ANGRY INDUCTIVE HEATER!

[Farfle]

Heya all, I had an interesting experience with the heater in the Porsche. It seems to have a significant amount of inductance, and at 100v and 18ish amps, it's very difficult to switch.

It DESTROYED the 20A OEM relay I started with, so I went back and got a big honking 50A switch with large contacts, and an aftermarket 24VDC 50A relay and made a flyback circuit as per this diagram, with three Parallelled 6A10 1kv 4A diodes+ :

It didn't seem to have any effect, and melted down that big 50A switch, which sustained an arc for about 10 seconds

So then I made this circuit: with a 1000uf cap, a 5 ohm resistor and those same three paralled diodes.

With that circuit, it melted my 50A relay

A quick note though, hooking just a capacitor across the heater instantly fixes the contact-opening arc problem, but then the contacts weld shut when the switch closes .

SO, what the heck can I do short of an HVAC throw switch for this silly heater?

 

[Ricky_nz]

Sounds like you need a proper relay with 100VDC rated contacts.
An alternative is to build a switch using a FET and then switching the voltage on the gate to control the FET.

 

[liveforphysics]

Even a big 500v rated multi-hundred amp kilovac or gigavac will fail under breaking DC with big inductance. They will likely make it hundreds of cycles first though.

Gotta drive it with a cheapo brushed motor controller or something or build your own with a big FET.

Alternatively, you can make an arrangement of diodes and caps that clamps the flyback one contact break, but doesn't make it inrush across the caps ESR on make. Its still only going to be an on-off system though, if you start trying to switch it fast to control average power it will fail.

 

[John in CR]

Or be a man and don't waste the already reduced battery capacity on a heater. Wear extra clothes. Of course there's also the approach I adopted...move somewhere with a better climate.

Merry Christmas

 

[Farfle]

Or be a man and don't waste the already reduced battery capacity on a heater. Wear extra clothes. Of course there's also the approach I adopted...move somewhere with a better climate.

Merry Christmas

Thanks! Thats the current plan lol, (was -7F last week) the heater we have is a 5kw 240v heater from a maytag dryer, and a bilge blower fan. It is the most ass-kicking heater ever.

I had a thought, what about a big honking SSR? The heater draws around 19A cold, and tapers to 16a as it heats up.

 

[Ricky_nz]

I had a thought, what about a big honking SSR? The heater draws around 19A cold, and tapers to 16a as it heats up.

An SSR should work fine provided it is one designed for use with DC as these have FETs internally.

 

[fechter]

Most solid state relays work on AC only. I had a similar problem using a hair dryer on DC. The switch failed on the first cycle. A FET or IGBT module is probably the best approach.

 

[Ricky_nz]

Most solid state relays work on AC only. I had a similar problem using a hair dryer on DC. The switch failed on the first cycle. A FET or IGBT module is probably the best approach.

Yep AC ones using Triacs are more common but there are DC (FET based) ones out there.
I guess the advantage over a bare FET is the opto isolated drive which would allow switching from the 12V side of the electric car, isolated from the main battery

 

[Arlo1]

Maybe use a precharge circuit with the cap you used.

 

[Ricky_nz]

Maybe use a precharge circuit with the cap you used.

With inductive loads the biggest problem is the turning off rather than on.

 

[Punx0r]

Multiple relay poles wired in series?

 

[zombiess]

put a 0.25W 100 ohm resistor in series with it :lol:

I am thinking you need some sort of DC motor speed controller. The benefit of this you also get to adjust the heat so it's not just on/off. It also reduces the current drawn after it's warmed up. Then create a PID controller with a temp sensor and a digital read out for a nice auto climate control

 

[liveforphysics]

Perhaps the right Transorb, say something that clamps at 20v above the highest voltage you power the heater with could absorb the flyback and protect the relay.

A little brushed motor controller really seems ideal though.

 

[bigmoose]

Luke, do you think one of the GigaVac contactor products with internal magnetic arc extinguishment might work? It would be an experiment with a $100 ish contactor versus a $20 eBay IGBT or FET brick which we know would work... and would include the flyback diode in the package. Think I would just go with a surplus IGBT/FET 1/2 bridge FET brick.

 

[liveforphysics]

My hunch is that the magnetic arc blow-out system in high-end contactors would only help it last more cycles, but not be a long-term solution. I'm assuming his flyback from the heater is in the kV range.

 

[flathill]

Ditch it and use a heat pump
Way way more efficient
Dont create
Move

 

[Farfle]

Ditch it and use a heat pump
Way way more efficient
Dont create
Move

Well, the heater was 21 dollars, don't even want to know what 100VDC heat pumps run

 

[dnmun]

if you can use a brushed motor controller like luke said and get something like 20V average out of it then you could use a small windshield defroster/deicer from harbor freight to de ice the windshield.

they draw about 12A at 12V so maybe you would get 20A at 20V so you would have 400 W to defrost the windscreen. it may take a few minutes at -6 F. it has a fan so that might limit the voltage but i bet you could push it to 30V and maybe 800W.

 

[fechter]

Couldn't you just put a big diode reverse biased across the heater to catch the inductive spike (like you do with a relay coil). This would at least limit the voltage across the contacts when they open. Even if the load was perfectly resistive, you could still get a big arc though. Sort of like an arc welder.

 

[friendly1uk]

I'm a qualified auto electrician but have only ever seen resistive heaters. These would have a huge inrush, way higher than 19amps although most meters would miss it. I imagine this is what you have. Any induction problems being the fan and perhaps coiled element, but cured already with standard suppression.
I'm a normal spark these days. I have seen many relays fail first go. Mini contactors last forever doing the same job.

I have not seen a 50amp relay for this duty. Only 24v which would likely be the wrong contact material.

 

[liveforphysics]

Couldn't you just put a big diode reverse biased across the heater to catch the inductive spike (like you do with a relay coil). This would at least limit the voltage across the contacts when they open. Even if the load was perfectly resistive, you could still get a big arc though. Sort of like an arc welder.

That should help quench, but it's going to ring and likely re-strike a few cycles because it's trying to clamp at such a low voltage it doesn't dissipate the energy. TVS diode (typically called transorbs) will clamp the spike's while they've got a bunch of voltage, which should eat the inductively stored energy much faster IMHO.

However, perhaps a diode as you suggested, and a TVS clamping some safe voltage window above his heater drive voltage could make the spike ring in a fixed window of relative low voltage and keep it from drawing big nasty arcs that damage the contacts.



I would do something like this, but of course hack it's control side to be powered through a resistor divider stage that makes sense for your voltage, and replace the cap and FET with whatever they need to be for your pack's voltage. Then you can poke the knob through a panel on the dash and have variable heating control. Short trips that you've got plenty of pack for, you can crank the heat. Trying to get maximum range, you can still just take the edge off by feeding a few amps.

http://www.ebay.com/itm/12V-24V-36V-40V-10A-DC-Motor-Speed-Control-PWM-Controller-/221097853692?pt=LH_DefaultDomain_0&hash=item337a7582fc

 

[Farfle]

I like the ESC idea, so here is my solution:

split the heater into two parts, run two cheap 48v 30A ESC's, each controlled by seperate taps on a dual pot for isolation:

http://kellycontroller.com/kds48050e50a24v-48v-mini-brushed-controller-p-282.html

http://www.ebay.com/itm/2-x-5K-OHM-Linear-Dual-Taper-Rotary-Potentiometers-B5K-5KB-POT-ALPHA-/250956199510?pt=LH_DefaultDomain_0&hash=item3a6e27b656

and run them off of the pack such that each controller is on 12s:



linear heat control, even pack drain, and sub-150 dollars.

OR I can run one of these for 60 odd dollars, with the cap across the output to kill the inductive spike, which doesn't work for the standard relay, but does for the SSR because it can handle the inrush current:

http://www.mouser.com/ProductDetail/Continental-Industries/SVDD-1V40/?qs=sGAEpiMZZMtq49AUx5G3778uEJUSSwFqEM2%252bGVMQx1c%3d

 

[bearing]

24VDC 50A relay

You are using a 24VDC relays on 100V?
That's the problem. The air gap in the relay is too small. Connect many relays in series until they stop arcing (at least 4x24V as a start). You can probably use standard relays instead of the expensive ones. I had the same problem at work when doing harsh things to test electronics.

EDIT: or use an electronic switch, as suggested earlier in the thread. There are solid state switches for DC if you don't want to use a MOSFET.

 

[Arlo1]

24VDC 50A relay

You are using a 24VDC relays on 100V?
That's the problem. The air gap in the relay is too small. Connect many relays in series until they stop arcing (at least 4x24V as a start). You can probably use standard relays instead of the expensive ones. I had the same problem at work when doing harsh things to test electronics.

EDIT: or use an electronic switch, as suggested earlier in the thread. There are solid state switches for DC if you don't want to use a MOSFET.

nothing is really designed to break 100v with inductance in the circuit. I mean a good contactor will do it but only a few times till it burns the contacts up really bad.
Mosfets are designed for this.... A DC controller is one of the best ways to go. Other wise maybe just turning mosfets on and of as you need the heat on and off? You could series parallel the 2 heaters for a high and low....

 

[bearing]

The problem is not the inductance. The problem is simply the high DC voltage. The relay needs to have a large enough air gap to prevent an arc when closing. I suggested using low voltage relays in series, but a relay rated for the used voltage will of course do the trick as well. If it was the inductance, the problem would have been solved by the diode.

But sure, a IRF4115 or similar in place of the relay should do the trick.