](https://cad.onshape.com/documents/44a1dede6692df0b4c97b65a/w/2b560eddae1e061ad9ced0ca/e/8ad2d1cabe2cca58175f6d20?renderMode=0&uiState=68e90a754d9953904ca08514)!

***IMPORTANT**. The following results are not measured with the new test-bench 2.0! We used manometer only (air-flow) and a sealed plug with the anemometer attached!:**

**Motor 1** = 135W / **Motor 2** = 350W

- The Xiaomi G20 has it's cyclone compartment attached. This should definitely influence the flowrate, but not the maximum negative pressure.
- The Xiaomi G20 ran out of battery right after the test session. (Repeat with a full battery next time)
- The other two motors were directly attached to the manometer which makes it a somewhat unfair comparison to the Xiaomi.

Static pressure measurements (fully sealed inlet):

- Motor 1: 20.5 kPa
- Motor 2: 21.7 kPa
- Xiaomi G20: 25.2 kPa

**Findings:**
- The 350W impeller is extremely powerful but seems to run hot rather quickly when fully sealed. We should really focus a motor with controllable RPM and use a NTC thermistor on the air outlet.
- The G20 heavily controls it's rpm based on the suction force and heat generation - within seconds.



# Test-Bench Details

# Ad-hoc test with analog manometer (Holger)

## AliExpress vacuum
:exclamation: Completely blocking the airflow puts stress on the motor. Block only briefly and make pauses in between.
- Media: https://albums.ente.io/?t=k2uqwNCXPj#DVDSANb38qgqW5a2wPTaPRknfDUz8maPe2q4zjmFuyJ5
- Test subject: https://github.com/hkienle/osh-vacuum/issues/2
- 2 power settings
- Manometer (analog): https://de.aliexpress.com/item/1005005951430776.html
- thread: 1/4" NPT
- Adapter (3D print): https://www.printables.com/model/1439982-adapter-for-manometer-14-npt-to-28mm-hosepipe-flex

### Result (static pressure)
- weaker power setting: ~10kPa
- 
- stronger power setting: ~15kPa
- 

# Cyclonev0.1 Simulation

Fine dust cyclones working as expected

Inflow is currently going directly into the fine dust cyclones which is not ideal

Mid three cyclones see most of the airflow which makes the consideration of their inflow more important, also a better distribution would be preferable

The dominance of the middle cyclones is visible even further on top

Another perspective of the inflow going mostly into the fine dust cyclones instead of the main chambre

Particle Trace

# Cyclonev0.2

## Simulation
https://www.simscale.com/projects/jadelsberger/osh-vac-twincyclonev1/