Beiträge von Fred17

The proper solution is a "hybrid relay" -- where (for a non-inductive load) a solid state relay (triac) is turned on at zero-cross and then an electromechanical relay is closed -- and then when the circuit is desired to be opened the relay coil is deactivated and then after a suitable delay the triac is turned off. This prevents arcing at the relay contacts and prevents heating from the triac forward voltage drop.

If you want the ultimate in reducing power consumption, you can use a latching relay rather than one that requires the coil to be continuously powered. This is easy with microprocessor control but can also be accomplished with appopriate RC delay circuits.

Of course space constraints may make this difficult in a small shelly.

Just add an external 200k resistor and everything would be fine in switching to N?

I just remembered the Shelly Plus Uni also has a one-wire connection. So that and a DS28E18 or DS28E17 would seem to do the trick without requiring an add-on module. If you had the need for controlling several relays at the same time one could use one of the Shelly X boards instead. It seems there are break out boards with the DS28E17 available on Aliexpress.

Yes one would have to use an ESP32 with home assistant rather than a shelly to connect directly to i2c. I looked at the Shelly X-MOD1. It looks like it exposes a UART and 8 digital IO pins so if there were a digital pot with a parallel interface the io pins would work. I think Maxim or AD used to make such things but I don't see any now. However the shelly add-on and the X-MOD1 do have a 1-wire bus output, and so in theory one could use the (surface mount) DS28E17 or DS28E18 to convert 1-wire to i2c and thus connect the digital pot to 1-wire. Obviously this would also open up lots of other i2c devices for use with shellies.

JSE222 Not sure what is obsolete. There is a small breakout board from Adafruit (also available on DigiKey) that carries the DS3502 digital 10k pot --controlled by i2c with four selectable addresses. Allows analog voltage up to 15.5 volts.

https://www.adafruit.com/product/4286

It's probably obvious to others but it just occurred to me that another route to the same capability is to use the Uni Plus as a relay driver.

You could even use a latching relay and then have essentially no power usage other than the Uni. The Uni would provide a set or reset pulse to one or the other coils of a dual coil relay (or opposite polarities for a single coil latching relay) and the relay switches.

And if you want to have a SPDT changeover or DPDT or larger relay (high or low voltage on the contacts) with a shelly driven by a line voltage supply, you can use a small approved power supply to power the Uni with 5 or 12 volts (or the reworked Shelly 1 Plus with 12 volts). An example of a relatively inexpensive smallish power supply that would work is the MeanWell IRM series (for example the one watt IRM-01-5 or IRM-01-12).

Of course the packaging and connections are the responsibility of the user.

thgoebel Thanks for the information. I agree the creepage is insufficient.

Re SELV on the Gen3 mini-- Shelly statements are very confused -- The website says "Dry Contact: Allows switching on and off of lower voltage devices." But as dewaldo points out, the printed documentation says

⚠CAUTION! No SELV/PELV circuits may be connected to the terminals of the inputs and outputs, including the extension inputs.

⚠VORSICHT! An den Klemmen der Ein- und Ausgänge, einschließlich der Schaltereingänge, dürfen keine SELV/PELVStromkreise angeschlossen werden

Is it possible that it might be safe for ordinary low voltage circuits but not SELV?

If the insulator in the smallest gap shown in thgoebel 's teardown of the Mini is ceramic, then the mini might actually be safe for 120 volts (not 240). Better safe than sorry.

Very nice! Thank you thgoebel for this very clear presentation.

This was shared by the Doug Roberson, the Shelly USA CTO, in the Shelly USA reddit.

Seems odd that Dimmer 2 (Gen 1) works fine in the US, as does Dimmer 1 PM Pro and Dimmer 2 PM Pro.

I can't really think of any reason why, except perhaps

1) a design mistake made it not work at 60 Hz, or,

2) design choices that allowed 200 watt load at 230 volts produced a current-limited design that would only allow dimming 100 watts at 120 volts. ANd this seemed too puny to market.

Some time back I bought a dozen of the Plus 0-10 v dimmers to control Meanwell PWM LED drivers. The Meanwell dimming input (according to the datasheet) accepts a variable resistance (so it sources voltage) and sinking PWM from the dimmer but also accepts what it calls "additive DC voltage" and "additive 0-10 volt PWM" on the same input (not at the same time, of course). In other words it also sinks sourced PWM and DC voltage as additional control modes.

I want to be energy efficient and not have the LED drivers energized by 120 volts when they are not in use (even though standby power draw for each dimmer is less than 1/2 watt for each driver. ) Unfortunately the Gen 3 1/0-10 v dimmer is not available in the US and in any case I already have the Plus sinking version. I don't need power measurement or want the extra heat from a shunt resistor inside the Shelly case (even 1 milliOhm). The dimmers are in a central panel so it's no problem to add extra components. Assuming it works the components for the solution below will cost under $5 per dimmer/driver, including the LDO 10 volt regulator that cost under a dollar each. (It's rated at 150 ma so I don't need one for each dimmer/driver. Probably I'll put a small resistor on its output to ensure it's not overloaded if mutltiple dimmer/drivers are activated at once).

I have Panasonic 2 Amp Solid State AQG-22105 Rectifiers that can be turned on by 4-6 volts DC. They have an input imedance of 330 ohms (rather low I thought) and seem to want 10 ma at 5 volts. I also have a source of 10 volts to convert the Plus 0-10v dimmer to a source of PWM voltage. I can't use the PWM drivers as a source of 10v because they will be off normally when the Shelly dimmer wants to turn them on.

This is a little complicated to arrange given what thgoebel in the Plus 0-10v teardown thread advises regarding placing a 3.3kOhm resistor in series with the 10 volt source. Using the usual RC network with this resistor in series would seem to delay the turning on of the relay through the usual sort of RC network that is also needed to keep the relay turned on during as little as a 6% PWM duty cycle. (The Meanwell drivers dim to zero below 6% PWM). I explored using a PNP transistor fed with the inverted PWM signal from across the 3.3k resistor as suggested by thgoebel but I was not confident I could get the timings right and avoid all the drivers being turned on briefly on system power on -- also there would be a constant current flowing in the off condition in most of the configurations I examined, negating the purpose.

So to isolate the RC timing from the 3.3k resistor I decided to use a CMOS integrated circuit buffer -- 1/6 of a hex CD74HC4050E that happily accepts more than 10 volts on its input. It outputs 5 volts at up to 25 ma for charging a 2 microfarad capacitor through a 220 ohm resistor. Then this voltage is fed through a 2.2kOhm resistor to the base of an NPN transistor which is placed between the - terminal of the SSR and ground. The transistor only needs a milliamp at its base to cause it to reliably conduct the 10 ma needed for the relay. I've checked that there will be enough voltage and current at the transistor base to keep it on during the minimum 6% duty cycle and the relay should turn on within a full cycle at 60 Hz (one half cycle for the transistor to conduct even at the worst case 6% PWM input, and another to be sure the zero-cross relay turns on). This seems like a tolerable delay.

Here's a diagram. I'll report back after I've tested it.

.

Just by coincidence I happened to see this post on reddit in r/homeassistant.

It uses an access point set up on a wifi enabled computer running home assistant to keep contact with shellies when the main wifi fails. So this may be of interest to you if you are willing to consider installing home assistant as well.

https://www.reddit.com/r/homeassistan…g_in_case_of_a/

If you know your way around electronics you can use an optoisolator (for example PC817) so that the low voltage switch does not see any high voltage from the shelly. I do this with the Shelly Pro dimmer. Would also need a series resistor to limit the current through the optoisolator's led. But that's it.

Thanks!!!

Sorry if i wasn't clear. I was asking about the new I4 Gen3 version. I knew about the Plus I4 teardown. Perhaps you are saying that it's equivalent in this way to the Plus I4.

It would be good to know. The specs seem to suggest AC only operation.

Thanks!

I need to replace fuse F1. Does all the load current go through this fuse? in this case I can calculate an appropriate value based on my expected loads. Or does it only protect the electronics in which case it should be smaller. It seems to have died when I had a load connected that was only supposed to be 50 watts (120 v) but it may have had a lot of inrush current (LED transformer or driver) or something else may have been wrong with the wiring of the load. (Chandelier was installed by an electrician who wasn't available when I connected dimmer).

Thanks! I suppose that if one were willing to do some soldering one could supply it with 12 volts, similar to what has been shown for some of the relays.

Has anyone tested the minimum DC voltage needed across L and N (with N positive) for the 0-10 v dimmer to work reliably?