Fridge start current peak & soft start

Redpacket

Active member
Joined
Feb 28, 2018
Messages
1,534
So hearing the inverter pulse & seeing the current spike my fridge draws at start up got me thinking, and you know that feeling ;-)

The fridge says 430W 240VAC 0.7A on the label & it's only a 420L medium size one.
Can my inverter (Victron 5kVA Multiplus) handle it, yes - but I still don't like the startup peak.
This type of fridge has a single phase compressor with a "run" winding & a "start" winding.
Like many similar induction motors, the pulse of current flows when the start winding is given mains for a second or so to kick the motor off.

Here's the schematic:

image_wmsnbr.jpg

The PTC starter is the bit to replace.

So I started looking & saw there were a few like dometic & emerson units but they looked like >$400, too much for me.
https://www.actrol.com.au/product/universal-compressor-soft-start-32amp-2711261

image_papcks.jpg


then found this for an OK price:
https://www.ebay.com.au/p/6871A20052A-LG-Aircon-Outdoor-Pcb-soft-Start-Control/1273939409 ~$90 AUD

image_pfdboo.jpg

It should work like the emerson unit.
It's a simple looking device with only a few terminals but can't find the connection diagram for it (searched the a/c unit models this fits) so I'm going to have to check out the circuit vs the emerson unit before I connect it.

Wish me luck!
 
Good luck and please document how you make out. Ive been interested in lowering the peak spikes in my system as well. I have had it where my fridge and freezer both start at the same time and over load my 1000W inverter. I'm in the market for a new one but my normal loads are 100w so 1000W inverter should be fine
 
So the unit arrived today & connected it up with three incandescent lamps to safely check it out without blowing anything.
Here's what I've figured out so far -
The board pin out is like this:

image_bldpxn.jpg


It seems to make this waveform & switch the relays roughly like this:

image_uxftri.jpg


I'm thinking of connecting it like this (my fridge doesn't have a start capacitor, only a run cap):

image_sglxcu.jpg

image_gdplsv.jpg


I'll try & draw a diagram of the board later.
 
The waveform will still have a bit of a nasty peak current draw if it is switching at the top of the wave... Does it ramp up the current in this time - switch from 0% duty to 100% duty in a short time ? Thought the switching may have been more smooth rather than 50% wave output ?

Brilliant post though...
 
As far as I can see on the 'scope, the triac part just does a fixed duty cycle in the "chopped part" of the waveform for a few cycles (didn't count the exact cycles & the 0.7sec timing is a bit of a guess).
It didn't look like it was ramping up like you might find by eg turning up a light dimmer. More of a "stepped current" start.

Trouble with triacs is they're either off or on, there's no in-between like a transistor or mosfet can do - they are pretty robust devices though. This one is rated 25A, 600V & switching a stopped motor compressor without a heatsink!

The good thing with motors is they are inductive & so that sharp peak on the voltage won't = a sharp peak in the current, that'll lag behind a bit & be rounded off a bit.
There are the two NTC in-rush devices in parallel which if connected as I'm planning ATM, would also be limiting the start up current.

This board is simple, older tech - at least as it is now. I might see if I can make it ramp up with a modification. The chip on the top left looks like a simple CPU.
 
Interesting.. I was thinking of just the start coil resistance (partly ignoring the inductance) and the FET peak handling in some inverters that does not leave much room and only a few mS will blow them. The back emf should soften the peak though as you say. Suppose with 25A rated triac at the output (5.7kW @ 230V) when comparing input rating for an invertter at say 48V the aggregate FET ratings would need to be around 120A on a simplistic view.

I have an inbuilt soft start on a 3hp circular saw, which limits the start inrush to around 60A until the relay switches and that is when the actual spike occurs at 90A@48V, which surprised me a bit as I was expecting the peak before the relay switched. (230V saw through inverter - induction motor no separate start coil). Have you been able to measured the current profile through the switching process ?
 
I haven't got a spare compressor (yet!) to play with so haven't measured currents on a real motor yet. ;-)

The problem is there's no back emf at the beginning of the start up, this only builds up with speed. The initial current is winding resistance + inductance.

There's also the run winding which may or may not be getting power and being a heavy load itself too during start up. The way my fridge's compressor is wired, the run winding is already powered & the start winding is in parallel until the original PTC device drops the current. But the board don't appear to power the run winding until after start up via the start winding &/or start cap.

After start up in my fridge, the run capacitor is connected to the start winding. Other fridge/aircon units are likely to be different - per this board's apparent original connections.
 
Barron, It would be interesting to hear what your view is of them. The spec sheet looks interesting but does not say how the power limiting is ramped (guessing just chopping with progression to 100% duty) and the pedestal is a minimum voltage ?

I was contemplating using something like these - no adjustment, just cross your fingers the switching delay time works for the application and the NTC is good enough..
https://www.aliexpress.com/item/Ass...-Soft-start-Board-High-power/32826093896.html
 
BaronVonChickenPants said:
I have a similar issue with a water pump, pulls almost 7A at 230v to start then settles to 3A to run.

I purchased this unit:https://au.element14.com/united-automation/ssc-25/control-module-softstart/dp/1213104

The data sheet says all the right things but so far I haven't had much success, granted I've only spent 5 minutes fiddling with it.

What does the pump do with this unit in circuit?

Is this pump a "house water pressure"type pump eg with a reservoir tank?

I'd imagine you'd set the ramp up time to be short eg 2 secs max eg to get a "heavy" impeller + motor rotor up to speed pronto before the water pressure began to build too much.


I'm thinking that for my fridge, the compressor is small & the piston + rotor is not nearly as big as an aircon compressor would be.
So I'm wondering how to best mod this to do maybe 1/3 sec of chopped pulses, then switch to full power as per normal run conditions.

At the moment, just directly connected to the house mains via inverter (& previously on the grid), my fridge seems to start very quickly, just a clunk & it's running in way <1 sec, eg no grumble, then going.
 
Ok so I tried using PICAXE pre-programmed PIC chips (can use easy flowchart to code & easier programming) but they proved to be way too slow & not able to run commands needed in millisecond timings here.
I changed to standard PIC chips & settled on using an easy to get 16F84A that I have a few of.
I've managed to get a socket wired in just above the board & the pins connecting to where they need to.
Once I got the programmer hardware working (windows seems to have tightened some security stopping older chip programmers running), then I was able to test & debug some live code.
Now I've got the code running it pretty close to what I think I'll need.

It starts the triac at approx 40% & ramps it up quickly to 100% in under 1 sec, then switches on one of the relays & leaves it on.
So the circuit will end up feeding the run winding via the NTC first while the triac ramps up energy into the start winding, then cut off, & last the relay shorts across the NTC so full power t to the run winding + run cap connected as normal.

Bit more tweaking of the code in mind, then have to convince the wife it's time for some live testing on a fridge full of food - unless I can find a similar compressor to test on ;-)
 
Well i got around to trying the soft starter circuit live on the fridge.

So the original circuit has a base line like below (taken with a Mastech clamp/voltmeter).
I had to wait for the PTC device to "reset" (cool down again) & gas back pressure to drop between starts (did this most times...!).
I typically measured a peak current of ~7.4A at start, varied a bit, sometimes down to only ~5A.
When running normally, the current was ~0.7A per the nameplate.
- mains in to neutral was about 239VAC, 50Hz here (also = run to neutral)
- between run & start was ~267VAC
- between start & neutral was ~128VAC
(ie the capacitor was providing a phase shift & reduced voltage to the start winding)

During tests below when the motor attempted a start but stayed stalled, current was about 5.4A

The motor & capacitor arrangement in mine is called a "permanent split capacitor" (PSC)-style compressor.
Mine does have a "run" capacitor fitted (4uF, rated 470VAC).
In the original fridge circuit, the PTC device (when "cold") is low resistance & connects the start winding to the run winding.
The capacitor isn't involved yet as it's across the PTC (which is a short at first).
The PTC quickly heats (in approx 0.5 sec) & goes open circuit leaving the capacitor there.

Here's a link to a similar PTC device: http://www.carbonbrush.com/8ea.pdf

Trouble is, this is quite different to how the ramped start would be working (or even the soft starter board's original working as posted above).

So I connected up the soft starter "Mk1".
The soft starter circuit put an NTC (resistance starts higher, goes low) in series with the run winding then shorting that with a relay after ~0.4secs.
At the same time it ramped up the start winding phase from about 40% to full wave, then going open to leave the capacitor there just like the original circuit.
I tried multiple starts & made the ramp time quite fast (~0.4 seconds) - it started only properly once out of at least about 10 attempts - not good!
Because the compressor is rubber mounted, I could see it "kick" by varying amounts as it tried to start.
It also did a weird one where it grumbled to a rough shaking but was not "running" - turned that off pronto!....
Had to give it a few minutes to cool between attempts & did click its thermal overload a few times.
Unfortunately it seems this type of PSC motor does not like a ramped up start - at least not how I've tried so far!.
Bit disappointing, but hey, learning ;-)

So now I'm considering should I leave the start winding connected at 100% for a short while (eg 0.5 sec) before turning off the triac?

And also thinking about how sometimes when it starts (original circuit) it doesn't seem to dip the lights & other times it does.
It's operating just by simple thermostat contact closure normally so "luck of the draw" as to phase of the mains when juice is applied.
I think I'll try modifying the connections & CPUs program so it can control the phase point on the AC waveform that it attempts to start via the original PTC circuit & see if starting at eg zero crossing is best or maybe some other position on the cycle gets a reliable start with least "current pulse" impact.

I might also try ramped up phase into the whole original circuit too.
 
Redpacket said:
Well i got around to trying the soft starter circuit live on the fridge.

So the original circuit has a base line like below (taken with a Mastech clamp/voltmeter).
I had to wait for the PTC device to "reset" (cool down again) & gas back pressure to drop between starts (did this most times...!).
I typically measured a peak current of ~7.4A at start, varied a bit, sometimes down to only ~5A.
When running normally, the current was ~0.7A per the nameplate.
- mains in to neutral was about 239VAC, 50Hz here (also = run to neutral)
- between run & start was ~267VAC
- between start & neutral was ~128VAC
(ie the capacitor was providing a phase shift & reduced voltage to the start winding)

During tests below when the motor attempted a start but stayed stalled, current was about 5.4A

The motor & capacitor arrangement in mine is called a "permanent split capacitor" (PSC)-style compressor.
Mine does have a "run" capacitor fitted (4uF, rated 470VAC).
In the original fridge circuit, the PTC device (when "cold") is low resistance & connects the start winding to the run winding.
The capacitor isn't involved yet as it's across the PTC (which is a short at first).
The PTC quickly heats (in approx 0.5 sec) & goes open circuit leaving the capacitor there.

Here's a link to a similar PTC device: http://www.carbonbrush.com/8ea.pdf

Trouble is, this is quite different to how the ramped start would be working (or even the soft starter board's original working as posted above).

So I connected up the soft starter "Mk1".
The soft starter circuit put an NTC (resistance starts higher, goes low) in series with the run winding then shorting that with a relay after ~0.4secs.
At the same time it ramped up the start winding phase from about 40% to full wave, then going open to leave the capacitor there just like the original circuit.
I tried multiple starts & made the ramp time quite fast (~0.4 seconds) - it started only properly once out of at least about 10 attempts - not good!
Because the compressor is rubber mounted, I could see it "kick" by varying amounts as it tried to start.
It also did a weird one where it grumbled to a rough shaking but was not "running" - turned that off pronto!....
Had to give it a few minutes to cool between attempts & did click its thermal overload a few times.
Unfortunately it seems this type of PSC motor does not like a ramped up start - at least not how I've tried so far!.
Bit disappointing, but hey, learning ;-)

So now I'm considering should I leave the start winding connected at 100% for a short while (eg 0.5 sec) before turning off the triac?

And also thinking about how sometimes when it starts (original circuit) it doesn't seem to dip the lights & other times it does.
It's operating just by simple thermostat contact closure normally so "luck of the draw" as to phase of the mains when juice is applied.
I think I'll try modifying the connections & CPUs program so it can control the phase point on the AC waveform that it attempts to start via the original PTC circuit & see if starting at eg zero crossing is best or maybe some other position on the cycle gets a reliable start with least "current pulse" impact.

I might also try ramped up phase into the whole original circuit too.

maybe the motor is cogging due to phase shift?
 
rtgunner said:
Redpacket said:
Well i got around to trying the soft starter circuit live on the fridge
........
I might also try ramped up phase into the whole original circuit too.

maybe the motor is cogging due to phase shift?

Did you have any more info you could share on that? Eg explain what you're thinking a bit more? - I'm still figuring out how it works, especially at start up!
When running normally, I like found above, the start winding does have quite a bit of phase shift on it.
The pdf I linked (& others I'd read) also talk about phase shift during start but then this motor has the start & the run windings directly in parallel at least initially (ie no phase shift in the applied power).
So maybe the start winding is physically offset vs the run winding?
Or maybe the magnetic field moves differently/lags differently at startup?

Any thoughts would be welcome....!
 
Back
Top