transformer powered heating system no load / load voltages

Hi all,

My current project is a heating system with 5 coils independently driven by 20n5560 mosfets. (rough schematic attached) They share a common power supply (not pictured). That supply is a large transformer + rectifier(full bridge) + bulk cap. 2 additional power supplies:a 9v ps powering the arduino and a 12v driving the opto outputs. With no load on the system, the bulk cap charges to 60vdc. I understand that to be the peak voltage coming off the rectifier. Turning coils on, drops that voltage to 40vdc under typical use. That part makes sense to me as that is close to the average voltage from the rectifier.

Here is where I get confused: with a coil at full bore (pwm 255), the voltage across is coil is at most 25vdc. Why don’t I get the full 40vdc? My first though was I might be overloading the transformer and it just cant keep up. So I looked at the bulk cap supply voltage & the coil voltage side by side and sure enough the bulk cap maintains ~40vdc while only 25v is seen across the coil itself.

Is this typical or indicative of some other problem? If it is typical, why does this happen? where has that 15vdc gone? If not, any troubleshooting ideas would be welcomed. Thanks all!

Please provide a link to a datasheet for the mosfet you used, it’s not a common part number.

Your schematic is unfinished, it lacks the how and where the mosfet gate voltage is sourced. Lots of missing grounds. No free wheeling diodes across whatever the coils are.

Chances are you’re not driving the mosfets correctly. Do they get hot when turned on?

Ultimately, the lack of a usable schematic and device datasheet puts this post in the difficult to help category.

Transformer, heating coils, rectifier...
Wouldn't it have been easier to drive the heaters with straight AC, and switched with a relay.
Leo..

The partial schematic doesn't make sense. Nothing connected to pin 3 of the opto couplers, but I suppose ground. Nothing to pull the gates high. What does the '12V' indicate against R5 to R9? There is nothing there to supply 12V.

As stated by WhatsThat, impossible to give much help without proper and complete information.

I really should have been more clear, I am not looking for specific corrections to my circuit, its built and the arduino--> optos-->mosfet gates work fine. By that I mean, I have control over the voltage on each coil by way of PWM. And yes the heaters get hot. I am in the middle of tuning the system when I took note of the lower than expected maximum voltage.

mosfets are 20N6055, not 20n5560 as I prev stated) heres a link

Yah its a half assed the diagram, never used KIcad before and I found it rather rage inducing...There is a lot more to the setup than whats pictured (thermistors, buttons, display, ect) taking the time to make a complete schematic is just not worth it.
it was really more of a conceptual drawing. And the answers I am seeking are also more conceptual.

The partial schematic doesn't make sense. Nothing connected to pin 3 of the opto couplers, but I suppose ground. Nothing to pull the gates high. What does the '12V' indicate against R5 to R9? There is nothing there to supply 12V.

just assume those pin connection are correct. The 12v is an independent(common gnd) power supply and it should be present on all 5 iterations of that circuit. When an arduino pin is written to 255, I see 5v @ the input to the opto and the voltage measured at the mosfet gate is 10.7vdc...I am not great with data sheets, but that should be about right, no?

Transformer, heating coils, rectifier...
Wouldn't it have been easier to drive the heaters with straight AC, and switched with a relay.
Leo..

the thought has certainly crossed my mind...though a relay would not be suitable for PWM, unless you meant an SSR. But there are 5 coils (orig 6), all that gets a bit bulky and expensive....I built this out of spare parts. In the future I would like to learn how to drive triacs through an arduino, I have thousands of them sitting around.

...it lacks the how and where the mosfet gate voltage is sourced.

its a standalone 12v smps, with ground common throughout the whole system. when a write a pin fully high (255), I see 10.7vdc at the gate.

No free wheeling diodes across whatever the coils are.

I was hoping someone would bring this up. Those have been a common practice of mine in the past, though generally when driving dc motors. My current (and likely flawed) understanding is that with a purely resistive load, as is the case with a coil of nichrome wire, that diode is not required. Even as I type that, it sounds wrong. please let me know what I have wrong there. Also this system has 10+ hours of heating time, why hasn't their absence caused a problem (probably a shorted mosfet?) While think this may be a problem, I doubt its accounting for the voltage drop.

also, are the coils not clear in the drawing?

thanks for all the quick responses

I too would drive the heaters from A/C .

With a high load you will gets lots of ripple on the DC

I really should have been more clear, I am not looking for specific corrections to my circuit, its built and the arduino--> optos-->mosfet gates work fine.

Very difficult to give advice which is both helpful and not specific, especially if you've built it and it works. The questions you asked in your first post were very specific so need specific answers, so cannot be answered in general terms, they can only be answered in specific terms for a specific circuit. The circuit posted has problems that might or might not be the source of the problems.

I would not drive heaters with high frequency* PWM. Heaters respond slowly, generally a few seconds to minutes or hours, and the things they heat respond even more slowly, think of how long it takes to heat a kettle or your living room. I use PWM with a cycle time of 720 seconds for my central heating, I suggest you use something similar. Once you do that the relays are OK, no need for a rectifier or diodes or capacitors. The other sensible way to drive them is with a triac from AC using burst control, so some number of cycles on, some number of cycles off. For example 1 cycle on 3 cycles off gives 1/4 power.

And yes the heaters get hot

Ermm, the question you were asked was do the MOSFETs get hot....
(They shouldn't)

Inductors need fly-back diodes, or rather their drivers need them. ANY coil of wire is an inductor, even a single length of straight wire is an inductor. The more relevant question is to ask if the inductance is enough to be significant in the situation at hand. I would think the inductance of a heating coil would be low enough to be ignored, but the only way to know for sure is to put an oscilloscope on the connection to the heater and watch what happens when the power is interrupted.

*In this case 'high frequency' is vague. I mean certainly above 1Hz, certainly above the frequency a relay can handle. Heaters can be switched slowly, maybe every few seconds to minutes.

mtraven:
Here is where I get confused: with a coil at full bore (pwm 255), the voltage across is coil is at most 25vdc. Why don’t I get the full 40vdc? My first though was I might be overloading the transformer and it just cant keep up. So I looked at the bulk cap supply voltage & the coil voltage side by side and sure enough the bulk cap maintains ~40vdc while only 25v is seen across the coil itself.

Is this with ALL your MOSFETS turned ON?
What are the specs/data of your “coils” or are they heater elements?

What is the DC voltage across the drain - source of the MOSFET.
Switch the DMM to AC mode and measure the DC power supply with and without load.
A full schematic including your power supply and transfomer will help?
What value filter capacitors are you using?
Can you please tell us your electronics, programming, arduino, hardware experience?

Thanks… Tom… :slight_smile:

You never said what the resistance of the heater elements are. Assume you have a 48VAC transformer, what is the VA rating (Volt Amps)?

Trouble sshooting 101.

ASSUME NOTHING

Wire one coil direct -test

Rewire with mosfet then drive mosfrt full on hard. That is max powersupply voltage to the base

Rewire base from power suppply to have 12v power direct drive base as a switch full on 12v to the base- test

Rewire add resistor and opto back in. Drive opto with full 20mA direct full on

Rewire, drive opto from arduino
Drive full on- test

Run pwm - test

I would expect the heaters to be restive loads not inductive so there would be no back emf No need for diodes
Need datasheet.

The OP said his DC rectified and filtered voltage was was 60, 60 / 1.414 = 42.4V, Full wave rectified DC = 0.9 VRMS, VRMS = 42.4 / 0.9 = 47.15. Thus I assumed the standard value of 240 / 5, 48.

I would expect the heaters to be restive loads not inductive so there would be no back emf No need for diodes

Expecting the unknown is pretty much assuming. Could be an induction heater...?

Yah I know im being a pain in the ass and yall want a perfect schematic, but the reality is the problem (if there is one) is probably not in a theoretical drawing, but some error in fabrication that would not show up on a schematic.

Ermm, the question you were asked was do the MOSFETs get hot....

mis understand the question, no they do not get hot...i have used these on many projects, in the early days I did not understand to drive them correctly & they did tend to get hot. They are also way over rated for the application driving 30ohm coils @ even 60v is only 2 amps. If someone could confirm that driving those gates @10.7v is within a reasonable range.

You never said what the resistance of the heater elements are. Assume you have a 48VAC transformer, what is the VA rating (Volt Amps)?

No idea, bought at surplus, rewound the secondary & added taps for several different voltage outputs. Currently running on the 45vac tap. Ask about VA rating suggests you think I might be overloading (saturating?) the transformer...I can't prove thats not the case, but the thing weighs about 10lbs and is wound with ~10 gauge enamel wire. Also have not seen at temps beyond 38c.

he OP said his DC rectified and filtered voltage was was 60, 60 / 1.414 = 42.4V, Full wave rectified DC = 0.9 VRMS, VRMS = 42.4 / 0.9 = 47.15. Thus I assumed the standard value of 240 / 5, 48.

Not sure what the notation "240 / 5, 48." means, but the primary gets 120vac 50hz and the secondary is as described above. The math you showed is constant with what I found theoretically and by measurement.

I would not drive heaters with high frequency* PWM. Heaters respond slowly, generally a few seconds to minutes or hours, and the things they heat respond even more slowly, think of how long it takes to heat a kettle or your living room. I use PWM with a cycle time of 720 seconds for my central heating, I suggest you use something similar. Once you do that the relays are OK, no need for a rectifier or diodes or capacitors. The other sensible way to drive them is with a triac from AC using burst control, so some number of cycles on, some number of cycles off. For example 1 cycle on 3 cycles off gives 1/4 power.

this is a cast iron heated print bed, not central heating. my PWM frequency is like 500 hz, I think one is linked to timer0, so its more like 900hz.

Someone asked about the coils, they are custom wound lengths (~30 ohms) of kanthal wire (sorry I said nichrome earlier)

Is this with ALL your MOSFETS turned ON?

yes and no -- with fewer, say just 1, mosfet running, the supply voltage measured at the bulk/filter cap is higher (~50vdc), but there is still a 10-15v less across the coil itself.

What is the DC voltage across the drain - source of the MOSFET.

that is the voltage in question and it shows up 10-15v less than the supply voltage.

Switch the DMM to AC mode and measure the DC power supply with and without load.

I have tried that in the past, as I recall it did not pick up anything, I took that to mean the ripple was sufficiently smoothed. This was a while ago, I will re-run that test this evening and confirm the findings.

A full schematic including your power supply and transfomer will help?

transformer --> full bridge rect --> cap (270 uF). I don't have any more information than that.

Can you please tell us your electronics, programming, arduino, hardware experience?

I would say I am a well versed amateur. Significantly better with on the programing side than hardware. Probably have 10-15 arduino projects running simple things like my soldering iron, to the most complex system I built for extruding 3d printing filament. And I am always looking to learn, so feel free to talk over my head, if I dont understand I will ask.

Wire one coil direct -test

by direct, you mean ps straight to the coil? That's a good idea, I will try it tonight.

Rewire with mosfet then drive mosfrt full on hard. That is max powersupply voltage to the base

are asking me to drive the mosfet directly? That I have not done, but when I drive a pin high (255), I see 5V at the opto & 10.7v at the gate of the mosfet. At that point the filter cap reads ~50vdc and the coil reads ~40vdc. Does that not provide the same information as the test you purposed? If not, it sounds like you want me to hook 12v straight to the base, not sure how that's going to be different, but I will give it a try.

Rewire base from power supply to have 12v power direct drive base as a switch full on 12v to the base- test

not sure I understand the difference between this test and the prev one.

I see the incremental approach you are taking to troubleshooting, but doesn't the test I outlined above confirm those gates are driven properly?

Also are you asking me to drive the mosfet with 12v & switch 12v (as apposed to the 40-60v supply) ? why? to ensure too much current doesn't hit the coil during the test?

Need datasheet.

what component do you need a data sheet for? If its the heaters, they are hand wound lengths of kanthal, ~30 ohms each. That's all I have to give you.

--worth mentioning: I have thoroughly looked for shorts between coils or the frame and found nothing.

looks like people want more specific question, understandable, so let me ask this:

is the 10.7v @ the mosfet gate(base) an appropriate voltage to turn the mosfet all the way on while maintaining low resistance (prevent mosfet from heating up).

is the 10.7v @ the mosfet gate(base) an appropriate voltage to turn the mosfet all the way on while maintaining low resistance (prevent mosfet from heating up).

Based one the datasheet, yes, as the on resistance is given as 0.16 ohms with a Vgs of 10 volts.

With the output = 255, what’s the voltage across drain to source?

Have you measured the ac ripple on your supply?

Hi,

transformer --> full bridge rect --> cap (270 uF). I don't have any more information than that.

270uF will not give you anywhere needed the filtering you need.

What is the DC voltage across the drain - source of the MOSFET when the MOSFET is turned ON?

Switch the DMM to AC mode and measure the DC power supply with and without load.

Without sufficient filtering your DC readings will only be the average of the power supply output, which will have a significant amount of AC ripple on it?

5 x 30 Ohms in parallel is a total load of 6 Ohms.
60V / 6 Ohms = 10 Amps needed from your power supply.
Did you do any calculations for the value of filter capacitor required to provide an adequate level of DC?

Tom... :slight_smile:

CORRECTION: cap is 470uf --sorry I built this back in march so I don't remember every detail and its hard to visual see.

Based one the datasheet, yes, as the on resistance is given as 0.16 ohms with a Vgs of 10 volts.

thank you, at least I got something right. :slight_smile:

Have you measured the ac ripple on your supply?

if this is done by using the ac setting on a dmm, then yes. Once prior to posting on here and I swear it was 0v measured on AC (that's what I would be look for right?). But since some of you asked me to, I measured it again right now, with no load at all, the filter cap shows 63.0vdc and 58.5vac. That does seem like a problem!

With the output = 255, what's the voltage across drain to source?

if its just 1 coil, my filter cap shows 46vdc and Vds =36vdc.
With all the coils fully on, filter cap =42vdc, vds =25v

note: In saying 5 coils x 30 ohms, I was simplifying. The resistance varies on each coil. The effective resistance of all 5 coils is 3.92 ohms.

Now 60v is the no load voltage, I am never going to see that in operation, right? I believe 42vdc is the most I will get on that tap... so 42 / 3.92 = 10.7 amps. I think the transformer is more than capable of producing that.

270uF will not give you anywhere needed the filtering you need.

does the corrected value, 470uf, do me any better?

Did you do any calculations for the value of filter capacitor required to provide an adequate level of DC?

I can hear you snickering at me as you ask that question, lol. No, I have no idea how to approach calculating that.

Without sufficient filtering your DC readings will only be the average of the power supply output, which will have a significant amount of AC ripple on it?

ok so that's ~42vdc, with ripples extending to to 60v correct? That would be fine if 42v got to the coil, but I only see 25ish.

thanks for all the help guys, I feel like I am getting someone and if nothing else, learning.

so I just checked for shorts again and now I have coils touching one another :frowning:

when I lit one of the coils that was not touching another, I saw 51.2v @ the cap and 51.2v drain-source, which seems totally reasonable to me. Though I did see a voltage of 1.6 on two of the coils that were off at the time, is it possible that is an induced voltage from the adjacent coils? and the table is all cast iron, so maybe its like a crude transformer?

I think what is happening is as it heats, those vague shorts between adjacent coils come and go leading to a difficult to diagnose problem.

might be day or two, those coils are cast in silicone, so Ill have to dig em out and recast it.

thanks again for the help, I am always blown away at the willingness to help in forums like this!

If you have supply voltage 51.2V from drain to source, the MOSFET is NOT conducting. I don't see how your opto coupler connection could work anyway. This should work.
New-Project.png

New-Project.png

Hi,

What is the DC voltage across the drain - source of the MOSFET when the MOSFET is turned ON?

Put the negative of the DMM on the source or gnd/negative of your power supply and positive on the drain of the MOSFET and measure the DC volts with MOSFET ON and OFF.

Thanks.. Tom.. :slight_smile:

finally finished the repair work on the coils. I think all my problems were coming from there. I probably tried to pack to much wire into the heater causing sporadic shorts within coils and between adjacent coils. The more interesting problem i found was several breaks in the nichrome....but the silicone was keeping the broken ends more or less aligned creating a shotty / changing connection.

So now I have my 60vdc peak power supply, loading the ps drops me to a very stable 42vdc....the math on that checks out. And on my coils I can range the voltage between 0-42 vdc.

Still running through some test sequences, but things are looking good.

Thank you all for putting some thought into it, much appreciated.

ps what was the verdict on flyback diodes?

What was the verdict on flyback diodes?

You don't need them, while any coil of wire has inductance the inductance of your heating coils is insignificant in the application you have.

I have reservations about the way you've done this as per my previous comments, however, it's easy for me to say that with my experience of electronics. I guess what is more important if you got it working to your satisfaction and you learned something.

++Karma; // For a successful project.