First: Wow - what a great resource this site is. I'm impressed.

I'm in need of replacing the run capacitor in my air conditioning unit once again - it's a recurring event that's been going on for a few years now, each spring.

I'm not sure how much background info is required so I'll keep it to a minimum and hopefully not bore anyone with irrelevant details.

Re my skill level: I've only just rekindled my passion for electronics this year and never really received any formal training in it other than an introduction in high school. I have monkeyed with it some over the years getting more into it with Xbox console modding and then this year discovered the wonderous world of microcontrollers.

In the past I have not used a bleed resistor on the replacement but I recall that the original capacitor had one soldered in place (although I had no idea why it was there at the time). But I'm now thinking that this is the reason that I have been having to replace them so often since (we have cold dry winters here and I'm thinking that the cap is discharging itself catastrophically at some point (see attached pictures).

It's a dual cap with 35uF and 5uF values. It's rated at 370 volts but the unit runs on 220 (operating min:197, max:253).

I've used the capacitor discharge calculator at http://www.welwyn-tt.com/products/resistors/calculation-tools.asp to try to determine an appropriate bleed resistor but I was hoping to get confirmation from some pros who understand the formulas involved and the assumptions that I'm (obviously) ignorant of.

I used the following values: 35 uF capacitance, 370 volts inital charge (or should I use 240-260?), 24 volts safety threshold (should this be lower?) and a 60 second discharge time. This indicates a resistance of 626K Ohms and less than a 1/4 watt. Replacing the resistance value with a standard 560K Ohm value drops the discharge time to 53.6 seconds and the power rating is still (just) under 1/4 watt.

So, is this a sane bleed resistor value (560K Ohm, 1/4 watt) for this scenario? What questions am I not asking that I should be?

Thanks for your time and consideration.

... Al

EDIT: In the picture, it's the common post that's blown out. The Fan post is the upper right and the Herm post is the upper left.

 

Hi Al,
Welcome to the forums.

You should be fine with the 560k Ohm resistor. It's doubtful that the cap has any kind of a charge on it, but it pays to be cautious.

I like to use incandescent light bulbs for discharging caps. They're rather unique in their non-linear resistance characteristics. A bulb's filament has a very low resistance when cold, but increases dramatically when the bulb is at normal operating temperature.

So initially, the discharge rate of the cap will be quite high until the filament comes up to temperature. The power stored in the cap is dissipated as heat and light in the bulb's filament. As the cap discharges, there is less power to dissipate in the bulb's filament, so the filament cools - decreasing it's resistance, causing the rate of discharge to increase.

 

Thanks Sarge. I'm actually intending to solder the bleed resistor in place on the new one so that whenever the unit is powered off, it discharges the cap within a minute or so. I'll use the light bulb idea though for one-time discharges. Thanks.

 

Oh, don't solder it in permanently.

You really shouldn't need a bleeder resistor on it. Your cap is probably getting "zapped" by transients on the 240v line.

 

Hi Al,
Welcome to the forums.

You should be fine with the 560k Ohm resistor. It's doubtful that the cap has any kind of a charge on it, but it pays to be cautious.

I like to use incandescent light bulbs for discharging caps. They're rather unique in their non-linear resistance characteristics. A bulb's filament has a very low resistance when cold, but increases dramatically when the bulb is at normal operating temperature.

So initially, the discharge rate of the cap will be quite high until the filament comes up to temperature. The power stored in the cap is dissipated as heat and light in the bulb's filament. As the cap discharges, there is less power to dissipate in the bulb's filament, so the filament cools - decreasing it's resistance, causing the rate of discharge to increase.

As a broadcast engineer, I often had to work with 25KV+ power supplies. We never relied on the bleeders, but always went in there with big shorting sticks. Bleeder resistors DO burn out, occasionally. One time one of my proteges went in there with a shorting stick and got a huge SNAP! from a 10uf oil capacitor. I told him that's what happens if you don't have a bleeder, and why you use a shorting stick. "It's safe now, isn't it" he asked. "No it's not," I said, "You just reversed the charge on the capacitor. Hit it again and see!"

He did just that....and behold, another big snap! I told him he had to HOLD the stick on the terminals for about 10 seconds or so. NOW it's safe!

More than one engineer has gotten knocked on his keister after zapping with a shorting stick!

Eric

 

Oh, don't solder it in permanently.

Is there a problem with soldering it in? Is it dangerous or just inefficient?

... Al

 

It's dangerous for the capacitor to have an amateur holding a soldering iron to it. Solder the resistor to a couple of .250 crimp connectors and install it that way, then run down to the Home Store and spend $30 for a surge protector that goes on your mains box.

 

Agreed. You should install the surge protector.

Also, check to BE SUPER SURE that is the proper CAP for the motor. If you got the unit used, or with the house, it is totally possible that the initial cap you saw was a "rig" job to get it working with the wrong cap.

 

It's dangerous for the capacitor to have an amateur holding a soldering iron to it.

Could you please define "amateur"? I wouldn't consider myself an amateur but I guess that's a matter of perspective.

... Al

 

Agreed. You should install the surge protector.

Also, check to BE SUPER SURE that is the proper CAP for the motor. If you got the unit used, or with the house, it is totally possible that the initial cap you saw was a "rig" job to get it working with the wrong cap.

I am the original owner of the house built in 1997. The whole home surge protector was installed the day after I moved in. The air conditioner was install new a month later. The original cap with a bleeder resistor lasted until 2005. Since then I've been through 3 or 4 of them. They always fail between the last use in the fall and the first use in the spring.

Perhaps the whole home surge protector needs to be replaced. Is there an easy way to test them?

... Al

Edit: The replacement caps have thus far all come from Carrier dealers.

 

Thats odd. Have you contacted Carrier regarding the bleeder? I would, and you should. They most likely have a mention of it in the warranty info. But if not, an email to their tech support should result in an answer. You may want to see if they changed anything since your unit was manufactured.

You most likely are not the only customer with this problem. And most folks pay a few hundred dollars to have them replaced, so complaints to Carrier were probably high.

They may have a pigtail to connect to the .25in spade on the caps themselves.

At least get the resistor value they recommend. They may even require it in the warranty.

[ed]
I just looked through some carrier wiring diagrams, and all three show bleeders across the caps..but no values.

Contact them for the values.
http://www.residential.carrier.com/apps/contactus.jsp?b=c
[/ed]

 

Sorry if you're feeling prickly about being called an amateur. My perspective is as a State Certified Air Conditioning Contractor for 25 years and an electronic design engineer for 36 years.

From that, I can tell you that the cap has a protective fuse in the common terminal, a 35/5 is a common value for 2 ton centrals with an E.E.R of 12, 370 volts is normal because the capacitor forms a series circuit with the inductance of the motors and that causes higher voltages than what you get off the power lines. A diagnostic method might be to use a seperate 35 uf/370 and a 5uf/370 to seperate the fan circuit from the compressor circuit. The next time the common terminal "safety's out" you will know if the compressor is having a problem or the fan is having trouble. (Much more likely, the compressor.)

Turn off the disconnect box at the condenser, turn on the thermostat, then go outside and turn the power to the condenser back on. Watch closely to see if the compressor is struggling to get started and, if you have an "amp clamp" meter, check the run current against the name plate rating. You can allow the refrigerant pressures to develop, turn the power off, and right back on to get a reading of the "locked rotor" current. Only do this stall test long enough to read the meter.

To be thorough, check for any leakage from any of the three compressor terminals to the case. I use a $300 ohm meter, but the proper tool is called a "megger".

Be aware that the fusable link in the capacitors is literally a fuse. It can be falsely failing because of pre-heating caused by corroded connections. If there is any sign of heat damage on the insulation of the wires, replace them.

Check for a "start helper" of any sort. The recent designs use a solid state device that allows current when cold and stops assisting in a matter of seconds. If this is defective, the compressor will start slower and the stress will be inflicted on the capacitor.

I hope this is helpful.

 

Thanks, -rm-. I have tried to get the information from several sources including the manuals, wiring diagrams and service guides, Carrier dealers (who just want me to get them to come out and inspect it for a bazillion dollars) and suppliers of the capacitors all to no avail.

I have not - until just now - used the form on the Carrier website or contacted them directly to try to find this information (slaps himself on the forehead ). Hopefully that turns up something to confirm/deny what I've estimated.

Having said that, it should be a value that can be calculated based on the known values (voltages, capacitance) and the desired result (time to discharge), thus the online calculator that was referenced in the beginning of this thread.

My only concerns with the calculations that I've done are whether 60 seconds is too long or too short and whether I should really be using a higher wattage resistor (which it seems is based on the length of time to discharge the cap).

... Al

 

Sorry if you're feeling prickly about being called an amateur. My perspective is as a State Certified Air Conditioning Contractor for 25 years and an electronic design engineer for 36 years.

So do you have access to Carrier documentation/information that will identify the proper value for a bleed resistor on a Carrier model # 38TKB024310 run capacitor?

Turn off the disconnect box at the condenser, turn on the thermostat, then go outside and turn the power to the condenser back on. Watch closely to see if the compressor is struggling to get started and, if you have an "amp clamp" meter, check the run current against the name plate rating. You can allow the refrigerant pressures to develop, turn the power off, and right back on to get a reading of the "locked rotor" current. Only do this stall test long enough to read the meter.

To be thorough, check for any leakage from any of the three compressor terminals to the case.

Thanks - I'll give this a try once I have the new cap.

If there is any sign of heat damage on the insulation of the wires, replace them.

No signs of heat damage on the connectors or insulation.

Check for a "start helper" of any sort.

There is no start capacitor in this unit.

I hope this is helpful.

Yes it is, thank you.

... Al

 

I don't have any bleed off resistors laying around today, but I can calculate.

Bleed off resistors are almost an after-thought. That's why some have 'em..some don't. They do not make capacitors go bad, and this won't fix your problem. The time limit is about 3 minutes because any manufacturer with good sense will put a time delay on a compressor restart, and Carrier has good sense.

correction: 250VAC will produce 354 volts peak without any inductance. Good reason for a 370 volt cap! Why would I say 250 VAC? Because that's what happens at my house.

The real limit is the voltage a resistor can survive. That's what puts you in the 2 watt range...500 volts high limit.

Mouser part number 71-CPF3100K00FKE14 = 100k, 3W, 62 cents.
Mouser part number 594-5083NW100k = 100k, 2W, 32 cents.

Both of these form a time constant of 3.5 seconds with the 35uf cap. Ten time constants = 35 seconds to bleed off ninety-some percent of a worst case charge. Both of 'em are flameproof. I think the standard values I find in air conditioners are around 150k, but I've never replaced one because they don't burn out.

 

Thats odd. Have you contacted Carrier regarding the bleeder? I would, and you should. They most likely have a mention of it in the warranty info. But if not, an email to their tech support should result in an answer. You may want to see if they changed anything since your unit was manufactured.

Here's Carrier's response:
"Thank you for contacting Carrier. Your dealer is in the best position to assist with identifying correct, current part numbers. In the event you are not familiar with dealers in your area, I have listed several below, for your convenience".

The dealers in my area have definitely not been helpful in this matter.

... Al

 

To bring this thread full loop: The new cap arrived, the AC is working again. I haven't added the bleeder yet but plan to use a 150K Ohm, 2 Watt resistor soldered to two female disconnects at some point this season (I'll take a picture when I do).

I'm also planning on replacing the whole home surge protector since the existing one has been 13 years in operation.

Thanks for all the help.

... Al