Which capacitor should it be?

[mgenius33]

This!

Unless there are other electrical paths available in parallel to where you are taking your readings, it is not possible for the sum of the run and start winding currents, as measured with your amp meter clamped around both leads, to differ from the current in the common lead.
Likewise it is not possible for the common minus run winding current, as measured with the meter clamped around both leads, to differ from the start winding current.

If you do get unequal readings, it just means your clamp on meter is being affected by other electromagnetic fields in the area.
This will happen to even the best true RMS meters.

That won't work. Just install the cap specified by the compressor manufacturer.

Kirchoff's Law requires those two amp readings to be equal regardless of cap size.

That was the answer I gave on "hard starting scroll" in the pro resi forum.
Tom R gave an answer in post #25, refuting my statement, stating the phase shifting may cause current to flow through the start to the run instead of through to common.
I'm kinda busy so I don't have much time to research. I thought I knew where he was coming from at first, but was thinking about it today at work. I believe his findings are based on a failing capacitor allowing for a high resistance path to ground, not magnitude of capacitance.

[mark beiser]

That was the answer I gave on "hard starting scroll" in the pro resi forum.
Tom R gave an answer in post #25, refuting my statement, stating the phase shifting may cause current to flow through the start to the run instead of through to common.
I'm kinda busy so I don't have much time to research. I thought I knew where he was coming from at first, but was thinking about it today at work. I believe his findings are based on a failing capacitor allowing for a high resistance path to ground, not magnitude of capacitance.

The phase shift is why the sum of the current measured in the start lead and in the run lead is higher than the common lead if you clamp your meter around each lead individually.
When you clamp your meter around 2 leads at the same time, the effect of the phase shift is canceled out, so to speak. The measured current with your meter clamped around 2 leads at the same time should equal the reading from the remaining lead, regardless of the size of the run capacitor.

If they are not the same, either your meter is not getting accurate readings, or current is leaking somewhere.

[beenthere]

cool. ive got a unit that kills capacitors twice a summer. everybody ive talked to has said i have the correct size. im actually going there tomorrow and will try that out!

if its off do i just keep swapping caps until i get the correct reading?

Use 2 capacitors wired in series to equal the uf you need. Less heat build up in the caps.

[John Markl]

Use 2 capacitors wired in series to equal the uf you need. Less heat build up in the caps.

Series, or parallel ?

[ar_hvac_man]

Are you using American Caps? I've added a hard start Cap on a few units that were eating Caps and it seems to have helped.

Im using Packard. They boast the american flag but who knows nowadays. I added a hard start after my second cap blew in the same summer, a number of years ago. Still devouring them.

[ar_hvac_man]

Quit putting Chinese/Pakastani/Mexican run caps in it and put an AmRad made in USA capacitor in it and see how it does.

Didnt realize you could see my inventory from Texas. You must have me mistaken for someone else...

[ar_hvac_man]

Use 2 capacitors wired in series to equal the uf you need. Less heat build up in the caps.

Good idea I will try that. You do mean parallel though right?

[rundawg]

I added a hard start after my second cap blew in the same summer, a number of years ago. Still devouring them.

Are you using 370v or 440v capacitors?

Have you ever measured the voltage between the Com and Herm terminals on the capacitor when the unit is running?

Just a 8% increase in voltage, over the rated voltage of the capacitor, will decrease the life of the capacitor by 75% - see attachment.

[ar_hvac_man]

I was using 370's which it calls for but thought maybe going to 440's would help.

I havnt measured that voltage but will when I get there today. Thanks for the literature. Ill try anything at this point.

The system runs great, when it runs. Its an old Ruud cube. 1990 model. Clean, amp draws all good, alll measurements within specs.

[chuckcrj]

Series, or parallel ?

Parallel I hope.

[chuckcrj]

Can someone explain this. A Wiki article on Kirchhoff's law states;

KCL is only valid if the charge density remains constant at the point to which it is applied. Consider the current entering a single plate of a capacitor. If one imagines a closed surface around that single plate, current enters through the surface, but does not exit, thus violating KCL. Certainly, the currents through a closed surface around the entire capacitor will meet KCL since the current entering one plate is balanced by the current exiting the other plate, and that is usually all that is important in circuit analysis, but there is a problem when considering just one plate.

Doesn't this say that a circuit can violate Kirchhoff's law if a capacitor is involved?

[chuckcrj]

Can someone explain this. A Wiki article on Kirchhoff's law states;

Doesn't this say that a circuit can violate Kirchhoff's law if a capacitor is involved?

Never mind, I got it now.

Question, how did this idea get started if it doesn't work? I mean the idea that you can diagnose a run cap by comparing the amp draw of the run and start, and comparing it to the current in the common?

[Wendo]

Good idea I will try that. You do mean parallel though right?

I think he actually means series. If you need a run cap of 35 mfd you can wire two 35 mfd caps in series. You'll end up with the same 35 mfd but you'll double the voltage rating. IMHO a lot of run cap failures are because of

A: Cheap junk!

B: Heat buildup