Motor start capacitors?

The 250 volt rating is the max RMS voltage applied to the terminals of the capacitor. That is a typical rating for a 120 volt motor, and 330 volt is common for a 240 volt motor. The start windings produce a higher voltage than what is applied to the motor, and that is what the capacitor needs to be rated for.

I am not sure of the relative reactances, but I believe that the capacitor produces a large percentage of the phase shift. That is the reason that electrolytics are used for starting, since an oil-filled capacitor with high enough capacitance would be physically prohibitive.

I agree with you that this sounds like a start switch problem. I refered to a starting relay in my earlier post, but your motor may have a centrfugal switch. If those contacts don't open fast enough, it will kill the capacitor. A typical rating for start capacitors is "3-second start, 20 starts per hour"

Electrolyte

None, I don't think Digikey stocks capacitors intended for current handling purposes. That's why you can't find current ratings. There are all kinds of high power, high current caps but you won't find them at Digikey, Mouser or any other electronic supply house. Caps for motor starting, motor running, power factor correction, high current RF work, energy storage, and a whole host of power applications are available from suppliers in those fields.

Hmmm? I never thought of it as a resonator. A simple calculation shows the impedance to be about 19 Ohms. This means the winding resistance will likely swamp the inductor reactance and soften any resonance (low Q). I don't know what the typical winding inductance is, I suspect it would be higher than

50mH. What we need is actual values from a real motor of L, R and C to answer the question.

Of course if it were truly in resonance there would be "0" deg of phase shift and the field would not rotate. That means it has to be off resonance to get phase shift.

In a so called split phase motor there is no capacitor and all required phase shift is provided by the starter winding's L and R.

I get most of that sort of thing from a place called Surplus City, I forget the URL but you can google it, they sell a lot of HVAC stuff. Need to email or call for a customer number to get the catalog but anyone can do it. They have a huge selection of dirt cheap motor start and run caps.

| None, I don't think Digikey stocks capacitors intended for current handling | purposes. That's why you can't find current ratings. There are all kinds of | high power, high current caps but you won't find them at Digikey, Mouser or | any other electronic supply house. Caps for motor starting, motor running, | power factor correction, high current RF work, energy storage, and a whole | host of power applications are available from suppliers in those fields.

Not even for construction of (large) PSUs (AC to DC) and inverters (DC to AC)?

---------------- Actually the capacitor may be such as to bring the start winding somewhat leading. Ideally the start winding current should be in quadrature with the main winding and as the main winding current at start would typically be about 75 degrees lagging, the start winding + series capacitor should be leading by about 15 degrees. In practice, the actual capacitance would likely be matched so that the start winding is still inductive (lose some torque but stay away from possible resonance) . In this case, a larger capacitor may be the wrong choice. In any case the capacitor and the start winding are not designed for continuous duty. Some motors have start/run capacitors which are designed for near quadrature currents at start- and then partially switched out so that near quadrature conditions occur at low slip.

As others have suggested, try a capacitor designed for motor starting duty and also check the common cause of failure- the centrifugal switch. --

Don Kelly snipped-for-privacy@shawcross.ca remove the X to answer

Hey, if you know the capacitance and the voltage (& frequency) then the current is also known. If a cap is rated for the frequency and voltage then it can handle the current. The reason that the AC electrolytics are rated as either "starting" or "running" is that the "starting" caps can only handle the voltage (& current) for a short time before they start getting warm/hot. "Run" type caps can handle the current indefinitely. IOW: you can safely hang them across the AC mains (with some overcurrent protection in case the cap fails.)

** Posted from **

Except that neither 'starting' or 'running' capacitors are applied 'across the mains'. They are both in series with one of the motor windings and don't 'see' full line voltage.

daestrom

On Tue, 13 May 2008 11:10:14 -0400 John Gilmer wrote: | | |>

|> None, I don't think Digikey stocks capacitors intended for current |> handling |> purposes. That's why you can't find current ratings. There are all kinds |> of |> high power, high current caps but you won't find them at Digikey, Mouser |> or |> any other electronic supply house. Caps for motor starting, motor running, |> power factor correction, high current RF work, energy storage, and a whole |> host of power applications are available from suppliers in those fields. | | Hey, if you know the capacitance and the voltage (& frequency) then the | current is also known. If a cap is rated for the frequency and voltage | then it can handle the current. The reason that the AC electrolytics are | rated as either "starting" or "running" is that the "starting" caps can only | handle the voltage (& current) for a short time before they start getting | warm/hot. "Run" type caps can handle the current indefinitely. IOW: you | can safely hang them across the AC mains (with some overcurrent protection | in case the cap fails.)

Capacitors aren't rated as to frequency, either. So for a capacitor, that is not known. A capacitor used for smoothing DC would have to have a high voltage for the DC, being able to carry a high charge voltage, but yet not have a lot of current due to limited ripple at that point. You can build at capacitor for a given value of microfarads and a given peak voltage, with varying sizes for heavier or lighter conductors (and conducting foil inside).

There are some very high farad capacitors intended to store power in parallel with batteries to allow changing the batteries. They work on CMOS or other circuits that draw very little power. Apparently my universal remote control for the TV has one, as it says I have 10 minutes to change the batteries, but if I press any button without batteries present, the stored codes are lost. The circuits I have seen for these have resistors between the batteries and capacitor to avoid them charging up too quick if batteries are inserted with no charge. Sounds like capacitors with extremely thin plates that cannot take a high current.

Actually, they might see voltages even higher than line voltage because they resonate with the inductance of the motor winding. It's up to the folks who make the motor and starter to determine what voltage cap is necessary. But if a cap is rated for a certain voltage, it's designed to take that voltage.

Most "run" caps are rated will above 240 VAC and can be placed across the mains.

** Posted from **

Actually, many caps ARE rated as to frequency. It's a backwards way of rating the cap for current. The higher the frequency at the same voltage, the greater the current and the greater the heat build up in the cap.

Caps designed for ralatively low frequencies will overheat at higher frequencies. Folks who design switching power supplies soon learn this.

** Posted from **

|> Capacitors aren't rated as to frequency, either. | | Actually, many caps ARE rated as to frequency. It's a backwards way of | rating the cap for current. The higher the frequency at the same voltage, | the greater the current and the greater the heat build up in the cap. | | Caps designed for ralatively low frequencies will overheat at higher | frequencies. Folks who design switching power supplies soon learn this.

But I don't see capacitors rated this way in the catalogs from which they can be selected. If I have a projects, such as a PSU design, and need to have a capacitor of a specific current, or need to adjust the design around available capacitors, how can I know what their current maximum is?

If you are only building a handful of units for non-critical applications you might just use the "smoke test." If you are designing for a longer production run it might pay to varify with the manufacturer that the cap is suitable for the application.

When in doubt use a conservative design which in the case of caps mean you a significantly higher voltage rating than you calculate or measure. When you have things working you run an extended "smoke test" than then shut down and give the components the "digital temperature test." If you say "ouch" when you touch a component, it may be running a little hot but not necessarily and that's why you read the spec sheets.

You will often see "hints" in the suggested applications section of the data sheets or catalog description such as "suitable for switching power supplies" or "recommended for capacitor input filters."

You have to pay attention to "frequency" effects as you go to RF as at some point the inductance of the leads will resonate with the capacitance. All this is why when you want a power supply designed you go to a guy who has done it before, etc.

Just for fun, I dug up a "control unit" for our old deep well pump that failed. This is just a motor starter with a current relay. Interesting enough, this cap isn't marked as a starting cap so I suspect that many replacemnet caps are only sold in the "run" version. The label includes:

88-108UF 330 VAC 50/60Hz +65C. Clearly this cap is designed to not fail pre-maturely when it runs a little warm. "Doing the math", 65C is 65*(9/5)+32=149F. That's up in the "ouch" range but just barely. Above 140F is considered dangerous for domestic hot water. ** Posted from **

| If you are only building a handful of units for non-critical applications | you might just use the "smoke test." If you are designing for a longer | production run it might pay to varify with the manufacturer that the cap is | suitable for the application. | | When in doubt use a conservative design which in the case of caps mean you a | significantly higher voltage rating than you calculate or measure. When | you have things working you run an extended "smoke test" than then shut down | and give the components the "digital temperature test." If you say "ouch" | when you touch a component, it may be running a little hot but not | necessarily and that's why you read the spec sheets. | | You will often see "hints" in the suggested applications section of the data | sheets or catalog description such as "suitable for switching power | supplies" or "recommended for capacitor input filters." | | You have to pay attention to "frequency" effects as you go to RF as at some | point the inductance of the leads will resonate with the capacitance. All | this is why when you want a power supply designed you go to a guy who has | done it before, etc.

The idea I had in mind was designing a DC to AC inverter for whole house use that would support a high enough fault current to ensure the breaker would trip on a short circuit. It would have to be able to deliver the high current briefly, but that would not be a long term expectation.

I suppose commercial inverter designers don't just pick caps out of a Digikey catalog.

Designing a large inverter is not a trivial task, it's a case where you'd be better off buying a commercially made unit, it's a much safer approach at any rate. I'm by no means trying to discourage one from designing and building their own hardware, but this is a case where even I'd try to find something commercial. Maybe pick up a surplus server farm UPS and modify it for your purpose.

What I am about to say could easily be easily rebutted by someone who actually does some measuring of a real motor.

Assuming the main winding has low resistance compared to reactance at stall, then the start winding current will lag the applied voltage by almost 90 deg. If the starting capacitance resonates with the starting winding inductance, then the current in that winding will be in phase with the applied voltage. If the start winding is oriented a quarter cycle mechanically from the main winding. This is exactly the condition required to generate a rotating magnetic field. The resistance of the start winding must be made large enough to prevent high resonant currents from flowing. Under such resonant conditions, there is sufficient damping to prevent high currents flowing through the reactive components.

Bill

There are all kinds of limitations, explicit, implicit or even unmentioned. For example, how many cheap capacitors are rated for high frequency inductance? Energy storage capacitors such as used for large pulse lasers have all kinds of limitations. RMS current, ringing, etc. Sometimes these have to be checked out by the user. Even then, change in manufacturing can upset results after years of satisfactory performance.

Bill

On Thu, 15 May 2008 20:13:58 GMT James Sweet wrote: | | snipped-for-privacy@ipal.net wrote: |> On Thu, 15 May 2008 13:58:50 -0400 John Gilmer wrote: |> |> | If you are only building a handful of units for non-critical applications |> | you might just use the "smoke test." If you are designing for a longer |> | production run it might pay to varify with the manufacturer that the cap is |> | suitable for the application. |> | |> | When in doubt use a conservative design which in the case of caps mean you a |> | significantly higher voltage rating than you calculate or measure. When |> | you have things working you run an extended "smoke test" than then shut down |> | and give the components the "digital temperature test." If you say "ouch" |> | when you touch a component, it may be running a little hot but not |> | necessarily and that's why you read the spec sheets. |> | |> | You will often see "hints" in the suggested applications section of the data |> | sheets or catalog description such as "suitable for switching power |> | supplies" or "recommended for capacitor input filters." |> | |> | You have to pay attention to "frequency" effects as you go to RF as at some |> | point the inductance of the leads will resonate with the capacitance. All |> | this is why when you want a power supply designed you go to a guy who has |> | done it before, etc. |> |> The idea I had in mind was designing a DC to AC inverter for whole house use |> that would support a high enough fault current to ensure the breaker would |> trip on a short circuit. It would have to be able to deliver the high current |> briefly, but that would not be a long term expectation. |> |> I suppose commercial inverter designers don't just pick caps out of a Digikey |> catalog. |> | | | Designing a large inverter is not a trivial task, it's a case where | you'd be better off buying a commercially made unit, it's a much safer | approach at any rate. I'm by no means trying to discourage one from | designing and building their own hardware, but this is a case where even | I'd try to find something commercial. Maybe pick up a surplus server | farm UPS and modify it for your purpose.

But they don't make them. I've searched. Available currents on the ones I have found online (quite many) have been a small percentage above the rated normal loading.

The largest inverter I found was an 18kVA unit. That would be sufficient to provide my normal needs. But it's short circuit rating was a mere 90A. If you have this feeding a panel with say a 60A or 70A main breaker, how long do you think it would take for that main breaker to trip on a short circuit inside the panel bus bars? How long would it take for a 30A 2-pole branch breaker to trip on a short circuit on that branch?