 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : Any ideas about limiting inrush current starting BIG ac motors?
  Page 2 of 3   |
|||||
| Author | Message | ||||
| Volhout Guru  Joined: 05/03/2018 Location: NetherlandsPosts: 5091 |
When the compressor is not running, and you apply power, it will consume 144A until the rotor starts to move. Your generator cannot supply that. Once the rotor is moving, the current goes down. Let's assume your generator can supply 240V at 10kW, that is 40A. To convert that into 144A (say 160) you need to halve the voltage in a transformer. So you take a heavy (say 8 or 10kW) conversion (auto)transformer from 240 to 120. With a relay your start the compressor at the 120V tap, once it is running you switch to the 240V tap. In essence you could remove the transformer once switching is done. A 8kW autotransformer, and a 40A mechanical relay (typically relays can handle 4 x I_nominal for seconds)should not cost you the top, especially not when bough on ebay. Success... Volhout PicomiteVGA PETSCII ROBOTS |
||||
CaptainBoing Guru Joined: 07/09/2016 Location: United KingdomPosts: 2170 |
very nice worked-up example. I think this is precisely what the OP is after |
||||
| Paul_L Guru Joined: 03/03/2016 Location: United StatesPosts: 769 |
@lizby --> Very nice job Lance! That should work nicely for your sump pumps, well pump, which are all fractional Hp motors running from your small portable generator. Now all I have to do is figure out what Triac will be beefy enough to carry the load of my monster compressor running from my 10KW natural gas generator bolted to a concrete slab, and how much drive current I will need to turn the bigger Triac on. Paul in NY Edited 2020-06-16 07:30 by Paul_L |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
Paul--one of the links I found on "Electronic Software-Driven Soft Starters" said their unit included a big capacitor (like a "hard start" module), in addition to the cycle slicing. For your unit, any idea of the size of the capacitor and when it would be cut into the startup sequence? PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
| Paul_L Guru Joined: 03/03/2016 Location: United StatesPosts: 769 |
It turns out that there are at least four gadgets that will do this job. First there was this from Hyper-Engineering selling for $259. https://www.electricgeneratorsdirect.com/Hyper-Engineering-SS1B16-32SN-V230/p99614.html?gclid=Cj0KCQjwuJz3BRDTARIsAMg-HxVrpG98EaboLPiZTuMJHoLySKfWQd2Q12e28IeSZKtKWGQYJqpDRDwaAh3oEALw_wcB Then Emerson began buying the Hyper-Engineering gadget and sells it for $568. https://climate.emerson.com/documents/css-compressor-soft-starters-product-leaflet-4-pages-en-gb-4849010.pdf Then Emerson apparently found a cheaper thing from Poland which sells for $292.43. https://www.ebay.com/c/8005779452?norover=1&mkevt=1&mkrid=711-153677-346401-4&mkcid=2&keyword=&crlp=307364301300_&MT_ID=&geo_id=&rlsatarget=dsa-19959388920&adpos=&device=c&mktype=&loc=9004846&poi=&abcId=&cmpgn=1615587334&sitelnk=&adgroupid=66643712052&network=g&matchtype=b&gclid=Cj0KCQjwuJz3BRDTARIsAMg-HxVBZdz82kGubpuUahfu35otwPmbjeVnt6Q4hXX2ZdBCL_aKQ7aTBj8aAqlYEALw_wcB Then a small company in NJ, MicroAir.net, re-engineered this and sells it directly for $338. https://www.microair.net/products/easystart-368-6-ton-soft-start-motor-starter-for-115-and-230v-ac-motors?variant=29181662283 All of these gadgets seem to adjust the timing of the Triac turn on point depending on the acceleration time of the individual compressor during the first four or five startup cycle of the individual compressor. They must somehow sense the current draw of the compressor, possibly with a toroidal core, and change the delay time to the initial turn on pulse which is then committed to memory somewhere. This is a practical and profitable product with a ready market among the owners of boats / campers / RVs / caravans which one of you younger guys might want to fool around with in an effort to make some bread for yourself or for the continued support of the development of the Micromite and Geoff's wonderful language. Lance is probably too old, and I am definitely way too old to get involved directly but I think somebody here should think about it. In the states the regulatory guys, Underwriter's Labs, (a private company owned by the fire insurance companies), only get involved with permanent installations of building wiring which is fed by utility mains. These gadgets, by definition, would only be used for limiting starting currents from private emergency generator supplies when installed in a building, or from generators installed in campers, RVs, caravans, or boats, which are not permanent buildings. It is likely that UL would only be involved in looking for obvious shock hazards and proper grounding provisions which would be much simpler than guaranteeing that the thing actually worked. I have no idea what regulatory or bureaucratic nonsense would be required in Oz, NZ, or the continent, but it definitely sounds like something someone would be interested in investigating. Paul in NY |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
Who you callin' too old? Oh, you mean too old to wish to be involved in product development. Well, ok, then. I envy the many people in OZ who appear to be able to just throw more solar panels at a problem of powering a home under varying conditions or in the event of loss of mains power. In Canada, we have not only the problem of solar panels being far more expensive, but even if we had the panels we don't have the sun when it's needed most. So I'm looking at generators and batteries. What I am definitely too old for is running around trying serially to hook up the most needed appliances to a small generator in a (probably stormy) time of emergency. I looked at a whole-house auto-start propane generator, but was quoted over $10,000 for a 10KW unit. Well, over a period of some months, I averaged about 24KW hours a day. So it seems that if needs could be properly sequenced, no more than a 3KW or 4KW generator should suffice. So there's load-shifting, and I've theorized about a round-robin sequencer to run sump pump, then well pump, then boiler, then fridge, then freezer, then repeat. I think it would work, and I've prototyped software and hardware but I'd like for someone else to (inexpensively) commercialize it, since as a DIY project, I'm afraid I'd end up making it more prone to failure. Maybe the whole house generator would be the best solution, especially in a time of plague, when we don't want to be forced to look for accommodations if we have an outage in the midst of a winter storm, when all we would absolutely need is the relatively small amount of power required to run the oil-fired boiler and associated pump. Then, too, the house was built in the 1840's and made to be heated with wood to a certain degree of comfort. I'm too old for that, too, handling wood in that quantity. PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 7937 |
I've not spotted this in any of the other posts, apologies if I've missed it. Some of the compressors I've come across start off-load by venting the output line back to the input via a solenoid valve or by using a mechanical system that lifts the valves in the compressor head. The solenoid is on a timer to let the motor get up to speed before applying load. Without this (or a similar system) compressors can be a pig to start as they have a tendency to stall easily. Generators are always a more difficult supply to start motors on. It's not unusual to need a set at least 3 times bigger than the maximum motor load. If you'd been able to start a 6kW set on a 10kW generator I'd have been very surprised. A soft-starter (not a VFD) is almost certainly the only way you'll manage this on a single phase system. Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
If you're familiar with soft-starters, can you explain what characteristics they should have? Are there refinements in the cycle slicing routine shown above (for instance, it would be possible to ramp from the peak of the AC half-cycle instead of the end of it); and do you know about the sizing and timing of adding a capacitor to the start-up routine. PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
Chopperp Guru Joined: 03/01/2018 Location: AustraliaPosts: 1097 |
Hi Paul Awhile back, matherp wrote a Cfunction to do this. Some blurb following. I used it & it worked quite well driving SCR's for battery chargers It's included in the MM downloads under Cfunctions Triggered Pulse.txt This function takes a trigger pulse on pin-15 (28-pin) or pin-42 (44-pin) MM2 and uses it to create an output pulse on a user-specified pin with a user-specified delay and of user-specified duration. See this post on The Backshed forum: http://www.thebackshed.com/forum/forum_posts.asp?TID=10572 You'll need to read the whole thread Brian Edited 2020-06-16 23:12 by Chopperp ChopperP |
||||
| Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 7937 |
Oh, I'm by no means an expert on them. :) I've come across them a fair number of times though as I used to design electrical control panels for a living. Basically they clamp the voltage, and hence the torque, during the startup sequence. That's almost always done with back to back thyristors (or triacs for smaller loads) triggered by a feedback system that's watching the motor current or back emf. They are taken out of circuit at the end of startup by shorting the power device(s) with a contactor. The power to the motor is controlled by the area inside the sine wave, between the zero crossing point and the cycle slice point. Hence you can have a leading edge control or a lagging edge control. In the former case the slice point moves up the leading edge with increasing power. In the latter case it moves forward up the lagging edge from the trailing zero point. There are benefits to both approaches. Leading edge control means that you close the control devices at zero current and open them later in the cycle. That tends to cause less interference. The lagging control means that you close onto higher voltage, so the torque is better. The control loop will have to have some sort of "slugging" - probably as you would think a capacitor would behave. That's because the motor needs *some* initial current simply to overcome the initial inertia of the armature. If there is too much slugging though the inrush current will take the overload/breaker out. A VFD can be used to soft start, but it does so by controlling the supply frequency. They won't always be as good as a soft start unit, it depends on the device. They will usually tend to be bigger, with bigger heatsinks on the ones that are left in circuit during normal running. You tend to pay a premium for them too! Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
I am a bit late to the party, but.... Volhout is right, from standstill its going to draw a 144 amp surge no matter what with 220 volts applied. But a soft starter introduces a whole lot of other problems. Refrigeration compressors usually start up with zero or almost zero pressure differential, and there is very little volume in the pipework and condenser. That means after only a very few revolutions, the compressor is going to have near full head pressure to start up against. If you try a very long slow soft start, the head pressure is going to very quickly build up, and it may never be able to reach anything like final running speed with only fractional voltage and power applied. Soft starting is really self defeating in a compressor application, and its not going to be a neat solution. What really large industrial compressors sometimes do, is unload the compressor by opening a path between suction and discharge, and then use a star delta starter to bring the huge machine up to full running speed, before closing the bypass. As you are in the US, the logical way to do it would be to first unload the compressor, start up initially with 110 volts, then switch over to the full 220 volts, then close the bypass solenoid valve, and you are away. Cheers, Tony. |
||||
palcal Guru Joined: 12/10/2011 Location: AustraliaPosts: 1993 |
The compressor should soft start OK, years ago I worked on GE split systems up to 2 1/2HP. It was mandatory in some states of OZ for a soft start to be fitted so we fitted them to all the units. You don't need a long slow start, with the ones I used you could hardly tell it was there. Many years ago I built a soft starter for a meat mincer and that was a long slow start to avoid belt slip and breakage. Edited 2020-06-18 12:03 by palcal "It is better to be ignorant and ask a stupid question than to be plain Stupid and not ask at all" |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Yes, it needs to be very fast. Cheers, Tony. |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
How fast is very fast? In the sample code I provided above, I stepped up from 323 microseconds to the full 8.33 millisecond half-cycle time in 250 microsecond steps with each step repeated 32 times, so the full ramp took about 4 seconds. Without the repeats, the ramp occurs in 26 half-cycles (as coded), or 216+ milliseconds. Assuming an attempt to start a 1/2HP well pump using a 1200 watt inverter (2400 watt surge) with a 12V battery, is 216ms fast enough, or is 4 seconds adequate, or somewhere in between? (I can't easily test the well pump, so I'm setting up to test with a 1/4HP pump that I put in the river so I can water the garden during the typical late August drought here in south shore Nova Scotia. Boiler heating and running water from a battery bank in a wintertime power outage is what I'm ultimately looking for.) What type of interference? Is there a significant difference in interference between turning on the triac at the zero-crossing point and turning it off at that point (or rather, the reverse, interference in turning off mid-cycle and turning on mid-cycle). PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
| palcal Guru Joined: 12/10/2011 Location: AustraliaPosts: 1993 |
I'm no expert on ac current but my take on this is that the inrush current is only there until back EMF is generated by the collapsing magnetic field to produce sufficient Impedance to limit the current. The DC resistance of the run winding in this case is probable around one ohm, so you can imagine on the first half cycle with no back EMF it behaves like a DC circuit hence the huge current flow. I suspect it does not take many cycles to build up the full back EMF and so the Impedance of the motor will rise from the original DC resistance of one ohm to perhaps 10 ohms (I'm plucking these figures out of thin air). I suspect that the fact the rotor is stalled at start up also causes some extra current until it gets moving. "It is better to be ignorant and ask a stupid question than to be plain Stupid and not ask at all" |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
Thanks, palcal. That provides a useful conceptualization. I'll see what my test rig shows in a couple of days. Not that I can look at the details of what you have described. PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
| Paul_L Guru Joined: 03/03/2016 Location: United StatesPosts: 769 |
We seem to be talking about apples and oranges here. Freon compressors come in different varieties. Scroll compressors start up almost totally unloaded because the moving scroll element drops away from the fixed scroll on a cam when it is stopped. They accelerate VERY quickly, probably within five rotations, then the scroll slowly lifts up and begins compressing the gas. My monster is a scroll with a locked rotor current of about 4.6 times the running current (140A vs 30A). They really don't need a phase shifting capacitor to delay the current sinusoid after the voltage sinusoid, but most of them come with a small capacitor. Piston compressors need a lot of torque to start. The piston must compress the gas on the first rotation. They might draw a locked rotor current up to 10 times the running current. The same situation exists with water pumps. You will find piston pumps, axial flow turbine pumps and centrifugal pumps. The piston pumps will have higher starting torque requirements and will draw higher locked rotor currents. Thyristors, Triacs and gated SCRs will all switch off and block when the instantaneous current goes to zero, probably a little after the line voltage goes to zero. They must be gated on sometime during each half cycle. With a 60 Hz line frequency the full cycle period is 16.66 ms and each degree of the full cycle has a period of 46.27 us. If you pulse the gate of a Triac at 80 degrees or 7.4 ms. into the half cycle the instantaneous line voltage available will be Sin(10) * (240 * (2)^-2) or 0.174 * 339.4 or 58.9 v and will decrease to zero in 462.7 us when the power line sinusoid goes to zero. Using my monster compressor as a load if we assume that the instantaneous current draw at the turn on point will be proportional to the 140 A current observed at 240 vrms then the instantaneous current at the Triac gating point will be 58.9v * (140A / 240v) or 34.35 A which will decrease to zero in 462.7 us. along with the power line voltage. A good initial condition for a soft start would be to trigger the Triac on during the first half cycle of the power line after a delay of 8.2 ms from the zero crossover time, and then decrease that delay by 0.1 ms during each successive power line half cycle (i.e. 8.1, 8.0, 7.9, ... 0.1, 0.0) it will require 82 half cycles of the power line, (or 82 / 120 = 683 ms), for the motor to receive full time power after which time the mechanical relay could be pulled in bypassing this soft starter. This would result in an initial instantaneous voltage of 2.828 v resulting in an initial instantaneous current of 1.650 A. My 10 KW generator should have no difficulty handling the resulting current ramp from 1.65 A up to 30 A. It should be noted that the mechanical relay contacts will never have to snap shut at the 90 degree point in the power line sinusoid with the motor stationary. That would be the worst case condition for the mechanical relay without the soft start circuit and would result in an instantaneous current of about 197 A with my monster compressor. In connection with Lance's fractional horsepower motors, say a 1/2 Hp, running on 120 vrms, will draw 373 W or 3.1 A running. If it is lightly loaded with a centrifugal pump it should have a nominal locked rotor current of about 6.8 A. Gating a Triac on at the 80 degree or 8.2 ms point would result in an initial instantaneous voltage of 1.414 v resulting in an initial instantaneous current of 0.235 A. In reviewing Lance's code for the Picaxe I see that he is using 32 steps starting with an initial delay of 320uS * 32 or 10.24 ms then ramps that in groups of steps over 4 seconds until he is delivering continuous power to the motor. I would think that a 4 second ramp might be too long and that the steps should be finer, but I'm unfamiliar with how the picaxe chip can time pulses. I know that the MMBasic PAUSE command will halt the running program in integer steps with a resolution of 1 ms., so it should be possible to have an initial delay of 8 ms, but it does not seem to be possible to decrease the delay in 80 steps of 0.1 ms. That is what led Lance to generating a group of gating pulses at one delay time, then generating additional groups of gating pulses at decreasing delay times. Has anyone got any idea how we could step the decreasing delay time in finer steps using MMBasic? Could we use a string of JK flip flops to divide the clock frequency down? Is the internal clock frequency of a MM chip available on an external pin? It would be ideal to have a 100 us clock pulse available. Lance's comment that the ramping of the brightness of a LED on the output of a bridge rectifier does not appear dramatic does not surprise me. A LED, once excited, will not terminate its light output very quickly ... it exhibits an afterglow instead of a sharp cutoff. A DVM on the output seemed to ramp more smoothly from 0V to 16V. Paul in NY |
||||
| lizby Guru Joined: 17/05/2016 Location: United StatesPosts: 3378 |
Paul--thanks for the detailed explanation. I think I'm coming closer to an understanding. Does that not mean that we need to be talking about slicing 8.33ms half cycles? Running the PICAXE at SETFREQ 16 (16mHz clock), PAUSE 1 takes 323 microseconds and PAUSE 2 takes about 587 microseconds, and each increment in the PAUSE number adds 250us. I guessed that this might be sufficient resolution, but if you believe that 100us steps would be better than 250us, there's the PAUSEUS command. PAUSEUS 1 takes 112us, and each additional PAUSEUS number adds about 2 and a fraction microseconds. With experimentation, I could get very close to 100us steps. Actually, I coded 26 250us ramp steps, with initial pause of (26*250us+overhead = about) 6.623ms into the 8.33ms half-cycle before turning on the triac. This is repeated 32 times and then 250us is subtracted from the delay time (and that delay time is repeated for 32 half-cycles). I will turn the process around to turn on the triac at the beginning of the half cycle instead of the end, and make 100us ramp steps with no repetition. This should complete to full power in about 83 half-cycles, or 2/3rds of a second. ~ Edited 2020-06-19 22:37 by lizby PicoMite, Armmite F4, SensorKits, MMBasic Hardware, Games, etc. on fruitoftheshed |
||||
| Paul_L Guru Joined: 03/03/2016 Location: United StatesPosts: 769 |
I don't think that will work. I believe that pulsing a Triac's gate will turn the Triac on after which it will then continue conducting until the current goes to zero at the end of the current half cycle. It then waits for another pulse to turn it back on. The problem is that with an inductive load (a motor) the current will not be in phase with the voltage. When the current reaches zero causing the Triac to switch off the instantaneous rate of change of the voltage across the Triac (dv/dt) will spike which may cause the Triac to switch back on prematurely when the differential voltage between the gate and the current terminal spikes. An RC snubber network may be needed to control the dv/dt spike the design of which will depend on the motor characteristics. Paul in NY |
||||
RFburns Regular Member  Joined: 21/07/2017 Location: AustraliaPosts: 43 |
This maybe of interest; a paper discussing IGBT and Thyreistor soft start complete with an IGBT single phase soft start .RF Soft start.zip Strong like horse smart like tractor! |
||||
| Page 2 of 3 |
|||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2025 |
