Hello Al
In your search of a big Watt Resistor, the most common to use is a radiator/owen ore a big Watt Lamp, a wall mounted radiator of 2000W with several steps is absolutely the best, but the most fancy is to use a big Bulb Lamp/Halogen, then you can see how much ampere the motor is using.
Across the Radiator/owen ore bulb you can use a start relay on the AC Volt side or a FET on the DC Volt side.

It was very common in the old time to use Radiator/Owen elements as resistor, by the way you can also use a Iron,

If this explanation is a bit to short, i have more info.

All the best
Sverri Jensen

 

@R!f@@ Sorry, wasn't thinking. I am used to calculating UTC from my local time, in which case at 0500 local I add 5 to get UTC.
John

 

Dammit! Just blew the 80N60! just the first second of switchon was enough to blow it, no smoke or heat but its now an expensive jumper.
Looks like maybe I will have to try measuring the actual current as Max said. Can you recommend a voltage to test at? Would a 12v Pb battery do or do i need 5v or whatever?
l

I am not sure why you would have the problem if you are controlling the motor with a PIC via PWM if you use a ramp up speed control?
There is no need to use wastfull resistors etc, these would only Maybe be used where there is no control in place, the alternative if determined to apply full PWM at switch on is to place a current sense detection in the source of the Mosfet.
You only need a couple of volts to test the armature resistance.
Also you did not confirm if the flywheel was removed?
Did you test the motor on a source such as an automotive battery to test.
Max.

 

Hi to all,

I have a treadmill motor from a mate that I want to use on my lathe with speed control. It is 180vdc brushed perm mangnet (2 wires only) and 2.5hp continuous rated. I can do the control of speed ok using a pic micro and a pot driving PWM to a FET Driver and then 600v 40A FET without issues and can probably cobble together some kind of feedback for speed stability under load. BTW I'm in the UK and power here is 220vac.

My problem is that the startup current from static on the motor is hundreds of Amps, seeing as the windings are only about 5ohms. The motor is happy to run on unsmoothed full bridge rectified dc and the PWM works as it should but only with a 10 or so ohm resistor in series or the fet would die immediately on startup.

I know that many industrial motors have multiple series resistors that are switched in as needed mainly by mechanical means but I aslo know that on the original treadmill boards this is not how it works. I can't find out how the treadmill boards manage to limit the startup current and there seems nothing on the web about it. Anyone able to help?

Regards, Al

I think if I were to do it I would use hardware. Part of that is that I don't have a lot of experience with micros.
I would do a couple of things:
Isolate the high voltage from the circuits with opto isolators.
Do soft start at both power up and with the adjustment.
Add current limit.
I will draw one up if you think you might want a hardware solution.

 

Max, Yes the flywheel is off and the motor works fine on battery - tried it up to 96v (8 x 12v lead acids).

Ron, If you have a different slant on this I would happily try your way. I'm on my last big IGBT now too. Doh! So if you feel like drawing something up I would be pleased to try it. I'm always open to different ideas.

Thanks, Al

 

I'm on my last big IGBT now too.

Thanks, Al

Are you ramping it up or switching direct?
PWM?
Max.

 

Hi Max, I have the duty cycle for the PWM starting at 0 and even the speed pot was at the zero end which gives about 1%, but even that instantly killed the IGBT 80A 600v. Doh!

I just found both my meters out and managed to test the current with motor locked. At 3.86v it draws 0.93A What does this mean in relation to the expected current from my earlier 5ohm measurement? Is it much off?

Here is my basic code for the driver.

dim speed as byte
dim runup as byte
dim l as byte

main:
' Main program

ANSEL = %00000001 ' Configure AN pins
TRISA = %00000001 ' configure PORTA pin 0 as input
TRISB = %00000000 ' designate PORTB pins as output

PWM1_init(20000)
PWM1_START
runup = (ADC_Read(0) >>2) ' Get 8-bit results of AD
for l = 0 to runup
delay_us(1)
PWM1_SET_DUTY (l)
next l


portb = %00000000
while (TRUE)
speed = (ADC_Read(0) >>2) ' Get 8-bit results of AD

PWM1_SET_DUTY (speed)

wend

end.

Al

 

Hi Max, I have the duty cycle for the PWM starting at 0 and even the speed pot was at the zero end which gives about 1%, but even that instantly killed the IGBT 80A 600v. Doh!

I just found both my meters out and managed to test the current with motor locked. At 3.86v it draws 0.93A What does this mean in relation to the expected current from my earlier 5ohm measurement? Is it much off?

Al

There must be something else going on, can you post the circuit?
The armature resistance sounds about right.
Max.

 

Hi Max, I will draw it out. I just make things on vero board and figure out as I go.

Al

 

Hi Max, I will draw it out. I just make things on vero board and figure out as I go.

Al

No pre-sketch?
You could also try the original with a small load such as a low wattage lamp.
p.s I only program in Assembly.
Max.

 

Hi Max, No, never did get the pre sketch discipline, just tend to draw what I want on something called VeroDes since I usually build on vero before going to the trouble of etching a (wrong) board. I usually muddle through as its fairly easy to move and change things on it. Its only drawback is you cant zoom in so the markings get crazy with more than a dozen components close together.

I have tested the controller with varying loads, 25w bulb, 200w bulb and both worked flawlessly and caused no damage or even warmed anything up!

I am now struggling to find a drawing package that I can do a quick setup on to post here but so far have failed, tried KiCad (no pic16f88 and it doesn't have a generic 18pin component that I can find) and SmartDraw 2016 (couldn't find ANY components that were even vaguely useful) and just trying Express Sch now. Looks most promising so far but don't hold your breath for it. LOL

EDIT: Also if all works ok with a simple drop resistor, what can be wrong? How can 10ohms make everything work perfectly even on a much smaller spec FET? I mean the ONLY thing that even got warm was the resistor, the FET stayed cool and so did its driver and associated components. Makes no sense to me!

Hope to post diag shortly,

Al

 

Kicad has a few 18pin Picmicro either to use or very easy and fast to reassign to a new part. 16f54 for example.
Your right, doesn't make sense if works with other loads.
Max.

 

Well ExpressSch was pretty good! Missing Max627 but easy enough to make components so that's ok.

Here it is.

 

Hi Max, No, never did get the pre sketch discipline, just tend to draw what I want on something called VeroDes since I usually build on vero before going to the trouble of etching a (wrong) board. I usually muddle through as its fairly easy to move and change things on it. Its only drawback is you cant zoom in so the markings get crazy with more than a dozen components close together.

I have tested the controller with varying loads, 25w bulb, 200w bulb and both worked flawlessly and caused no damage or even warmed anything up!

I am now struggling to find a drawing package that I can do a quick setup on to post here but so far have failed, tried KiCad (no pic16f88 and it doesn't have a generic 18pin component that I can find) and SmartDraw 2016 (couldn't find ANY components that were even vaguely useful) and just trying Express Sch now. Looks most promising so far but don't hold your breath for it. LOL

EDIT: Also if all works ok with a simple drop resistor, what can be wrong? How can 10ohms make everything work perfectly even on a much smaller spec FET? I mean the ONLY thing that even got warm was the resistor, the FET stayed cool and so did its driver and associated components. Makes no sense to me!

Hope to post diag shortly,

Al

What is the part number of the clamp diode across the motor? All the other load are resistive.

 

Looks like I managed to upload the image albeit a bit large.

The diode is STTH8L06 and was the best I had.it seemed ok with the sheet unless I read something wrong.

Al

 

Looks like I managed to upload the image albeit a bit large.

The diode is STTH8L06 and was the best I had.it seemed ok with the sheet unless I read something wrong.

Al

Looks good to me.
The only other thing I can think of is decoupling of Vdd for the driver and perhaps a 10 or 20 ohm resistor between the driver and the gate to reduce ringing.
Do you have a scope?

 

Hi Ron, Sorry mate, I simply forgot the decoupling on the schematic but not in the real thing. There is decoupling on the 5 and 12v lines of 100n and 1u with an extra 120u on the 12v to boot. Also decoupled are the 7805 and 7812 not shown on the drawing, I am fairly well trained over the years on only a few things but decoupling is one of them, mainly as it is cheap to do and can save problems.

As for the resistor, you had me wondering if I'd left it of the schematic too but it is there. 10ohm as recommended by most of the FET sheets.

Got to a stage now where if someone can't find the reason for my problem I will have to stop experimenting until I get some more big FETs or IGBTs. this circuit blows them like an army bugler at reveille!

EDIT: No scope either, lent it to a mate who killed it a while back.

Need a miracle I think, divine intervention at the very least, Al

 

Hi Ron, Sorry mate, I simply forgot the decoupling on the schematic but not in the real thing. There is decoupling on the 5 and 12v lines of 100n and 1u with an extra 120u on the 12v to boot. Also decoupled are the 7805 and 7812 not shown on the drawing, I am fairly well trained over the years on only a few things but decoupling is one of them, mainly as it is cheap to do and can save problems.

As for the resistor, you had me wondering if I'd left it of the schematic too but it is there. 10ohm as recommended by most of the FET sheets.

Got to a stage now where if someone can't find the reason for my problem I will have to stop experimenting until I get some more big FETs or IGBTs. this circuit blows them like an army bugler at reveille!

EDIT: No scope either, lent it to a mate who killed it a while back.

Need a miracle I think, divine intervention at the very least, Al

And when the load is a lamp you can see that it takes a second or so to brighten up?

 

Hi Ron, Yes the slope is easily visible. My first test was actually too long a ramp so I sped up the loop until it took about 2 seconds to get from 0 to full.

Al

 

Try running the motor on 24 volt as it probably will turn ok if it is unloaded at that voltage. It may in fact run with no load on the motor at 12V.
Get it all working ok on the lower voltage first. Then see how the high volts goes, but first via a lamp to limit the current.
A handy trick taught to me years ago by a TV repair man is to insert a 240V light globe in series with the supply to test. If you have a short, the lamp will light and hopefully protect your FET and diode. I always run switch mode power supplies I fix via a lamp before running them directly on the mains. It saves lots of smoke and shrapnel (not to mention it is easier on the nerves)
Did only your FET blow or the diode too? If the diode is dead, it is not rated at high enough current.