How can I get the most from my TIP121 transistors?

[AlainB]

Hi,

As you can see in the picture, this controller is quite rudimentary. The basic unipolar phase sequence is provided by a L297 IC, a marvelous and very versatile chip that every beginner should know about in my opinion as a newbie myself too, and 4 TIP121 darlington transistors are used for the power stage. There is no external resistors or diodes.

As it is, with this 8V, 0.8A motor driven full step at 12 volts and a ballast resistor of 2,4 Ohms I can obtain a G00 speed of 180 RPM with a good usable torque. I don't know the rating but at that speed, I cannot stop it from spinning with my fingers. I can't stop it either at 200 RPM but it is starting to spin a little rough. At 225 RPM I can stop it and at 250 RPM there is about no more torque.

Now for the questions!

With this particular transistor, would it be better to put a resistor between the logic and the base of the transistor. A computer power supply (5V) is feeding the logic part. If needed, what value should have this resistor?

I tried to put external diodes (1N4001) between the collectors and emitters, in the proper polarity, but all I got with this setup was a small humming from the motor with no spinning at all.

What kind of protective diodes should I use and where should I put them. Please indicate if possible the values or numbers.

Ultimately, I would like to use this design with some 2,25V, 2,25A, 70 on. torque recycled Genicom printer motors. I would like to feed them at 12 volts with ballast resistors of around 4,3 Ohms per phase, that is if the ratio is not too much for a 2,25 volts motor. By the way, what is the maximum ratio? I know that it is not the same as the ratio for a chopper board.

Thank you for reading this long post up to here!

Alain

TIP121:
http://www.ortodoxism.ro/datasheets/...onics/4128.pdf

[vtxstar]

Looks much like something I have on my bench - without the L297 though. I also breadboarded a drive set for my BOE (Board of Education) and wrote some code to have it turn a motor. I guess that qualifies us for the "Motor Spinner" award. I digress...(ahem).

If you can manage it, try putting the diode across the motor windings rather than the collector and emitter of the transistor. This is same principle as putting a diode across the windings of a relay to eliminate the inductive "kick" due to the collapsing field. The cathode of the diode should be connected to the positive voltage side of the winding. The anode the other end toward the lower potential (ground, return or earth - most likely the collector of the transistor).

That should allow your motor to spin again and protect your transistors while you work out the rest.

The base resistor needed is a function of the beta of the transistor. This is more important in a biased amplifier arrangement and not with the saturated switch condition that most of us hobbyists use. Now you've got me thinking about those transistor classes I had years ago. Most important in this situation is Ic max, maximum collector current, and the maximum power dissipation of the device. Not only does the resistor in the collector path limit the current, it also improves the time constant of the circuit providing a better current waveform.

An excellent text is "Electric Machines , Drives, and Power Systems" by Theodore Wildi. I have an older 5th edition. Stepper motors are discussed in chapter 19 of that book.

I will have to dig a little more for a better analysis of your circuit.

EDIT: I looked over the datasheet for the transistor. You don't need the external diode. They are internally protected already. This is like the Rohm D1783 transistors I am playing with.

[sswitaj]

Just quickly glancing at the L297's data sheet, it looks like you'd want current-limiting resistors in each line between the L297's output and the transistor bases.

Right now, to the L297, the load presented by the transistors looks like two diodes to ground (about a 1.4v drop) there is no current limiting element I can see (unless there's one intrinsic in the chip, but it looks like a conventional totem-pole output to me), so when the IC tries to go high, it tries to pull it's output to 4-5 volts. The transistor input acts like a clamp that sinks a lot of current above 1.4v.

This is not optimal for either the IC or transistor.

A ballast resistor, say 510 ohms, between the IC outs and transistor bases would limit the current flowing in that branch to about 5mA, plenty enough to get a TIP 121 (which has a whole lot of gain) to turn on solidly, but a small enough value to not stress the L297.


As Vtxstar pointed out, some "freewheeling" diodes, connected in "reverse" so that their anodes went to the transistor collectors and their cathodes (the striped end) went to the +12v supply, would go a long way toward increasing the service life of any transistors you use, because they would shunt the high-voltage spikes generated every time the transistors switch off.

One other factor you might want to consider in such a low-voltage design is that the TIP121 is a darlington transistor. That means that in effect, it is internally two transistors wired in series. While this makes for great gain, it also means that the saturation voltage (the voltage across the device when it switches on) is fairly high, about 1.4v minimum.

If you were to substitute a more modern transistor, like the BDW93, with a single stage hfe of 1000 (below 1 A), you'd get a Vcs(sat) of about .2 volts.

In english, that means 1/6th of the voltage drop and power dissipation you're seeing now. Your dynamic voltage headroom would suddenly get 35 % better, from 2.6v (12v supply - 8v motor - 1.4v transistor) to 3.7v (12v - 8v - 0.3v).

I don't know if this would result in a significant performance difference, but if you were building a "real" product, the dissipation reduction in the transistors might allow you to loose the fan and heatsinks, a significant cost element in many designs.

[vtxstar]

To add to sswitaj's comments....after looking at the L297 app note (AN470) it states that the device "...can be used with any power stage, including discrete power devices (it provides 20mA drive current for this purpose)."

The ST author did not provide an example diagram for driving discrete devices - exactly what you are doing AlainB. More reading shows an example of driving a ULN2075B Darlington array - a piece part type of what you are doing. The array chip does have a base resistor in series with the pin connection. In other words, inside the chip between the pin and the transistor. "Rin" is 350 ohms for the ULN2075B. sswitaj's 510 ohms gives a good safety margin.

Geoff Williams book "CNC Robotics: Build you own workshop bot" has several of the app notes for the L297 (and L298 sister chip). His controller is essentially straight out of the SGS-Thomson app notes.

This got my curiosity up more and I have been paying more attention to the old HDD and FDD boards piled up around the shop here. They are loaded with assorted motor drive chips, many with house numbers from ST or who ever filled the order for the board vendor.

Some careful scrounging on ancient FDD should provide useful parts. Then again I just saw some NOS motor drive chips on eBay listed for 99 cents. <sigh>

Revisting the "damper diode" I found this post by a very active member...
POST TO READ

This is probably noticed by those trying to make their steppers really zing along.
I think I would go with a servo drive if I needed high end rpm.
Heck, it's a hobby anyway.

[AlainB]

Hi,

Thank you for your very elaborate replies.

I tried one of my 2,25 volts 2,25A motor at 12 volts with 2 ballast resistors of 4,3 Ohms/25 watts. I can get a good usable torque at 350 RPM but the motor is kicking every now and then, maybe one little kick per two or 3 revolutions. But it is not loosing step. I had it run a 100 line g-codes file for hours and the marks that I made on the shalft still enline perfectly with the one on the body.

At 325 RPM it is running very smooth and strong, no kick.

While running, at different speeds, the motor get hot, not very hot, maybe 40-50 C. It would get very hot in about 30-45 minutes if on standby, 2 phases energised. The transistors are not heating a bit, that is with the fan and heat sink. I did not tried without them.

I suppose I could get more from the motor but I am happy with that for now. I use an antique 486, 25Mh laptop, with no math co-processor wich limit me to TurboCNC Ver.3. So 350 rpm at 400 steps per revolution make 2333 steps per second. I wonder if the laptop can cough up much more than that.

Anyways, I would like to continue with these TIP121. I have 20 of them and I want to use them. I will put the 510 Ohms resistor as suggested. For the diodes, if you could tell me what kind of diodes I should use, more specifically what value or what reference. I got this 1N4001 diode number from various designs on the web, not because I know something about diodes values. Any help to improve my driver will be much appreciated.

Alain