DC power problem (devices losing power intermittently)

Forums › Gadgets, robotics, home automation, electronics (including wearables) › DC power problem (devices losing power intermittently)

SirHumphreyAppleby

2844 posts

Uber Geek

#293226 7-Jan-2022 17:25

A while back I posted asking for assistance with supplying DC power to a greenhouse. I found some inexpensive (relative term) wiring to achieve a <3% voltage drop over the required distance, and all seemed to be going well until I got it all wired up. Trouble is, I am now seeing power interruptions which at a minimum take out the temperature sensor for a second, and at worse causes the system to reset multiple times in a row.

The greenhouse has three independent systems which run off a single 12V supply. These include a timer (mosfet) which controls a solenoid (<1.7A), a temperature sensor (relay) which turns on a fan (0.15A) and switches a DPDT relay to open a window using a small linear actuator (0.4A), and a pressure switch (relay) which controls a secondary relay that switches a pressure pump (10.5A - run well below max pressure, so this is probably a significant over-estimate).

I have had the problem occur with the timer, temperature sensor and solenoid disconnected from the system, so I'm rapidly running out of parts to debug. I have also checked all the connections, and everything (now) appears firm. The only commonality between when the faults occur seems to be a change in load, although at times it appears to be random. I have managed to get the issue to occur multiple times triggering the temperature sensor, which only slightly increases the load, and multiples times when the solenoid switches off (never on). Oddly, the biggest load is the pump, and that doesn't seem to cause a problem switching on or off (possibly once switching off). The pump is by far the biggest power user and has a direct connection (via the relay) to the bulky wiring coming into the greenhouse, so that may explain why that doesn't cause an issue.

No photos of my messy wiring, but I can get some tomorrow if they would help.

Does anyone have any theories on what may be happening here?

I note the power supply documentation states "Do not switch Secondary feed". This wasn't in the online description when I ordered, but could there be something the power supply is doing that is interfering with my use case? It is rated for 0mA up to the stated maximum, so it should be fine when the load is very light.

1 | 2

1063 posts

Uber Geek

Trusted

#2844957 7-Jan-2022 17:32

What are the specs of the power supply?

SirHumphreyAppleby

2844 posts

Uber Geek

#2844959 7-Jan-2022 17:37

Mehrts:

What are the specs of the power supply?

12V DC 264W

richms

28172 posts

Uber Geek

Trusted

Lifetime subscriber

#2844960 7-Jan-2022 17:38

If the power supply has over voltage protection and you are switching off an inductive load then its back EMF kick can be enough to trigger that protection on some power supplies and have them shut down till it resets. Make sure that any motor or solonoid has a diode across the relay or transistor driving it to control the spike at switch off.

There may be issues with RF on the cables if they're long from other nearby loads switching on or off causing something to trigger protection. Or it could just be a bad connection.

Too hard to remotely troubleshoot without pictures or details of what the gear is and how its connected.

Richard rich.ms

Ge0rge

2052 posts

Uber Geek

Trusted

Lifetime subscriber

#2844961 7-Jan-2022 17:38

I wonder if it's something to do with a back-emf, being that it seems to happen when solenoids are turning off?

edit - @richms beat me to it.

richms

28172 posts

Uber Geek

Trusted

Lifetime subscriber

#2844964 7-Jan-2022 17:41

Ge0rge: I wonder if it's something to do with a back-emf, being that it seems to happen when solenoids are turning off?

edit - @richms beat me to it.

Must be by nano seconds because I hit submit and the page came back with my post and yours below it. Photo finish type situation.

Richard rich.ms

SirHumphreyAppleby

2844 posts

Uber Geek

#2844966 7-Jan-2022 17:50

richms:

Ge0rge: I wonder if it's something to do with a back-emf, being that it seems to happen when solenoids are turning off?

edit - @richms beat me to it.

Must be by nano seconds because I hit submit and the page came back with my post and yours below it. Photo finish type situation.

You were both pretty quick.

The components I am currently using are these. If they worked well, I intended to buy replacements for when they died. If they are possibly causing issues, I will find alternatives.

Timer... (Note, this is solid state, so does a diode do the same thing here?)

https://www.aliexpress.com/item/4000106760211.html

Thermometer...

https://www.aliexpress.com/item/4000111155904.html

Pressure switch...

https://www.aliexpress.com/item/32957301645.html

Solenoid... (A bit beefier than the crappy little plastic ones)

https://www.aliexpress.com/item/4000268827489.html

The supply to the electronics and relay control for the pump run in parallel for around a meter, but otherwise the rest of the cabling is either separated or very short. I will post photos tomorrow.

richms

28172 posts

Uber Geek

Trusted

Lifetime subscriber

#2844973 7-Jan-2022 18:06

I cant see any diode across the output on the first one where the board is visible. You may see talk of a body diode in the mosfets but that is not appropriate for shunting the back EMF as the back EMF diode has to go across the load in reverse polarity. Ideally as close to the load as possible so that you minimize the radiation from the supply cables to it.

Richard rich.ms

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

SirHumphreyAppleby

2844 posts

Uber Geek

#2844974 7-Jan-2022 18:08

richms:

I cant see any diode across the output on the first one where the board is visible. You may see talk of a body diode in the mosfets but that is not appropriate for shunting the back EMF as the back EMF diode has to go across the load in reverse polarity. Ideally as close to the load as possible so that you minimize the radiation from the supply cables to it.

Thanks. The EMF theory sounds like the most likely. I'm looking in my box of tricks to see if I have any diodes (doesn't look likely), so I will pick some up tomorrow if possible and give it a go.

SirHumphreyAppleby

2844 posts

Uber Geek

#2844982 7-Jan-2022 18:29

No diodes in my box. Any suggestions on appropriate parts? 1N4001?

There are lots of recommendations online for solenoids and motors, and almost as many comments saying they are wrong.

k1w1k1d

1519 posts

Uber Geek

#2844986 7-Jan-2022 18:45

Fit diodes across all the inductive devices in the circuit. Don't forget the relay coils. Fit the diodes as close to the device as possible.

I use 1N4007 diodes for low current devices, and 1N5408 diodes for devices that draw more current. Both are available from Jaycar.

richms

28172 posts

Uber Geek

Trusted

Lifetime subscriber

#2844989 7-Jan-2022 18:58

1n4007 is all you will need for small loads, the 1n4001 would do for very small loads like relays. If the load is huge like a motor than the 4007 will probably do fine since it has a high surge rating but as you are not designing for manufacturing you dont need to cut costs so going for something larger would be a good idea for piece of mind only.

Look the part numbers up on element14 or somewhere to get an idea of the real specs of them for surge capability as that's what you are using for back EMF on something seldom switched. Usually no need to go to a massive high current rectifier diode for something that's hit seldom so will never heat up from it. The power dissapation and similar specs are more important when they're used for rectification with a constant current, not a small one that dies out in milliseconds.

To be honest I have just pulled diodes out of old power supplies and stuff and used those when needed. All seem to work ok. Not sure if using a schottky yanked out of a cheap chinese non compliant wallwart power supply matters or not, but it solved the problems so I call that a solution.

Richard rich.ms

SirHumphreyAppleby

2844 posts

Uber Geek

#2846811 11-Jan-2022 16:49

I managed to get my hands on some 1N4007s and installed them across the solenoid, the coils for both stand-alone relays, on the load side for the pump relay and even the fan. That, unfortunately, didn't resolved the issue. I didn't put one on the linear actuator as that depends on polarity for direction, so it would have go to on the supply side of the relay. It's also controlled by internal limit switches, so can't be what's initially triggering the issue.

I'm still seeing issues when the solenoid switches off, but also the occasional issue when nothing should be happening. This appears almost random.

The only thing I can think of is the proximity of the coil to the rest of the setup, but that wouldn't explain the random issues.

I've triple (or more) checked the wiring and everything is tight and as short as possible. Of course, I'm a software person, so I wouldn't trust that means they're connected well.

Below is how I have it setup at present. The wiring that disappears out the top goes to the fan and linear actuator. The wires at the bottom are supply and output to the pump relay. The black wires seen at the bottom are just the thermometer wire.

EDIT: Now that it's cooler, I can also confirm the issue is still occurring when the relay switches on to open the window using the linear actuator. This stays on until the temperature drops.

MadEngineer

4274 posts

Uber Geek

Trusted

#2846848 11-Jan-2022 18:55

What was the distance for the power? Depending on the distance I’d be considering a higher voltage transmission then a regulated power supply local to the gear. That would reduce voltage drop over the cable.

You're not on Atlantis anymore, Duncan Idaho.

SirHumphreyAppleby

2844 posts

Uber Geek

#2846851 11-Jan-2022 18:59

MadEngineer: What was the distance for the power? Depending on the distance I’d be considering a higher voltage transmission then a regulated power supply local to the gear. That would reduce voltage drop over the cable.

I looked into that option and even asked on here for advice. The distance is around 7m. The supply cable is 10mm^2 copper, which should keep the voltage drop to around 3% even with the pump running at its maximum rated current. The issue occurs without the pump running, so voltage drop should be negligible.

SirHumphreyAppleby

2844 posts

Uber Geek

#2851090 15-Jan-2022 10:07

I ended up replacing some of the interconnecting wires with thicker copper wire. This seems to have greatly reduced the issue. I've only observed one random glitch in 1.5 days (not watching constantly), but it's not repeatable like it was previously. Hopefully replacing a bit more of the wiring (for the solenoid and relays) will completely eliminate the issue.

1 | 2