I have a cook/siring setup for stirring tomatoes while they cook down into sauce....the motor i'm running is a surplus store item, so I dont have much info on it, but it seems to run well at 24Vdc and draws ~2.5A. Even with a small, open air(not ducted) fan, It got pretty hot in an hour of continuous (185F on the IR therm) and needs to run for 6-8 hours at a time. I figure I can run it on a duty cycle to keep it cool, maybe even throw a thermistor on the motor and monitor /adjust it that way...and as a side benifit I can PWM to adjust speed....cool....simple project right? not for me
I am using a board with the atmega328 chip
power supply is 32V (i intend to PWM down to 24V max)
Initially, I was using a TIP120 NPN 5A transistor: base connected to pin10 via a 220ohm resistor & an LED indicator, collector out to the motor, emitter--> ground, a fly back diode from e->c and +Vps --> motor. I have burnt 5 or so of these chips screwing around with it and that was the last of those chips. I have on hand some TIP31 NPN 3A transistors, given that I had burnt a 5A TIP120, I figured I would run a gang of 3 of them to handle the load. so I branched the pin10 output to the 3 bases, which did not run until I pulled the resistor, bridged the 3 collectors via a heat sink and added the flyback diodes to each. That heat sink got very hot, very quickly and fried those chips!
when they are burnt, it seems to perm. close the connection between the collector & emitter as I get a voltage just a bit less than the full power supply & the motor runs....this leads me to believe I may be running too much current through the base? At one point, I did try a larger resistor on the base, but it still burnt.
just looking for any advice, I'll bet I'm doing something stupid.....
if additional info is needed, please ask and thanks for your help!
Nowadays its so cheap to find a decent high current buffer I would not bother with building one myself.
More serious, you can burn your arduino without opto-isolation.
I am not a fan of the transistors either. Great little devices back in 1947, but with an age of 67 I think its time to start thinking about Mosfets, unless the characteristics are not really ideal for some particular reason.
Using a larger base resistor on the base will limit current at the expense of operating in the linear region - this means heat.
Using an opamp with a sense resistor to drive them can limit inrush currents too. If you really need to use transistors and can use one ampop to drive each, set to a given current brilliant! Thats saves a lot of power on the emitter resistor too!
To finish I assume you do know how to parallel transistors and have followed the steps to ensure the current is distributed evenly between them.
More serious, you can burn your arduino without opto-isolation.
Crap!
I am not a fan of the transistors either.
Your a prize idiot. No one has yet beaen the transistor. The chips used in the arduinos have 100,000,000 of transistors in them along with umpteen diodes. A MosFET is a transistor - look it up oh stupid one - the 'T' is transistor!!!
Using a larger base resistor on the base will limit current at the expense of operating in the linear region - this means heat. Using an opamp with a sense resistor to drive them can limit inrush currents too. If you really need to use transistors and can use one ampop to drive each, set to a given current brilliant! Thats saves a lot of power on the emitter resistor too!
Read what you have written is it opamps are constructed from transistors . What your trying to recommend? May be a Pentode or two!
Opps - why don't you look at the design of an opamp - Oh look it is made up from --- TRANSISTORS!.
Forgive my lapse, I meant Bipolar Junction Transistors in particular.
While you may have your point of view I think your post is less than constructive.
Perhaps we want to focus on solutions for the OP rather than criticizing others opinions. I use BJT's for many applications where convenient. Power switching is not one of them.
casemod:
Nowadays its so cheap to find a decent high current buffer I would not bother with building one myself.
More serious, you can burn your arduino without opto-isolation.
I am not a fan of the transistors either. Great little devices back in 1947, but with an age of 67 I think its time to start thinking about Mosfets, unless the characteristics are not really ideal for some particular reason.
Using a larger base resistor on the base will limit current at the expense of operating in the linear region - this means heat.
Using an opamp with a sense resistor to drive them can limit inrush currents too. If you really need to use transistors and can use one ampop to drive each, set to a given current brilliant! Thats saves a lot of power on the emitter resistor too!
To finish I assume you do know how to parallel transistors and have followed the steps to ensure the current is distributed evenly between them.
I am not dead set on transistors, its all I have on hand that will hand the load....meanwhile I have about 30 gallons(and increasing) of tomato pulp needing to be cooked, so I really dont have time to wait for comps to be shipped.
your last assumption is incorrect, Ive never done parallel transistors before, so I just did it in what I saw as the simplest way. I will look into that & the opamp setup you mentioned, both are pretty new to me.
so I guess I am interested in any additional suggestions for a) getting this circuit running now, with what I have and b) what the best way to control this motor is so I can build that circuit for the long term.
@OP,
My guess is that the coup de grace was removing the current limiting resistor on the base of the transistors.
The digital output pins can only source about 40 mA and that is less than the max current for the base of those transistors so that is a bit of a mystery. Mosfets work better in parallel than transistors but I would recommend one of these :
casemod:
Nowadays its so cheap to find a decent high current buffer I would not bother with building one myself.
More serious, you can burn your arduino without opto-isolation.
I am not a fan of the transistors either. Great little devices back in 1947, but with an age of 67 I think its time to start thinking about Mosfets, unless the characteristics are not really ideal for some particular reason.
Using a larger base resistor on the base will limit current at the expense of operating in the linear region - this means heat.
Using an opamp with a sense resistor to drive them can limit inrush currents too. If you really need to use transistors and can use one ampop to drive each, set to a given current brilliant! Thats saves a lot of power on the emitter resistor too!
To finish I assume you do know how to parallel transistors and have followed the steps to ensure the current is distributed evenly between them.
I am not dead set on transistors, its all I have on hand that will hand the load....meanwhile I have about 30 gallons(and increasing) of tomato pulp needing to be cooked, so I really dont have time to wait for comps to be shipped.
your last assumption is incorrect, Ive never done parallel transistors before, so I just did it in what I saw as the simplest way. I will look into that & the opamp setup you mentioned, both are pretty new to me.
so I guess I am interested in any additional suggestions for a) getting this circuit running now, with what I have and b) what the best way to control this motor is so I can build that circuit for the long term.
Well, BJT's have different gains. In the old days you used to buy matched pairs for radios (germanium type), but they all slightly vary.
This means that if you have more than one transistor the current wont be distribute evenly and you may end up with 10A in one and 1A each in the remaining two, instead of 4A each as you would originally assume.
To sort this you either use a emitter resistor or a opamp active driver.
With an opamp, the transistor is set in current mode, making sure the current never goes above the maximum limit for the individual transitors. If the current is small, the one with the highest gain may still be carrying more current that the remaining, but when the current passes that limit, aditional flow is routed to the other devices.
The resistor is a bit less sophisticated method. Basically VBE is decreased due to a floating GND and therefore the gain for that particular transistor is reduced keeping the current on a safe level. However you loose more energy on the emitter resistor that using one with an opamp that can act on mV
The opamp also sets a given current limit. If it is 5A per transistor it is 5A, while the resistor method will never be that accurate. You can more or less share the load, but you can also overload the three and burn them all. It is however much simpler.
You have to do some reading to learn how to calculate the actual value for your actual case with the method of your choice
This is relevant to start learning
To get going that is what you need to do. Also make sure you have enough base current so that the BJT's do not operate in the linear region. You can pre-amplify the base circuit with another transistor in a similar fashion as shown below:
Of course you could just use an PSMN5R6 MOSFET which is good to 100A @ 100V with an IR2121 mosfet driver set to operate with a current limit and of course the reverse diode for the motor. The whole package should cost you about $10. You can even opto-isolate the arduino as the driver current drawn at the input is negligible. I guess thats the most efficient way to control the motor in the long term.
Two examples above. Just change your components for the higher voltage (Mosfet and a 7815 regulator to feed the IR2121)
If it's a only matter of turning a "tomato stirring motor" on for several hours then you should use a relay.
Pretty simple.
Industry has been using them for years and still.
You didn't mention adequate heatsinking for your TIP-xxx transistors. (Probably "None", right?)
so I branched the pin10 output to the 3 bases, which did not run until I pulled the resistor, bridged the 3 collectors via a heat sink and added the flyback diodes to each. That heat sink got very hot, very quickly and fried those chips!
The point is that bipolar junction transistors simply do not share current well. It is all to do with the temperature coefficient.
When a transistor gets hot it will conduct more current, as it conducts more current it gets hotter and so conducts more. This is called thermal run away. When you have several transistors in parallel one ends up taking too much current and then pops, the very quickly the next and the next.
You can mess about with resistors like in the above circuits or mess about putting them on the same heatsink, but ultimately all those tricks will fail. The only sure fire way is to make the transistors on the same piece of silicon so they all run at the same temperature.
However there is no need to do this nowadays. If you use FETs they have a temperature coefficient that goes the other way. The hotter they get the large is the resistance so the current drops and so it does not get as hot. A collection of FETs will share current.
However most of the time you can get a higher current FET so there is little point in sharing, but you can if you want.
"Ganging" junction transistors, of whatever stripe, is not a good idea. (I guess I better add "IMO").
"Ganging" outputs on ULN2803s is OK because they are on the same die. (That willnot help the OP.)
Junction transistors are depicted on ICs as "ganged", but, again, they're on the same die.
"Ganging" discrete FETs is acceptable.
Different situations.
I see that someone points out that El OP did mention heatsinking, afterall.
I'd like to see a picture of it.
If it's a little tab-mounted fin job, then forget it.
I think that all of these thorny power electronics considerations can be dispensed with by using a relay.
One thing to remember with MOSFETs is that if you gang them each must
have its own gate resistor to prevent high freqency oscillation modes where
neighbouring devices interact(*). A single MOSFET and the gate resistor is
there to protect the Arduino pin (or can be omitted if driving the gate from something
able to handle large capacitive loads).
(*) So I'm led to understand by a rather elderly databook on MOSFETs.
Just a warning: While it is true that the MOSFET increases resistance with temperature it is also true that the I2R losses shoot up. Therefore, in Practice, it can run into thermal runaway much easily than a BJT if the heatsink temperature is allowed to go too high.
For a nominal RDS(on) of a few mOhm it is unlikely that the current decreases at any significant rate, however twice the resistance = 4x more heat losses for the same current.
MarkT:
One thing to remember with MOSFETs is that if you gang them each must
have its own gate resistor to prevent high freqency oscillation modes where
neighbouring devices interact(*). A single MOSFET and the gate resistor is
there to protect the Arduino pin (or can be omitted if driving the gate from something
able to handle large capacitive loads).
(*) So I'm led to understand by a rather elderly databook on MOSFETs.
Correct, the gate resistor is there to avoid ringing due to track capacitance and also to prevent the switch off event to be too quick. On high power devices switching off to fast, for example, in the event of a short circuit, will result in any inductive load generating a high voltage that may break the VDS insulation and ruin the device. This is the reason to use snubber capacitors. have a "google".
But this does not imply you can get away without it if using only one. Same principle applies.
Assume a suitable driver charging the gate at 1 or 2Amps with a 4.7Ohm resistor. If the FET fails, and they usually fail in short, I can assure you the driver wont survive. For an arduino at low frequency, perhaps one can get away with it, but rated RDS(on) will not be achieved.
