Rectifier + large smoothing capacitors are overloading a 20A breaker

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
Hello, long time lurker, first time poster here. I am having a problem that probably has a simple solution, but that solution is escaping me. I have a thyristor-based voltage controller feeding a full-wave bridge rectifier to provide approx 48VDC. The rectifier is loaded with 4x10000uf caps (rated at 63V) for voltage smoothing, and 8 100W peltier coolers. When I turn up the thyristor to 30-50VDC, everything works BUT the 14ga cord always gets warm/hot and the 20A breaker at the panel always flips after 20+ seconds. This makes no sense to me because I have a DC shunt measuring the current draw of the peltiers at *only* 500-700W, and an AC current measurement using a Kill-a-Watt device of only 7A....

With such a small DC wattage draw, what could be causing such a large (and miscalculated?) AC power draw? The only thing I can think of is reverse voltage breakdown at the rectifier and backfeeding the AC supply, but that shouldn't be happening with only 50VDC coming from the caps.... Hopefully my diagram is of some assistance.

Thanks for the help!

P.S. - In case it helps, here are the actual voltage controller/rectifier/caps:
Caps - https://www.amazon.com/gp/product/B07R1X6DPJ/
Voltage controller - https://www.amazon.com/gp/product/B087JBGHXG/
Rectifier - https://www.amazon.com/gp/product/B07JC3Z79N/
 

Attachments

drjohsmith

Joined Dec 13, 2021
852
Where did you find that circuit ?

My first thought,
when the thyrister turns on,
the capacitors are "a short" to the instant current rise,

Then the very large capacitance across the bridge, is shifting the current spikes with respect to the voltage spikes.

Your also direct coupled to mains,
so the pelties will be "at mains voltage"

not a good place to start,
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
I didn't find that diagram, I created it to illustrate what I wired up. I know that the caps are are presenting "a short" when it's first powered on, but that doesn't seem to cause a problem as it functions for 20+ seconds. I've tried fewer caps, but if I use fewer they get hot/overloaded.

Do you have a suggestion for how I could rework this to function properly? (Also, I don't understand why you say that I have the peltiers direct coupled to mains voltage as they are only being fed by the 48V from the rectifier)

Thanks!
 

Papabravo

Joined Feb 24, 2006
21,158
What is you mains voltage?
Multiply the RMS value by 1.414 (The Square Root of 2) to get the paek voltage out of the rectifier. Take an allowance of 1.4 V off of the peak voltage for the rectifier diodes. That will be the average DC voltage across the peltiers
 

MrChips

Joined Oct 2, 2009
30,706
The thyristors reduce the average power to the load.
The peak voltage is still 120V x 1.4 = 168V

Also the four capacitors appear as a short across the power source.

The reactance of a capacitor is given as:
Xc = 1 / 2πfC

Plug in the numbers:
Xc = 1 / (2 x 3.14 x 120 x 4 x 0.01) = 0.03Ω

So you have a 0.03Ω short circuit across the AC mains.
 

Papabravo

Joined Feb 24, 2006
21,158
The thyristors reduce the average power to the load.
The peak voltage is still 120V x 1.4 = 168V

Also the four capacitors appear as a short across the power source.

The reactance of a capacitor is given as:
Xc = 1 / 2πfC

Plug in the numbers:
Xc = 1 / (2 x 3.14 x 120 x 4 x 0.01) = 0.03Ω

So you have a 0.03Ω short circuit across the AC mains.
Wow! That's 5600 Amps Peak
You did omit the effect of ESR which would help, but probably not much.
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
Ok, I understand that you're saying there is a large inrush when powered on, but do you have a suggestion for a better design for smoothing the voltage provided to the peltiers? Or is this approach completely unworkable?
 

MrChips

Joined Oct 2, 2009
30,706
Peltier TEC devices operate on DC voltages. You don't have to smooth the rectified AC. A modest amount of capacitance will be fine.

What you need is a step-down transformer in order to reduce the applied voltage.
Don't forget that the peak voltage is the transformer AC RMS output voltage multiplied by 1.4.

For example, a 12V TEC needs about 9-10VAC output (taking into account the loss in voltage across the rectifier diodes).
 

MrChips

Joined Oct 2, 2009
30,706
I don't know what are your plans for eight 100W TECs.

A TEC is a heat pump. It moves heat from one side (cold) to the other (hot).
The efficiency of the TEC is a function of the temperature differential, i.e. the difference in temperature between the hot side and the cold side. For it to be efficient, you need to remove the heat from the hot side. If you don't you will quickly destroy the TEC.

Equally important, you need to be able to control the temperature of your device.
Hence some form of temperature regulation is required. It can be a simple ON/OFF thermostat-like controller or a PI controller (Proportional + Integral).
 

AnalogKid

Joined Aug 1, 2013
10,986
Please show us the schematic for the SCR phase control.

My guess is that the SCR turn-on is being delayed a portion of each AC cycle such that the average value of the remaining portion is approx. 48 V or whatever. If so, this means that there are very large current spikes each cycle.

NOTE - if the SCR is placed in the actual circuit where it is shown in the schematic, then you can delete the bridge rectifier. An SCR conducts only 1/2 of the incoming AC waveform, so two of the diodes never conduct any current. This increases even more the current spikes, the ripple voltage and current, and heating in the capacitors.

Is the circuit breaker strictly magnetic, or thermal-magnetic?

ak
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
I'm using this to cool a circulating liquid. I'm sufficiently cooling the hot side of the TEC, it's on a liquid cooling loop that can handle 1000W+, and temp regulation is not a factor as I don't care how cold it gets, only that I cool the hot side.

The problems I have a with a transformer are 1) It's big/heavy and 2) I haven't found a 110VAC to ~48VAC that can handle 20+ amps. That's why I went this route since my space is limited.

So it sounds like I need to go a different route other than thyristor+rectifier+smoothing caps. Also, when I reduce the capacitor count/values, they simply overheat trying to keep up with the TEC load, and TEC efficiency goes way down. I could get rid of the caps altogether BUT when I do, the TEC ceases to function properly, it ONLY generates heat, not cooling. So TEC's definitely do not like straight-up pulsed DC for some reason. I'm definitely not smart enough to know why that is haha.
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
Please show us the schematic for the SCR phase control.

My guess is that the SCR turn-on is being delayed a portion of each AC cycle such that the average value of the remaining portion is approx. 48 V or whatever. If so, this means that there are very large current spikes each cycle.

NOTE - if the SCR is placed in the actual circuit where it is shown in the schematic, then you can delete the bridge rectifier. An SCR conducts only 1/2 of the incoming AC waveform, so two of the diodes never conduct any current. This increases even more the current spikes, the ripple voltage and current, and heating in the capacitors.

Is the circuit breaker strictly magnetic, or thermal-magnetic?

ak
I'll try out eliminating the rectifier, thanks for filling me in on the SCR waveform! And the circuit breaker is indeed thermal-magnetic.
 

Danko

Joined Nov 22, 2017
1,829
Peltier TEC devices operate on DC voltages. You don't have to smooth the rectified AC. A modest amount of capacitance will be fine.
It should be minimum ripple, no phase control, no PWM:
https://www.meerstetter.ch/customer-center/compendium/71-peltier-element-efficiency
So TEC's definitely do not like straight-up pulsed DC for some reason.
Recommendations of the manufacturers
Manufacturers of Peltier elements suggest using direct current and limiting current ripple for the regulation of the output current. They explicitly advise against the usage of direct PWM control of Peltier elements:
  • Ferrotec: "However we recommend limiting power supply ripple to a maximum of 10 percent with a preferred value being <5%."
  • RMT: "TEC [Peltier elements] controlled by PWM operates less effectively than at DC current. The PWM control is always less effective than TEC operation at the same average DC current and power consumption."
  • Marlow: "Thermoelectric coolers require smooth DC current for optimum operation. A ripple factor of less than 10% will result in less than 1% degradation in ∆T. [...] Marlow does not recommend an ON/OFF control."
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
Please show us the schematic for the SCR phase control.

My guess is that the SCR turn-on is being delayed a portion of each AC cycle such that the average value of the remaining portion is approx. 48 V or whatever. If so, this means that there are very large current spikes each cycle.

NOTE - if the SCR is placed in the actual circuit where it is shown in the schematic, then you can delete the bridge rectifier. An SCR conducts only 1/2 of the incoming AC waveform, so two of the diodes never conduct any current. This increases even more the current spikes, the ripple voltage and current, and heating in the capacitors.

Is the circuit breaker strictly magnetic, or thermal-magnetic?

ak
And this is the best I can do for a circuit right now, I can't get in to tear it down right now (I didn't build it).
 

Attachments

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
What you need is a switched mode power supply.
That's exactly what I was afraid of. I can do that but it'll have to sit outside the device.

If you're feeling extra nice, can you please let me know why the "quick and dirty" method that I built will never work? Is there just too much power involved?
 

MrChips

Joined Oct 2, 2009
30,706
That's exactly what I was afraid of. I can do that but it'll have to sit outside the device.

If you're feeling extra nice, can you please let me know why the "quick and dirty" method that I built will never work? Is there just too much power involved?
Shouldn't that be obvious? You can't make it work as is.
 

Thread Starter

jwilcox767

Joined Jan 4, 2022
9
Shouldn't that be obvious? You can't make it work as is.
Not obvious to me, I only know enough to be dangerous. I figured there were some enhancements that could be made to the current design to fix the flaws, but sounds like that is not an option.
 

Ian0

Joined Aug 7, 2020
9,667
And this is the best I can do for a circuit right now, I can't get in to tear it down right now (I didn't build it).
You need an inductor between the phase controller and the capacitors (but it will probably be bigger and heavier than the switched-mode supply suggested by @MrChips )
 

drjohsmith

Joined Dec 13, 2021
852
Not obvious to me, I only know enough to be dangerous. I figured there were some enhancements that could be made to the current design to fix the flaws, but sounds like that is not an option.
My summary of the discussion

a) The cooler will be biased at mains voltage
b) The current to the cooler is very large pulses
c) The thyristors does not control the voltage, so the capacitors are seeing mains voltage
d) There is no way to protect the Peltier to stop it over heating and "burning out"
 
Top