It would seem that MOT - microwave oven transformer - is a common topic but I have not found definitive answers to what they can be reused for.
As Output:
According to http://www.diyaudio.com/forums/tubes-valves/108482-mots-output-transformers.html they are no good.
As plate chokes:
Shoog seems to be a advocate that, I am guessing, a unmodified MOT can work.
http://www.diyaudio.com/forums/tubes-valves/86263-another-cheap-fix-mot-plate-choke.html
http://www.diyaudio.com/forums/tubes-valves/101419-cheap-lunch-2.html
But other times Shoog states MOT do not have enough inductance.
http://www.diyaudio.com/forums/tubes-valves/222175-microwave-transformers.html
This is not a call out on Shoog but I simply wish to know if he/she tried using an MOT as a plate choke and simply decided they were no good.
As Power Chokes:
It seems like any other transformer to choke converison it should be gapped; however, Dr. Photon suggests that this isn't necessary in http://www.diyaudio.com/forums/parts/52779-microwave-transformer-choke-4.html.
I personally would like to know if there is more definitive answers to reusing a MOT. I believe Microwave Oven dissassembly is dangerous and should be forum wide discouraged. I should hate to see something happen like what happened at this link http://www.instructables.com/id/Build-a-Microwave-Transformer-Homemade-Welder/. A mother wrote about that during disassembly of a microwave, her son was killed (post 83, on page 3). If MOT's do have a use I would rather it be plainly stated. However, like many things I would rather see people buy - spend money - on proper designed materials for a project.
As Output:
According to http://www.diyaudio.com/forums/tubes-valves/108482-mots-output-transformers.html they are no good.
As plate chokes:
Shoog seems to be a advocate that, I am guessing, a unmodified MOT can work.
http://www.diyaudio.com/forums/tubes-valves/86263-another-cheap-fix-mot-plate-choke.html
http://www.diyaudio.com/forums/tubes-valves/101419-cheap-lunch-2.html
But other times Shoog states MOT do not have enough inductance.
http://www.diyaudio.com/forums/tubes-valves/222175-microwave-transformers.html
This is not a call out on Shoog but I simply wish to know if he/she tried using an MOT as a plate choke and simply decided they were no good.
As Power Chokes:
It seems like any other transformer to choke converison it should be gapped; however, Dr. Photon suggests that this isn't necessary in http://www.diyaudio.com/forums/parts/52779-microwave-transformer-choke-4.html.
I personally would like to know if there is more definitive answers to reusing a MOT. I believe Microwave Oven dissassembly is dangerous and should be forum wide discouraged. I should hate to see something happen like what happened at this link http://www.instructables.com/id/Build-a-Microwave-Transformer-Homemade-Welder/. A mother wrote about that during disassembly of a microwave, her son was killed (post 83, on page 3). If MOT's do have a use I would rather it be plainly stated. However, like many things I would rather see people buy - spend money - on proper designed materials for a project.
All MOTs are grossly undersized for cost reasosns, as they are only operated intermittently.
To thin copper wire => to much I2R loss
To little iron, or to few turns => high flux density => high iron loss
This means that any continous use probably needs new windings. That is true for both primary (230VAC) and secondary winding (~2kVAC)
they are constructed with a magnetic shunt , which takes a part of the magnetic flux decoupling the primary/secondary increasing the leakage inductance. This is what is used when they are used as ballasts (series inductor). When you do this you short the secondary, which means that the primary magnetic inductance is gone, left is only the leakage inductance.
The good thing about leakage is that in theory it is air cored , thus unsaturable. In this case it is through the magnetic shunt so there is a limit to the current it can take.
To thin copper wire => to much I2R loss
To little iron, or to few turns => high flux density => high iron loss
This means that any continous use probably needs new windings. That is true for both primary (230VAC) and secondary winding (~2kVAC)
they are constructed with a magnetic shunt , which takes a part of the magnetic flux decoupling the primary/secondary increasing the leakage inductance. This is what is used when they are used as ballasts (series inductor). When you do this you short the secondary, which means that the primary magnetic inductance is gone, left is only the leakage inductance.
The good thing about leakage is that in theory it is air cored , thus unsaturable. In this case it is through the magnetic shunt so there is a limit to the current it can take.
Thank you, rikkitikkitavi for you information. It does make a lot sense for these transformers are made as cheap as possible. Is it fair to say that MOT are useless (especially, if welded together), unless much laborious work is done to correct a flawed item.
I think it is logical to decide that Mircowave oven's should be recycled and left well enough alone.
I think it is logical to decide that Mircowave oven's should be recycled and left well enough alone.
I would rate the core to about 300VA continiously, and if adding a few more primary turns the split bobin makes a quite safe transformer. The magnetc shunt can often br knocked out without cutting the core open.
You have a fairly high degree of freedom rewinding the secondary.
Nowadays it gets more common with a switched powersupply, esprcially if the oven wants to meet the current demand on power factor etc etc.
MOTs have their uses, but it takes some efgort to get to know them.
You have a fairly high degree of freedom rewinding the secondary.
Nowadays it gets more common with a switched powersupply, esprcially if the oven wants to meet the current demand on power factor etc etc.
MOTs have their uses, but it takes some efgort to get to know them.
So you must increase the primary turns by about 25% to get to reasonable no load core loss and magnetizing VA current.
I have found most operate at 1.7T flux levels, some as high as 1.9T when operating at a high line voltage, 125 rather than 115vac. depending on the application and what you consider reasonable no load temperature rise, you may have to add 30-50% more turns to the primary to get rid of the humm and reduce temperature rise.. but this comes at a cost of copper losses.
I suppose it does make sense to use the stock secondary as a plate choke. The resistance is typically 60 ohms, 2200 turns but might be as high as 99 ohms for an aluminium secondary in a 500w microwave, or as low as 40 ohms from a very old kilowatt sized microwave.
given there are about 4 typical core sizes, it isn't hard to find a second microwave and fit both secondary coils in series or parallel for use as a choke.
I think its pretty typical to be able to fit two secondary coils on the same core.
I have found most operate at 1.7T flux levels, some as high as 1.9T when operating at a high line voltage, 125 rather than 115vac. depending on the application and what you consider reasonable no load temperature rise, you may have to add 30-50% more turns to the primary to get rid of the humm and reduce temperature rise.. but this comes at a cost of copper losses.
I suppose it does make sense to use the stock secondary as a plate choke. The resistance is typically 60 ohms, 2200 turns but might be as high as 99 ohms for an aluminium secondary in a 500w microwave, or as low as 40 ohms from a very old kilowatt sized microwave.
given there are about 4 typical core sizes, it isn't hard to find a second microwave and fit both secondary coils in series or parallel for use as a choke.
I think its pretty typical to be able to fit two secondary coils on the same core.
I have a MOT disassembled and it seems this is the specs on it. The center of the E is 35.5mm across. The voids are 16.9mm across. Total width is 105.3mm. Each lamination is 0.0218"= 0.55372 mm. The total length of the transformer is ~80mm (guessing based off measuring inside of primary windings.
The primary windings are 0.0752" - 1.91mm in varnish so am guessing 13 awg (0.072"). I count 10 windings across and 10 windings deep. Rough guess of 100 turns.
The secondary windings I am not concerned with.
By following the guide at:
Practical Transformer Winding
I have begun de-laminating the transformer and I will straighten and re-coat each lamination.
This information seems to match with what as been posted. As the primary turns seem to be 25 turns short (or 25%) for a 1 turn per volt on the primary. However, accounting for loses 150 turns would probably be best.
I have no practical goal with this transformer but a power choke or a 300-0-300 VAC with 6.3 VAC secondary would be cool.
Since I am a beginner it seems that maybe getting bobbins would be safest.
The primary windings are 0.0752" - 1.91mm in varnish so am guessing 13 awg (0.072"). I count 10 windings across and 10 windings deep. Rough guess of 100 turns.
The secondary windings I am not concerned with.
By following the guide at:
Practical Transformer Winding
I have begun de-laminating the transformer and I will straighten and re-coat each lamination.
This information seems to match with what as been posted. As the primary turns seem to be 25 turns short (or 25%) for a 1 turn per volt on the primary. However, accounting for loses 150 turns would probably be best.
I have no practical goal with this transformer but a power choke or a 300-0-300 VAC with 6.3 VAC secondary would be cool.
Since I am a beginner it seems that maybe getting bobbins would be safest.
Bando you have a larger than normal core and the 100 turns you counted make for a 1.6T field strength not accounting for the stacking factor of the core. --this 1.6T is close to or below average for MOT's
To get to 1.3T which I've found is acceptable for most mot re-purposing, (also not accounting for stacking factor) you only need to add 25 turns to the primary.
So I would keep it as is and just add more turns.
De-laminating the core is very labor intensive because you have to file off those welds. you shouldn't need to recoat the laminations, but if you do the stacking factor will be significantly less than it is now, and you'll either have to wind a new primary, or leave out some laminations.
also you may put it back together and find a higher core loss anyways.. this has been my experience taking apart APC ups transformer (but not recoating the laminations)
To get to 1.3T which I've found is acceptable for most mot re-purposing, (also not accounting for stacking factor) you only need to add 25 turns to the primary.
So I would keep it as is and just add more turns.
De-laminating the core is very labor intensive because you have to file off those welds. you shouldn't need to recoat the laminations, but if you do the stacking factor will be significantly less than it is now, and you'll either have to wind a new primary, or leave out some laminations.
also you may put it back together and find a higher core loss anyways.. this has been my experience taking apart APC ups transformer (but not recoating the laminations)
You will have found that the primary fills a lot less than 50% of the winding window.
Thus there is lots of room at add extra Primary turns.
If the existing is 10T wide then add whole layers of 10 T, i.e. add 30T + 10T + 10T. (or maybe 29T+9T+9T, if you want extra insulation at the edge of each new winding.)
Now series connect the extra 50T to the existing primary and connect via a Variac to the Mains. (Bulb tester first).
Plot the S curve of Ipri vs V prim from 0Vac to your maximum stated for your supplier, typically 127Vac in USA and 254Vac in UK & most of Europe.
You will see the primary current run away at the highest voltages. If the run away is not bad, then retest using the +40T tapping and plot the new S curve.
then with +30T.
Choose whichever tapping gives a reasonable maximum primary current at your max supply.
That gets the Flux into a good cool running level.
This can all be done without any secondary winding.
Once you have the primary completed and well insulated, (take special care to insulate the tappings well, they are at mains voltage) you can then wind on your own secondary of your choosing. It will FILL the remaining window to get the best regulation and best VA for your transformer.
If you underfill, you will get less performance.
Thus there is lots of room at add extra Primary turns.
If the existing is 10T wide then add whole layers of 10 T, i.e. add 30T + 10T + 10T. (or maybe 29T+9T+9T, if you want extra insulation at the edge of each new winding.)
Now series connect the extra 50T to the existing primary and connect via a Variac to the Mains. (Bulb tester first).
Plot the S curve of Ipri vs V prim from 0Vac to your maximum stated for your supplier, typically 127Vac in USA and 254Vac in UK & most of Europe.
You will see the primary current run away at the highest voltages. If the run away is not bad, then retest using the +40T tapping and plot the new S curve.
then with +30T.
Choose whichever tapping gives a reasonable maximum primary current at your max supply.
That gets the Flux into a good cool running level.
This can all be done without any secondary winding.
Once you have the primary completed and well insulated, (take special care to insulate the tappings well, they are at mains voltage) you can then wind on your own secondary of your choosing. It will FILL the remaining window to get the best regulation and best VA for your transformer.
If you underfill, you will get less performance.
Last edited:
I have been thinking of enclosing the transformer in a steel box with plastic top and have it filled with mineral oil. I have read this improves flux capabilities of the transformer itself. To my knowledge the greater the flux the better the voltage regulation.
I am going to re-coat each lamination but I have considered using polyurethane for electric insulation and also reducing lamination vibration.
Thank you all for your input thus far it has been very helpful and encouraging.
I am going to re-coat each lamination but I have considered using polyurethane for electric insulation and also reducing lamination vibration.
Thank you all for your input thus far it has been very helpful and encouraging.
fractional hp electric motors are known for using the thin steel shell as part of the magnetic circuit, given that they run the core fully saturated, so a significant amount of flux does flow through the shell.
however this will only reduce core loss a tiny bit for a transformer.. and actually may increase eddy current loss if the steel tank is significantly thick.. and if its not close to the transformer the only thing it will do is humm.
Note that for a lightbulb in series with the line you need a 500 watt bulb, its going to draw at least 5 amps no load current* on a stock primary.
once you add the extra turns to take the flux down to 1.3T you should be around 1 amp or less of magnetizing current.
*perhaps 20% of this is core and copper losses, the rest is inductive.
however this will only reduce core loss a tiny bit for a transformer.. and actually may increase eddy current loss if the steel tank is significantly thick.. and if its not close to the transformer the only thing it will do is humm.
Note that for a lightbulb in series with the line you need a 500 watt bulb, its going to draw at least 5 amps no load current* on a stock primary.
once you add the extra turns to take the flux down to 1.3T you should be around 1 amp or less of magnetizing current.
*perhaps 20% of this is core and copper losses, the rest is inductive.
Bando, immersing the transformer in mineral oil will help cooling it and thus it is possible to tolerate a higher power loss. In a well designed transformer the magnetic losses are low compared to resistive losses in windings, and the later has nothing to do with flux density. By increasing wire diameter and filling the window you will lower the resistive losses and improve regulation, get a more "stiff" transformer.
A normal MOT is driven close to saturation, but the regulation is shot anyhow due to the leakage shunt. It will always show a large idle loss due to magneics but this will go way up when power is drawn by the magnetron. Also remember that there is a fan in there cooling both magnetron and transformer. ¨
Is it common today with a switched power supply or is it still the old steel-n-copper chunk?
A normal MOT is driven close to saturation, but the regulation is shot anyhow due to the leakage shunt. It will always show a large idle loss due to magneics but this will go way up when power is drawn by the magnetron. Also remember that there is a fan in there cooling both magnetron and transformer. ¨
Is it common today with a switched power supply or is it still the old steel-n-copper chunk?
true, but, putting a transformer in a box filled with oil may not actually increase surface area available for cooling, especially if we can blow air between parts of the coils.
not true.. for a transformer seeing continuous duty, the optimal efficiency is at the point where they are approximately equal.--we can compromise this for a variety of reasons but only for small transformers. once you get to a few megawatts you are forced to build efficient machines because you would have to make the conductors hollow and pump water through them to cool them off.
Most factors, except leakage inductance favour large machines, but even wall warts deal with 30% leakage inductance.--on the small end this is a separate problem
basically the practical method to deal with very low intrinsic efficiency is to regulate away the primary copper losses by running the core saturated.
i think the highest coupling you can get for 1.9T iron and 60 HZ is in the range of 100 kw but don't quote me on that.
for gigawatt sized grid transformers, 1 watt per kilogram iron loss, and 1 watt per pound copper loss makes a 99.8% efficient transformer operating at 1 kilowatt per kilogram.. (hey, we can pray right?
)
true, fill it up full of copper.
true, but that was the point, to regulate the input line voltage away to deliver constant power to the magnetron.
no load idle losses are high because no one cares how many kilowatt hours per year it takes to heat up their coffee.
magnetic amplifiers, properly designed are very efficient and still used to this day. the magnetic shunts in between the primary and secondary are just the same as the magnetic shunts used in many arc welding supplies.. and if they made the transformer twice the size, 8 times the weight, and 8 times the price, it would be 95% efficient. but no one cares about the efficiency of their microwave so it is perhaps 75% efficient.
so today i had the chance to try and fix a mpja branded 0-30 volt 0-10 amp power supply. it never did work right from the beginning but recently the voltage/current loop died.
i was able to get it to work by disabling the current loop... meaning that if i shorted the output the current would rise to whatever the input impedance could deliver.
i set the supply to 1.1 volts.. shorted the output and got 29 amps on the meter.
so, output regulation? good enough for me.
Last edited:
And you can put a fan blowing on the oilbox aswell, or circulate oil through a external radiator. The oil WILL remove heat more efficiently as long as you have a fair amount of autoconvection inside the box preventing stagnant layers of hot oil unless you move up the air speed into significant Mach numbers.
About losses split- isnt the optimum point 50/50 shared losses resisitive/magnetic an economical optimization, as most "smaller"(a few hundred VA aboutish) transformer have a 80/20 or 90/10 spilt resistive/magnetic?
It is a matter of sufficient cooling aswell.
Well, you are probably right as I have not designed any transformers apart from a small DIY switcher only broken a few and dismantled a few more...
I dont remember, but if hte MOT wasnt such a lousy regulator (ie constant power) wouldnt the magnetron short our, hence the need for a high series impendance (created by the leakage) analog to how a arc welder works out?
About losses split- isnt the optimum point 50/50 shared losses resisitive/magnetic an economical optimization, as most "smaller"(a few hundred VA aboutish) transformer have a 80/20 or 90/10 spilt resistive/magnetic?
It is a matter of sufficient cooling aswell.
Well, you are probably right as I have not designed any transformers apart from a small DIY switcher only broken a few and dismantled a few more...
I dont remember, but if hte MOT wasnt such a lousy regulator (ie constant power) wouldnt the magnetron short our, hence the need for a high series impendance (created by the leakage) analog to how a arc welder works out?
no, optimal efficiency is at core losses=copper losses. but this is for 24/7 operation and no change in primary voltage..
microwave ovens are designed to include the 2 volts they lose in the 6 feet of 18 gauge wire they have for a plug..
yes we could argue about oil or air cooling but whatever.. just don't expect a passive 6 by 6 by 6 inch cube to cool itself better than a mot with a fan on it.
magnetrons are just like diodes. but with a 4 thousand volt voltage drop, and a 400+? ohm resistance.. it takes 5KV to push 1 amp through. but only 4 KV to push a tenth of an amp through..
those magnetic shunts are so you don't burn the transformer out... not the magnetron btw.
microwave ovens are designed to include the 2 volts they lose in the 6 feet of 18 gauge wire they have for a plug..
yes we could argue about oil or air cooling but whatever.. just don't expect a passive 6 by 6 by 6 inch cube to cool itself better than a mot with a fan on it.
magnetrons are just like diodes. but with a 4 thousand volt voltage drop, and a 400+? ohm resistance.. it takes 5KV to push 1 amp through. but only 4 KV to push a tenth of an amp through..
those magnetic shunts are so you don't burn the transformer out... not the magnetron btw.
No argument with me on the cooling, sometimes one is better yhan the other. It all depends on surface and convective heat transfercoefficient, the later is significantly increased by a fan of course.
I did some experiment with a 300VA toroid pushing DC through prim and sec at rated current, measuring resistance increase . A huge fan gave about 15-20% more current capacity for the same temperature increase. Limited by primary ( innermost) winding.
I guess a EI core has more surface so perhaps 30-40% benefit by fan cooling?
Interesting info about the magnetron. Didnt know that.
Regards Rickard
I did some experiment with a 300VA toroid pushing DC through prim and sec at rated current, measuring resistance increase . A huge fan gave about 15-20% more current capacity for the same temperature increase. Limited by primary ( innermost) winding.
I guess a EI core has more surface so perhaps 30-40% benefit by fan cooling?
Interesting info about the magnetron. Didnt know that.
Regards Rickard
So I messed up the primary windings. I wore the varnish off the wire and it was grounding to the steel (checked with DVOM). It turns out the wire is aluminium anyways, so I am going to start from square one. I am hoping you guys could check my math.
Cross section is 3.55 * 7.6 cm = 26.98 cm2 = 0.002698 m2
Finding Weber is Wb = T(esla) * Cross section
Wb = 1.3 * 0.002698 = 0.0035074 (Tesla value is as recommended by johansen post 7)
From what I understand you double this figure for it is only one direction; whereas, the voltage is a sine wave and having a negative and positive pull.
So the Wb = 0.0070148
To find the turns needed for 125 VAC
t = V*s/Wb (time [value s] used was 0.01s apparently this is for 50 hz but I wanted to over spec)
t = 178
I hope this is so far right.
If this is so then the inductance of the primary is about 1 H. Using the equation of H = Wb*t/A with an assumed amperage of 1.
Cross section is 3.55 * 7.6 cm = 26.98 cm2 = 0.002698 m2
Finding Weber is Wb = T(esla) * Cross section
Wb = 1.3 * 0.002698 = 0.0035074 (Tesla value is as recommended by johansen post 7)
From what I understand you double this figure for it is only one direction; whereas, the voltage is a sine wave and having a negative and positive pull.
So the Wb = 0.0070148
To find the turns needed for 125 VAC
t = V*s/Wb (time [value s] used was 0.01s apparently this is for 50 hz but I wanted to over spec)
t = 178
I hope this is so far right.
If this is so then the inductance of the primary is about 1 H. Using the equation of H = Wb*t/A with an assumed amperage of 1.
I'm not used to checking 110/120Vac primaries.
But from my 220/240Vac experience, 1H seems very low.
Xl = 2PiLF = 2*3.14*1*60 = 377ohms.
The primary will present that impedance to your mains voltage.
The off load current will be ~ 115/377 ~ 300mAac.
That is a factor of 10 to 30 too high.
I would expect 10mA to 30mA at nominal voltage.
That is what I would measure using the procedure in post8
But from my 220/240Vac experience, 1H seems very low.
Xl = 2PiLF = 2*3.14*1*60 = 377ohms.
The primary will present that impedance to your mains voltage.
The off load current will be ~ 115/377 ~ 300mAac.
That is a factor of 10 to 30 too high.
I would expect 10mA to 30mA at nominal voltage.
That is what I would measure using the procedure in post8
only a good quality toroid can deliver that specification.
read here: http://toroid.com/standard_transformers/isolation_transformers/DataSheets/749.1202.pdf
8.1 pound transformer, 2.2 watts no load loss, with 22ma no load current makes 2.64 VA.. i suspect this is not accurate, but 50mA is believable.
a stock mot will pull 4 to 6 amps no load, you can work out the inductance from that.
http://johansense.com/bulk/14gastats.PNG
once you add 27 turns to the stock primary, no load amps drops from 3.46 amps to .78 amps.. you can read the rest of the data from the chart.
the transformer in that photo was from a very old kenmore microwave and it was one of the much more efficient ones.
http://johansense.com/bulk/14ga2.PNG
and yes the no load loss and no load current runs away after 105 volts.
http://johansense.com/bulk/motlossbig.png
X axis is volts, orange is watts against the right Y axis, blue is amps in the left Y axis.
the step in the orange axis is because my killawatt meter was defective.
I have since bought a killowatt hour meter (house type, analog) and fitted the rotating spindle with a coil spring and i have to say i trust it more, even if i can only get perhaps 7 bits of dynamic range from it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.