Author Topic: MOT (microwave oven transformer) cooling.  (Read 11011 times)

0 Members and 1 Guest are viewing this topic.

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
MOT (microwave oven transformer) cooling.
« on: January 02, 2019, 10:24:29 pm »
I'm looking for some ideas to cool a single MOT. I've used a few single transformer in the past, but as most know they run hot as a single unit. Knowing that current limiting is the best option for better cooling and efficiency, it's not practical for the foot print I'm looking for.
As for cooling a single unit fan alone is not that great, I hit the high 80°C when loaded. I realise that's not that hot for a MOT, but the core temperature is sure to be far higher.

Having had a think on it, I've come up with modified air cooling using heatsinks attached to the transformer frame to increase surface area of the iron. But attaching the heatsinks is a bit of problem. I know there are thermal glues, but sure these are for small heatsinks, and there transfer temperature rating isn't that good. So I was wondering if its viable to drill and tap two M3 holes on each side to aid attaching the heatsinks ? My concern here is not wanting to upset the laminations with possible shorting. Any thoughts appreciated, my other idea is oil cooling in a suitable casing that can be sealed. It seems a messy headache to go through, and I'm unsure of longevity of the cooling oil.
I'm learning towards heatsinks, but don't want to degrade the iron core.
If I could current limit with out adding bulk I would do that, I thought about a triac type dimmer switch of up to 1000 watts, but gather these are not suitable as the modify the sine wave of the AC and will probably over heat the primary winding. Also I think there not suited to inductive loads.
Thoughts at ideas appreciated.
Thanks for reading.  :)
 

Offline joseph nicholas

  • Frequent Contributor
  • **
  • Posts: 408
  • Country: mx
Re: MOT (microwave oven transformer) cooling.
« Reply #1 on: January 02, 2019, 10:31:09 pm »
Without knowing what you're using it for, its hard to reply.  If you could elaborate more on your ultimate goal it would be useful.
 
The following users thanked this post: L_Euler

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
Re: MOT (microwave oven transformer) cooling.
« Reply #2 on: January 03, 2019, 12:06:40 am »
Thank you for replying. Yes sure, to drive high current linear RF amplifier.
Currents around 20 to 30 Amps at 12 to 13.8 Volts.
 

Offline thedoc298

  • Contributor
  • Posts: 36
  • Country: us
Re: MOT (microwave oven transformer) cooling.
« Reply #3 on: January 03, 2019, 07:37:42 am »
Why not put mot in bucket of mineral oil.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
Re: MOT (microwave oven transformer) cooling.
« Reply #4 on: January 03, 2019, 11:54:59 am »
I had considered oil, but would rather go with air as a first option. I was hoping to attach 3 x heatsinks with 6 x 3mm screws. Only not sure if drilling the EI transformer frame might upset the transformer magnetic circuit.  I'd use 2 x 3mm shallow holes on each of the three surfaces.
 

Offline joseph nicholas

  • Frequent Contributor
  • **
  • Posts: 408
  • Country: mx
Re: MOT (microwave oven transformer) cooling.
« Reply #5 on: January 03, 2019, 04:44:23 pm »
In order to get 20 to 30 amps for your project you will need to use 2 of these things in series.  The average microwave uses about 1200 watts or so.  You should also keep in mind that the average household circuits have 20 amp fuses or breakers for the mains.  I see you live in the UK but these limitations would apply to you as well.
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10172
  • Country: gb
Re: MOT (microwave oven transformer) cooling.
« Reply #6 on: January 03, 2019, 05:29:07 pm »
I would think that heatsinking the core and frame would be pretty ineffective and doesn't tackle the issue of heat buildup inside the windings.

I would approach it with forced air cooling, ducting the air so it hits the core edge-on and flows over the windings at each side. I think it's the best you'll do without messy oil cooling - where you need to arrange some circulation, either convection or stirred to be efficient anyway. You really don't want to be generating localized hot-spots in a can of mineral oil.

I know you've tried a fan already but it sounds as if you need to use a bigger one and use a duct to ensure that all the air gets 'squeezed' past the transformer. A properly ducted fan will shift a hell of a lot of heat.


Edit: The Copper (or Aluminium) windings will be a lot more thermally conductive than the steel core but there's no way to heatsink them... and you don't want to start adding things to the core that will impede airflow over them.
« Last Edit: January 03, 2019, 05:35:17 pm by Gyro »
Best Regards, Chris
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: MOT (microwave oven transformer) cooling.
« Reply #7 on: January 03, 2019, 06:02:32 pm »
In order to get 20 to 30 amps for your project you will need to use 2 of these things in series.  The average microwave uses about 1200 watts or so.  You should also keep in mind that the average household circuits have 20 amp fuses or breakers for the mains.  I see you live in the UK but these limitations would apply to you as well.
I agree that pushing a MOT to VA levels and duty cycles where simple fan cooling is insufficient is unwise, however you have a misunderstanding of UK wall sockets.    Subject to other loads on the circuit, any BS1363 UK wall socket can supply 13A at nominally* 240V, i.e 3.1 KVA.   

Unless the PSU design is ludicrously inefficient, 30A @ 13.8V DC shouldn't require more than 1 KVA AC in, a third of the max capacity of a BS1363 socket.

* The supply network in general remains at pre-EU harmonisation voltage levels, as the harmonisation was a paper exercise achieved by shifting the tolerance bands.
« Last Edit: January 03, 2019, 06:04:19 pm by Ian.M »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: MOT (microwave oven transformer) cooling.
« Reply #8 on: January 03, 2019, 06:09:01 pm »
In order to get 20 to 30 amps for your project you will need to use 2 of these things in series.  The average microwave uses about 1200 watts or so.  You should also keep in mind that the average household circuits have 20 amp fuses or breakers for the mains.  I see you live in the UK but these limitations would apply to you as well.
I agree that pushing a MOT to VA levels and duty cycles where simple fan cooling is insufficient is unwise, however you have a misunderstanding of UK wall sockets.    Subject to other loads on the circuit, any BS1363 UK wall socket can supply 13A at nominally* 240V, i.e 3.1 KVA.   

Unless the PSU design is ludicrously inefficient, 30A @ 13.8V DC shouldn't require more than 1 KVA AC in, a third of the max capacity of a BS1363 socket.

* The supply network in general remains at pre-EU harmonisation voltage levels, as the harmonisation was a paper exercise achieved by shifting the tolerance bands.
Assuming he's using the usual rectifier, capacitor and linear regulator topology, it wouldn't surprise me if he'll approach 3kVA, at full load, because of the huge regulator losses and poor power factor to boot.
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: MOT (microwave oven transformer) cooling.
« Reply #9 on: January 03, 2019, 07:16:28 pm »
I think that would fall under 'the PSU design is ludicrously inefficient'.
LTspice sim of a bridge rectifier with 30A DC load and 2.5V headroom at the ripple trough for feeding a 13.8V regulator, at 90% of nominal line voltage attached.  The secondary current is 47.9A RMS, which is slightly over 1KVA.

I'll grant you that if you don't carefully optimise your choice of secondary voltage, rectifiers, reservoir capacitance and regulator design, it will be hard to get right down to 1KVA in, but if it needs more than 1.5KVA after the switch-on surge, you'd better be building a room heater, not a PSU!
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: MOT (microwave oven transformer) cooling.
« Reply #10 on: January 03, 2019, 09:12:00 pm »
I think that would fall under 'the PSU design is ludicrously inefficient'.
LTspice sim of a bridge rectifier with 30A DC load and 2.5V headroom at the ripple trough for feeding a 13.8V regulator, at 90% of nominal line voltage attached.  The secondary current is 47.9A RMS, which is slightly over 1KVA.

I'll grant you that if you don't carefully optimise your choice of secondary voltage, rectifiers, reservoir capacitance and regulator design, it will be hard to get right down to 1KVA in, but if it needs more than 1.5KVA after the switch-on surge, you'd better be building a room heater, not a PSU!
Point taken: it's nowhere near as bad as I previously thought!
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
Re: MOT (microwave oven transformer) cooling.
« Reply #11 on: January 03, 2019, 09:21:14 pm »
Thanks for your replys.  I have reclaimed the MOT from a prevoius project, the transformer is wound with 12AWG silicon wire, and sleaved with shrink wrap tubing. The MOT puts out 14.70 Volts no load. In the previous project it put out 21.00 Volts rectified and filtered with 90000uf capacitance from 9 x 10000uf can electrolytics. I only ran it at 10 to 12 Amps as I was only using two MJ11015 power transistors. Noted at the time of use the output was very stable, although I wasn't that happy this 80°C surface temperature of the transformer.

From what I've read, class H 220 can go up to 180°C. So seems like plenty head room. Although the shrink wrap of the secondary will only go to 125°C. I know primary power efficiency isn't great, but for radio linear amplifier use it gets the job done. I'm just looking for cooler transformer running temperatures. Since posting I've found some thermally conductive epoxy, and have some decent area cpu heatsinks. Hopefully coupled with a couple of 100mm axial fans blowing over it, that I hope would keep things a bit cooler. The only negative is the epoxy rated to 150°C isn't cheap. I've also acquired another MOT after a new microwave oven upgrade.
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
Re: MOT (microwave oven transformer) cooling.
« Reply #12 on: January 03, 2019, 09:32:13 pm »
In the first project I used a low drop out voltage regulator on the regulator control board. Seems fine on input to the regulator voltage. I measured at the time of using the first incarnation. The output barely lost or dropped voltage under a 12 Amp load. From memory about 0.3 Volts of the 14.70 Volts going in to the bridge rectifier. The " Stiff" output AC was one of the positives with the MOT, baring in mind the primary is wasteful at full mains voltage.
Out of interest, can a triac type circuit control the primary voltage ? I know it chops the waveform AC, but wandered if such a circuit on the mains input would work.
 

Offline Wolfgang

  • Super Contributor
  • ***
  • Posts: 1826
  • Country: de
  • Its great if it finally works !
    • Electronic Projects for Fun
Re: MOT (microwave oven transformer) cooling.
« Reply #13 on: January 03, 2019, 09:45:58 pm »
I'm looking for some ideas to cool a single MOT. I've used a few single transformer in the past, but as most know they run hot as a single unit. Knowing that current limiting is the best option for better cooling and efficiency, it's not practical for the foot print I'm looking for.
As for cooling a single unit fan alone is not that great, I hit the high 80°C when loaded. I realise that's not that hot for a MOT, but the core temperature is sure to be far higher.

Having had a think on it, I've come up with modified air cooling using heatsinks attached to the transformer frame to increase surface area of the iron. But attaching the heatsinks is a bit of problem. I know there are thermal glues, but sure these are for small heatsinks, and there transfer temperature rating isn't that good. So I was wondering if its viable to drill and tap two M3 holes on each side to aid attaching the heatsinks ? My concern here is not wanting to upset the laminations with possible shorting. Any thoughts appreciated, my other idea is oil cooling in a suitable casing that can be sealed. It seems a messy headache to go through, and I'm unsure of longevity of the cooling oil.
I'm learning towards heatsinks, but don't want to degrade the iron core.
If I could current limit with out adding bulk I would do that, I thought about a triac type dimmer switch of up to 1000 watts, but gather these are not suitable as the modify the sine wave of the AC and will probably over heat the primary winding. Also I think there not suited to inductive loads.
Thoughts at ideas appreciated.
Thanks for reading.  :)

Some thoughts about your choice:

- MOTs are  made for a specific resonant circuit involving a rectifier, a reservoir cap and a magnetron, not for general use
- They are normally cooled by the fan of the oven, and they normally run short-term (< 30min).
- MOT cores have an iron shunt that creates a stray inductance.
- One HV side is grounded.

I have seen MOTs used for RF tube amps, but with several modifications:
- two MOTs were used, to emulate a split-winding secondary.
- The iron shunt was removed
- An inductor was added before the MOT to keep dissipation at line voltage low enough (without the resonant circuit, the core would saturate)
- A bridge rectifier was used for the heater circuit. Voltage is a few Volts depending on type, but you could use the MOT heater secondaries in series.

What to expect:
- low efficiency
- cooling is still neccessary
- cheap (you can get MOTs from the dump) but large

Alternatives
- if you just need a 12V heater source, why not try a halogen lamp lighting transfomer (up to 1kVA, not too expensive)
- or, you rip off the high voltage winding from your MOT and add just a few windings of thick wires as a secondary.
  This works even as a welding transformer if the wire is thick enough.
« Last Edit: January 04, 2019, 09:59:30 am by Wolfgang »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: MOT (microwave oven transformer) cooling.
« Reply #14 on: January 04, 2019, 09:47:52 am »
If you go for a halogen lamp transformer, make sure it's the iron core type. Most modern halogen lamp transformers have an oscillator driving a small high frequency transformer. The output is a 20kHz to 200kHz square wave modulated by double the mains frequency.

Why not just use a switched mode power supply? A 1kW unit isn't that expensive and will probably safe money on the electricity bill.
 
The following users thanked this post: thedoc298

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: MOT (microwave oven transformer) cooling.
« Reply #15 on: January 04, 2019, 04:32:53 pm »
You are trying to build a PSU that will supply up to 30 A at 13.8 V? Regulated? That sounds pretty easy and there are many designs out there. It is about 400 W which is not that much.

But you'd be better off going with a specific design than trying to reuse components that are ill suited for the job.

IMHO a SMPS is better suited for this end. In any case, a transformer or transformers followed by bridge rectifier and capacitor has a terrible wave form power factor which means everything will need to be oversized. I would go with a SMPS with power factor correction included. You will save bulk, heat, and probably cash.

You can search for Texas Instruments Power Factor Correction Circuits as they have resources online.

I guess I am getting older and lazier but I would not spend my time designing and building things which are very commonly sold commercially and can be bought for much less money than my time is worth.  You can buy a PSU with those specs which will be better designed and put together than anything you can do with repurposed microwave oven transformers.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: MOT (microwave oven transformer) cooling.
« Reply #16 on: January 04, 2019, 04:55:12 pm »
It would be silly to try building a SMPS like that. You can get surplus hot-swap server PSUs for not much more than the cost of shipping. These are mostly 12V at 30-60A but many can be tweaked to increase the output voltage slightly.
 

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: MOT (microwave oven transformer) cooling.
« Reply #17 on: January 04, 2019, 05:10:22 pm »
It would be silly to try building a SMPS like that. You can get surplus hot-swap server PSUs for not much more than the cost of shipping. These are mostly 12V at 30-60A but many can be tweaked to increase the output voltage slightly.

or just get one of these in 12V or 15V (adjustable +/- ~15%) https://www.meanwell.com/Upload/PDF/LRS-350/LRS-350-SPEC.PDF

$36.42 from Digikey
 
The following users thanked this post: Kilrah

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: MOT (microwave oven transformer) cooling.
« Reply #18 on: January 04, 2019, 05:15:59 pm »
1a. Reduce input voltage to, probably 80-100V (use another transformer in buck configuration).  Adjust secondary turns accordingly.
1b. Equivalently, add more primary turns (in the gap left from the magnetic shunts?) to the same end.

These both reduce capacity, but mind that an MOT is only about 500VA in size, for a typical (cool running, good regulation) design.  Don't be greedy about their ratings!

The reduction in V/turn will save input (magnetizing) current, and mechanical noise, because the core is not saturating every half-cycle. 

2. Sure, go ahead and bolt or clamp some heatsinks to it.  Iron's not the greatest thermal conductor, but there's a lot of it.  You'll need a lot of thermal goop to fill in the somewhat uneven laminated surface.

If possible, stick a thermistor inside the winding so you can monitor wire temperature.  Use high temp wire if possible, too.

3. Just forget about them.  More effort than it's worth?  A nice MeanWell switcher that size is only a few hundred bucks, and will perform very nicely indeed!

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline davelectronicTopic starter

  • Frequent Contributor
  • **
  • Posts: 772
  • Country: gb
  • Life's too short.
Re: MOT (microwave oven transformer) cooling.
« Reply #19 on: January 04, 2019, 07:47:08 pm »
Thanks again all for your replys. It's the something for nothing, or next to nothing that's appealing. They are far from perfect or efficienct I know this. A first stage transformer sounds interesting idea..
I have had other projects from the past with bought toroidal transformers for. Yes I could purchase another transformer. I've lost count of the smps units I have modified over the years, from good quality ATX units to very powerful server psu's.
It's just for the fun of it really. The MOT is never going to be fantastic, adding surface area creates bulk. Andding ballast creates weight, as does second transformers. First I will try heatsinking with thermal epoxy, see how that goes.
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: MOT (microwave oven transformer) cooling.
« Reply #20 on: January 04, 2019, 08:34:54 pm »
Personally if I needed a clean 13.8V supply with 30A peak load currents and a moderate to low duty cycle, I'd float a Lead acid battery.  As RF is involved, using a linear charger (or full cycle line frequency TRIAC controlled on the secondary side) would be advisable.  10A max current from the charger would handle any reasonable TX duty cycle for non-broadcast applications, and at about 350VA with a well insulated extra winding to buck 10% of the line voltage, any MOT should run cool enough without additional heatsinking.
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: MOT (microwave oven transformer) cooling.
« Reply #21 on: January 04, 2019, 09:20:34 pm »
Heat tends to be the limiting factor for power in a transformer. You can easily run transformers at 500% rated load, but not for long before it overheats, so it makes such operation reasonable only in cases where the duty cycle is very short.

Most of the heat comes from the copper wingdings so adding heatinsking to the core likely won't do a whole lot. It will certainly make the transformer feel like its running cooler since the core will certainly be cooler to the touch but deep inside the layers of copper coils there will still be a significant amount of heat. It does help, but not a ton. Having a powerful fan blow over the transformer is likely plenty enough cooling for how much heat can propagate out of deep inside the wingdings.

The problem is that its quite hard to tell when a transformer is running comfortably without having a temperature probe deep inside the wingdings. At a certain temperature the insulation on the wire starts to slowly degrade, this can cause a failure after a few years. If its getting hot enough that you can smell it then its really getting too hot and will probably start breaking down after days of such operation. Such breakdowns can be quite spectacular sometimes as they include lots of fire, sparks and smoke as turns short themselves and get hot causing the failure to propagate trough the wingdings until the whole thing is fried and the fuse blows.

When space and efficiency is a big factor its impossible to beat switchmode supplies.
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: MOT (microwave oven transformer) cooling.
« Reply #22 on: January 04, 2019, 10:36:20 pm »
Personally if I needed a clean 13.8V supply with 30A peak load currents and a moderate to low duty cycle, I'd float a Lead acid battery. 
Yup. I used to do that. Here you could get used car batteries for free. They were not good for starting cars but still plenty good for my ham radio rig. Nowadays used lead batteries are so tightly controlled you'd think they were heroin.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: MOT (microwave oven transformer) cooling.
« Reply #23 on: January 04, 2019, 10:59:32 pm »
Lead acid batteries are easy enough to get in the USA but I don't like keeping them in my house. They inevitably produce corrosive vapor, they release flammable hydrogen when charging, they will often leak if knocked over, they can supply vast currents for short periods of time which can cause a fire, they are heavy and bulky. Many of these issues are not too hard to mitigate but it's still not something I really enjoy messing with.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: MOT (microwave oven transformer) cooling.
« Reply #24 on: January 04, 2019, 11:12:16 pm »
It would be silly to try building a SMPS like that. You can get surplus hot-swap server PSUs for not much more than the cost of shipping. These are mostly 12V at 30-60A but many can be tweaked to increase the output voltage slightly.

or just get one of these in 12V or 15V (adjustable +/- ~15%) https://www.meanwell.com/Upload/PDF/LRS-350/LRS-350-SPEC.PDF

$36.42 from Digikey
Yes, that's what I'd do.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf